ArticlePDF Available

A Comprehensive Survey on Age of Information in Massive IoT Networks

October 2022
Computer Communications 197(4)

October 2022
197(4)

DOI:10.1016/j.comcom.2022.10.018

Authors:

Qamar Abbas

National University of Sciences and Technology

Syed Ali Hassan

National University of Sciences & Technology

Hassaan Khaliq Qureshi

National University of Sciences and Technology

Kapal Dev

Cork Institute of Technology

Show all 5 authorsHide

Ambient intelligence (AmI) represents the future vision of intelligent computing that can bring intelligence to our daily life through various domains. In such applications, AmI is often subject to the freshness of information collected, which is commonly quantified by a relatively newer metric called age of information (AoI). In the data aggregation and analytics for Internet-of-Things (IoT) AmI, AoI should be well managed, because information update should be as timely as possible to achieve optimal performances. AoI has been studied in various applications using different queuing policies, scheduling algorithms, and multiple access schemes, in which each component of communication and information systems are designed and analyzed to improve the AoI. This paper provides a comprehensive overview of literature on the AoI and its variants in large-scale networks. AoI in IoT systems depends on the arrival rate at the source nodes, queuing policy adopted at the nodes, the scheduling of nodes for information transmission and the access scheme adopted by the nodes. To better design and operate the AmI applications that require the freshness of information, we discuss the impacts of the queuing policy, stochastic modeling, scheduling, and multiple access schemes. In particular, non-orthogonal multiple access (NOMA), which is regarded as one of the key technologies in beyond 5G and 6G, and it hybrid version combined with the conventional orthogonal multiple access (OMA) are discussed in the context of AoI. In addition, we identify promising research opportunities in potential age-sensitive applications. Thus, compared to the existing surveys on AoI, this paper provide more practical and up-to-date design guidelines for the applications with the information freshness requirements.

Content uploaded by Syed Ali Hassan

Content may be subject to copyright.

A Comprehensive Survey on Age of Information in Massive IoT Networks

Qamar Abbasa, Syed Ali Hassana, Hassaan Khaliq Qureshia, Kapal Devb, Haejoon Jungc,∗

aSchool of Electrical Engineering and Computer Science, National University of Sciences and Technology, Islamabad, Pakistan

bNimbus Research Centre, Munster Technological University, Cork, Ireland

cDepartment of Electronic Engineering, Kyung Hee University, Yongin, Korea

Abstract

Ambient intelligence (AmI) represents the future vision of intelligent computing that can bring intelligence to our

daily life through various domains. In such applications, AmI is often subject to the freshness of information collected,

which is commonly quantiﬁed by a relatively newer metric called age of information (AoI). In the data aggregation

and analytics for Internet-of-Things (IoT) AmI, AoI should be well managed, because information update should be as

timely as possible to achieve optimal performances. AoI has been studied in various applications using diﬀerent queu-

ing policies, scheduling algorithms, and multiple access schemes, in which each component of communication and

information systems are designed and analyzed to improve the AoI. This paper provides a comprehensive overview

of literature on the AoI and its variants in large-scale networks. AoI in IoT systems depends on the arrival rate at

the source nodes, queuing policy adopted at the nodes, the scheduling of nodes for information transmission and the

access scheme adopted by the nodes. To better design and operate the AmI applications that require the freshness

of information, we discuss the impacts of the queuing policy, stochastic modeling, scheduling, and multiple access

schemes. In particular, non-orthogonal multiple access (NOMA), which is regarded as one of the key technologies

in beyond 5G and 6G, and it hybrid version combined with the conventional orthogonal multiple access (OMA) are

discussed in the context of AoI. In addition, we identify promising research opportunities in potential age-sensitive

applications. Thus, compared to the existing surveys on AoI, this paper provide more practical and up-to-date design

guidelines for the applications with the information freshness requirements.

Keywords: Ambient intelligence (AmI), age of information (AoI), Internet of Things, status update systems,

scheduling algorithms, queueing policies.

1. Introduction

With unprecedented increase in the number of con-

nected devices and sensors in massive Internet of Things

(IoT) networks, the amount of data handled may pose

huge computational challenges. In this context, ambi-

ent intelligence (AmI), which refers to the future vision

of intelligent computing without explicit input and out-

put devices, can be an eﬀective solution to bring intel-

ligence to our daily life through various domains com-

prising elderly assistance, preventive maintenance, and

video surveillance [1]. With AmI, sensors and proces-

sors will be embedded into everyday devices and the

∗Corresponding author

Email addresses: qabbas.phdee17seecs@seecs.edu.pk

(Qamar Abbas), ali.hassan@seecs.edu.pk (Syed Ali Hassan),

hassaan.khaliq@seecs.edu.pk (Hassaan Khaliq Qureshi),

kapal.dev@ieee.org (Kapal Dev), haejoonjung@khu.ac.kr

(Haejoon Jung)

environment will adapt to the users’ needs and desires

seamlessly [2]. Therefore, AmI is intrinsically and thor-

oughly connected with the freshness of information col-

lected, which is commonly measured by a newly coined

metric known as the age of information (AoI). In this

context, AoI is a critical metric in the data aggregation

and analytics for Internet of Things (IoT) in AmI, be-

cause such systems require the delivery of updates in a

timely manner for their survivability and stringent re-

quirements of lower latency in the data aggregation and

status update.

The modern applications of IoT are being rolled out

in almost every industrial and vertical applications [3],

whereas human intervention is decreasing due to the au-

tomation in the applications and obsolescence of man-

ual systems [3]. The latest generation of networks en-

able a variety of applications, which have diverse re-

quirements for their survivability. The performances of

Preprint submitted to Computer Communication October 22, 2022

applications such as smart vehicles in intelligent trans-

portation system [4], smart devices in a body area net-

work [5], sensors in smart agriculture [6], smart homes

[7], and industrial automation [3] are typically subject to

the status update systems, which are highly susceptible

to delays in the network. Such delay-sensitive applica-

tions are enabled through ultra-reliable and low-latency

communications (URLLC) in 5G and beyond, which of-

fer signiﬁcantly low delays up to milliseconds. The met-

rics used to characterize the delay performance of these

status update systems in literature are packet delays and

inter-delivery time [8]. The packet delay is the total time

elapsed from the generation of a packet to its delivery at

the destination, while the inter-delivery time is the time

between the reception of two consecutive packets.

Similarly, throughput is also used in literature, which

quantiﬁes the total amount of information delivered in

a unit time. However, these metrics are not suﬃcient to

characterize information freshness in a network due to

conﬂicting requirements. In fact, the latency in a status

update system can be minimized by reducing the update

rate at the monitoring sensors, whereas throughput of

the system can be increased by maximizing the packet

generation rate at the nodes. In other words, latency and

throughput performances cannot be directly related to

the freshness the information transmitted from the sta-

tus update system to the monitoring system. Most of

the status update systems demand fresh updates and the

information extracted is required to be delivered to the

monitoring system in a timely manner. Ideally, the in-

formation should be reached to the monitoring system

right after their generation at the source, but the delivery

of such information from the generating source to the

monitoring system is challenging in practice, because

of limited network resources and channel randomness.

Therefore, to better characterize the freshness of in-

formation, a new metric age of information (AoI) is in-

troduced in [9], which is deﬁned as the time elapsed

since the delivery of latest information about a monitor-

ing system. Compared to latency and throughput, AoI is

a more appropriate metric for the status update systems,

because the updates received at the monitoring system

must be fresh for the survivability of the applications.

Diﬀerent from delay, AoI quantiﬁes the freshness of in-

formation carried by each packet until it is delivered to

the monitoring system. Because of such eﬀectiveness of

the metric, AoI has been studied in numerous literature

on applications which demand fresh updates. AoI has

been studied as a concept, as a tool, and as a metric [10]

to capture the timeliness of information in information

sciences. It has been employed as a freshness metric

in variety of applications such as intelligent transporta-

tion system [4], URLLC [11], smart agriculture [6], and

industrial automation [3] to name a few.

Motivated by the latest advancements in intelligent

IoT and increasing demands for timeliness in status up-

date systems, this paper summarizes the recent research

eﬀorts in deﬁning, characterizing, modeling, and opti-

mizing AoI in various applications. This includes, but

not limited to, analytical modeling, scheduling policies,

access schemes and their roles in optimizing AoI for fu-

ture IoT applications. Note that this paper is distinct

from the existing reviews on the AoI and its variant in

[12,13]. While [12] is focused on the mathematical

framework for the AoI analysis based on information

theory, this paper provides a detailed and comprehen-

sive overview of the design components of the prac-

tical age-sensitive networks. In particular, to reﬂect

the signiﬁcant research attention on recent multiple ac-

cess techniques such as non-orthogonal multiple access

(NOMA), this paper presents the impact of the multiple

access approaches including orthogonal multiple access

(OMA), NOMA, and the hybrid version of the two.

Moreover, the literature survey in [13] introduces the

value of information (VoI) in wireless sensor networks

(WSNs) and presents how to employ the concept of the

VoI in collecting, processing, and forwarding data from

the sensors. Diﬀerent from [13], this paper is not lim-

ited to the VoI within WSNs, rather it provides better

guidelines on the practical system design and operation

based on the characteristics and requirements of the tar-

get applications using various notations of AoI. In ad-

dition, we have summarized the key information of 46

core papers in Table 1, including the objective, network

scenario, and methodology. The taxonomy of this sur-

vey paper is also depicted pictorially in Fig. 1.

The main contributions of this survey are:

•AoI, its variants and their related existing works

are explained comprehensively.

•Various parameters aﬀecting AoI in IoT systems

are listed intensively with related existing works.

•Techniques used to model AoI such as queuing dis-

ciplines, Markov chain and stochastic hybrid sys-

tems are explained in detail.

•Existing works on scheduling algorithms and ac-

cess mechanisms for AoI optimization are ex-

plained in depth.

•Signiﬁcance and potential of AoI in various IoT

applications are listed with supporting existing

works.

Section 7

Applications of AoI

Smart Agriculture

Intelligent

Transportation System

URLLC

AR/ VR

Section 3

AoI with

Queuing Theory

Last-In First-Out

(LIFO)

First-In First-Out

(FIFO)

Section 4

Modeling of AoI

Markov Chain

Markov Decision

Process

(MDP)

Stochastic Hybrid

System (SHS)

Section 6

Impact of Access

Mechanism

Time Division Multiple

Access (TDMA)

Frequency Division

Multiple Access

(FDMA)

Non Orthogonal

Multiple Access

(NOMA)

Hybrid Scheme of

OMA and NOMA

Section 5

Impact of Scheduling

Random Scheduling

Arrival-based

Scheduling

Age-based Scheduling

Weight-based

Scheduling

Section 2

AoI & Related

Metrics

Average AoI

Peak AoI

Value of Information

VoI

Factors Affecting AoI

Figure 1. Taxonomy of the current survey paper.

The rest of the paper is organized as follows. Section 2

gives an introduction of AoI and its variants. In Section

3, we consider AoI modeling using diﬀerent queuing

schemes. Section 4 highlights the stochastic AoI mod-

els based on Markov processes, whereas Section 5 gives

a brief overview of scheduling algorithms to minimize

the AoI. Section 6 describes the impacts of the multiple

access mechanisms, and then the future applications to

apply AoI modeling are discussed in Section 7. Lastly,

the paper is concluded in Section 8.

2. Age of Information and Related Metrics

Age of information is deﬁned as the time elapsed

since the generation of latest received packet at the mon-

itor and it is characterized as a freshness metric in status

update systems. Consider a simple network with mul-

tiple nodes, a central access point and a remote moni-

toring system, as shown in Fig. 2. Each node, which is

equipped with a queue, transmits their information after

extracting from an environmental source to the monitor-

ing system through the base station. Visualizing a single

node from the network, suppose that gnis the genera-

tion time of the latest packet at time nat a source node,

which has to travel through the network to the monitor.

In this case, the AoI at the monitor at time nis deﬁned

as ∆(n)=n−gn. The age of a packet is incremented

by unity after every time step from its generation until

its delivery to the monitor. The age process is depicted

in detail in Fig. 3, where g1indicates the generation

time of ﬁrst update and x1shows the departure of that

update. The age of update is incremented by 1 in each

Monitor

Node

Queue

Figure 2. A simple status update system model consisting of IoT

nodes, a collecting BS and a monitor.

time slot till its departure. In Fig 3,U1denotes the time

of the ﬁrst update in the system from its generation to its

departure, while V1corresponds to the inter-arrival time

of the updates. It is noted that the third update is gen-

erated before the departure of second update, therefore,

the second update is dropped, because the third update

is the latest update to minimize the overall AoI of the

system. Some variants of AoI used in literature to char-

acterize information freshness include the average AoI,

peak AoI, and value of information, which are discussed

in detail in the following subsections.

2.1. Average AoI, ∆avg

The average AoI, denoted by ∆avg, measures the cu-

mulative age of the whole system in a time range. From

Fig. 3,∆avg in the ﬁrst Ntime step is deﬁned as

∆avg =1

n=1

∆(n),(1)

where ∆(n) is the age process. Equation (1) is applicable

for both lossless ﬁrst-in ﬁrst-out (FIFO) and lossy last-

in ﬁrst-out (LIFO) systems [12]. Lossy systems replace

the older packets when a fresh packet arrives or discard

the new arrivals when the queue is full. While lossless

systems do not drop any packet by accommodating the

arrivals in their inﬁnite size queues.

In Fig. 3, the cumulative AoI corresponds to the area

under the age curve. Therefore, the average AoI of the

system up to time Ncan be computed as the area di-

vided by N. The area under then age curve is a function

of packet arrival rate and departure rate, as illustrated

Table 1

Systematic literature review of AoI.

Ref Year Objective Network Scenario Methodology

[9] 2012 Optimization of status update rate to

minimize AoI

Status updates in V2V communication The update rate is optimized according to the

service rate in a FIFO system

[14] 2012 Minimizing queuing delay to minimize

AoI

General source and server LIFO queuing system is used with single

queue length with preemption and without

preemption to reduce queuing delay

[15] 2013 Analyzing AoI for random service rate

systems

A general source-destination pair Random status update age with exponential

service rate is analyzed using stochastic hy-

brid system

[16] 2014 Improvement of AoI through packet

management

A general source-destination pair Arrivals are dropped when server is busy to

minimize average AoI

[17] 2015 Minimizing queuing delay to minimize

AoI of multiple sources

A single node with multiple sources and a des-

tination node

Queuing management is used to drop stale in-

formation and replace with latest update

[18] 2015 Studying trade-oﬀbetween frequency of

status updates, and queuing delay.

A network with multiple source nodes and a

destination

Optimized AoI by formulating the problem

using quasi-convex optimization

[19] 2016 Investigating eﬀect of multiple servers

on AoI

A source and destination with multiple servers Analyzed the eﬀect of multiple servers and

concluded that increasing servers minimize

AoI but at the cost of obsolete packets

[20] 2016 Analyzing peak AoI in the presence of

packet delivery error

A source destination pair Peak AoI expressions for diﬀerent queuing er-

rors are derived

[76] 2016 Optimization of scheduling to minimize

AoI

Multiple transmitter and receivers a steepest age decent algorithm with better

scalability is developed.

[21] 2016 Studying impacts of buﬀer sizes, packet

deadlines, and packet replacement on

AoI

A source sending information to a destination

node through a M/M/1/k queue.

An optimal buﬀer size for a given packet ar-

rival rate is derived using simulations

[22] 2017 Capturing freshness and cost of update

delay

Packets are scheduled using M/M/1 ﬁrst come

ﬁrst out queuing in a status update system

Value of information is formulated to sched-

ule packets according to cost of update delay

[23] 2017 Minimizing AoI of an information of in-

terest

A multi-server sending information of a

source to the intended destination such as,

real-time stock service

A pull model is proposed to schedule the in-

formation of interest determined by the repli-

cated requests

[24] 2017 Formulating AoI over an unreliable

multi-access channel

A multi-user network sharing a common

channel to send information to a destination.

Intended applications are environmental sens-

ing and health and activity monitoring

Scheduled access and slotted ALOHA-like

random access schemes are compared for

packets scheduling; Showed the ALOHA

scheduling is worse than scheduled access

[25] 2017 Minimizing AoI in a wireless broadcast

network with no buﬀer using scheduling

algorithms

A multi-source with distinct information sent

over a common channel to multiple recipients

in a wireless broadcast network

A novel scheduling algorithm is designed and

analyzed their age optimality leveraging MDP

and whittle index

[26] 2018 Formulation of uplink AoI in a two-way

data exchanging system

A two-way data exchanging system commu-

nicating through block fading channels

Average uplink AoI and uplink data rate are

derived as functions of downlink data rate in

closed form.

[27] 2018 Minimization of age and synchronization

of information across multiple ﬂows

Distributed cyber-physical systems A novel age-based scheduler is developed

with inter-arrival times in its decision

[28] 2018 Optimizing AoI using the switch or drop

of new arrived packets

A source sending updates to a destination

through a rate-limited link

A Markov chain is formulated and optimized

using MDP with value iteration.

[29] 2018 AoI minimization in a network with

wireless powered source with Rayleigh

fading

A wireless powered energy harvesting user

transmitting through a block Rayleigh fading

channel to the access point

A closed-form expression is derived for aver-

age AoI using queuing theory and probability

models

[8] 2019 Providing a summary of research on AoI,

its variants with various techniques used

to minimize AoI

A destination node and multiple source nodes Summary of approaches is used to minimize

AoI

[30] 2019 Performance evaluation of Multi-access

strategies for AoI minimization

Multiple source nodes communicating with a

common destination.

Various queuing strategies are adopted for

nodes having inﬁnite queue sizes.

[31] 2019 Optimization of scheduling policy to

achieve age optimality

A general system setting with arbitrary arrival

process.

Last generated ﬁrst out queuing policy with

packet replication.

[32] 2019 Minimizing age of information of up-

dates with distortion.

A destination requests information from an in-

formation provider source.

Age-optimal policies for the update request

and update processing delays

[33] 2019 Developing the age-optimal scheduling

policy

An interference-free multi-hop networks A preemptive last generated ﬁrst served

(LGFS) queuing policy to minimize AoI

[34] 2019 Designing the status sampling and updat-

ing process to minimize average AoI

A single IoT device and a real-time monitor-

ing system

A low-complexity semi-distributed sub-

optimal policy based on dynamic program-

ming

[35] 2019 Design of AoI minimization scheduler in

general setting

Multiple source nodes transmitting informa-

tion to a common base station

Tight lower bounds are derived for an opti-

mal solution by generalizing sampling behav-

ior, sample size, and transmission capacity

[36] 2020 Analyzing AoI in a carrier-sense multi-

ple access environment

Multiple source destination pairs sharing

medium for transmission

SHS is used to ﬁnd closed form expression for

average AoI

[37] 2020 Analyzing moments, distributions, and

sampling of an age process vector

A source transmitting status updates to multi-

ple nodes

A SHS is employed for a ﬁnite-state

continuous-time Markov chain

[38] 2020 Addressing AoI in an uncoordinated, and

unreliable multiple access channel

Multiple sensors transmitting information to a

common monitor

SHS approach is used to ﬁnd average AoI as a

function of the oﬀered load and the transmis-

sion success probability

[39] 2020 Deriving average AoI for various queu-

ing disciplines with Poisson arrivals.

A single-server multi-source queuing models Derived exact expressions for average AoI

with M/M/1 and M/G/1 queuing model under

FCFS

[40] 2020 Formulation of sensing time, transmis-

sion time, UAV trajectory, and schedul-

ing to minimize AoI.

A cellular Internet of UAVs, where a UAV

performs data sensing and transmits the data

to the BS.

An iterative algorithm algorithm is used to

minimize the delays

[41] 2020 Optimizing AoI in RF-Powered Commu-

nication Systems

A multiple source nodes are responsible for

sending update packets to a common destina-

tion.

A reinforcement learning framework is used

to optimize AoI

[42] 2021 Modeling trade-oﬀbetween AoI, quality

and energy and optimizing AQI (age and

quality of information)

A general cyber-physical system An online algorithm is proposed to optimize

a cost function, which is a linear combination

of age, quality, and energy

[12] 2021 An introduction and survey on AoI Summary of various network scenarios to an-

alyze AoI

Summary of multiple approaches used to

model AoI

[43] 2021 Maintaining freshness in a congestion-

prone network

Multiple mobile source nodes and a ﬁxed base

station monitor

Proposed WiFresh to improve average AoI

[44] 2021 Analyzing AoI in large-scale wireless

networks

Multiple source nodes and their correspond-

ing receivers

A general model that counts channel gain and

interference, dynamics of arrivals and queuing

[45] 2021 Characterization of spatial distribution

of the mean AoI observed by source-

destination pairs by modeling them as a

bipolar Poisson point process

A large scale network with source-destination

pairs

Using the accurate bounds on the moments of

success probability, tight bounds on the mo-

ments and spatial distribution of AoI is de-

rived

[46] 2021 Analyzing throughput and AoI in a

cellular-based IoT Network

A device-to-device network and a cellular

base station

Stochastic geometry is used to derive AoI and

throughput

[47] 2022 Designing a decentralized transmission

mechanism to minimize AoI

Multiple source nodes communicating with a

common receiver

Two general lower bounds on AoI are derived

for various transmission policies

[48] 2022 Analyzing AoI in an energy harvesting-

aided massive multiple access networks

An energy harvesting wireless power station

source and a data collection base station with

multiple massive machine type communica-

tion devices

Multiple vacation and start-up threshold poli-

cies are used to minimize energy consumption

with conventional multiple access schemes

[49] 2022 Minimizing AoI with query arrival pro-

cess and resource constraints

A pull-based simple time-slotted communica-

tion model

Modeled scheduling problem as MDP and

compared traditional AoI minimization and

query-aware schedulers

[50] 2022 Modeling AoI in energy harvesting en-

abled multi-source network

Multiple sources transmitting information to

an energy harvesting enabled transmitter

which forwards information to the destination

SHS is used to derive closed-form expressions

of the MGF of AoI

[51] 2022 Minimization of AoI in smart healthcare

system

Cellular-based IoT system and small cell base

stations

A closed-form expression of average AoI is

derived considering outage probability and

status generation probability

[52] 2022 Minimizing AoI in an IoT network

where same information is extracted by

multiple source nodes

Multiple information update nodes and a base

station

Greedy algorithm and polynomial-time op-

timal solution is proposed to optimize AoI-

aware update problem and AoI-reduction-

aware problems respectively

[135] 2022 Controlling AoI as well as traﬃc oﬄoad-

ing in 5G IoT Networks

Multiple 5G mobile IoT devices are deployed

on various locations

Price mechanism is used to control age and

cost of service provider by incentive based of-

ﬂoading traﬃc from highly congested loca-

tions

[138] 2022 AoI control with optimal resource allo-

cation and packet sampling

Multiple vehicles communicating through

V2V links

Formulated closed-form expression for AoI

outage probability and proposed resource al-

location algorithm subject to AoI outage prob-

ability.

[139] 2022 Analysis of AoI in clocked network Two agents and a communication network Derived a structural result on the distribution

of AoI process

[140] 2022 Minimization of the long-term average

weighted sum-AoI at the base station in

a multi-node WPSN

A wireless powered sensor network (WPSN)

where multiple sensor nodes send update

packets to the base station

Formulated the average weighted sum-AoI

minimization problem as a multi-stage

stochastic non-linear integer programming

(NLP) and designed an algorithm applying

Lyapunov optimization to decouple the

multi-stage stochastic NLP into per-frame

deterministic NLP problems.

[141] 2022 AoI minimization in a multiple access

setup with multiple antennas receiver

A two-user multiple access channel with

AoI oriented traﬃc with diﬀerent character-

istics. The receiver has multiple antennas

while transmitters have single antenna with

Rayleigh fading channel.

Investigated the average AoI for FIFO, LIFO

and queue with replacement and derived AoI

and average AoI. Formulated an optimization

problem to minimize the average AoI of the

ﬁrst user with a constraint on the average AoI

of the second user. To solve the proposed op-

timization problem, the interior-point method

is used.

[142] 2022 Design and Analysis of AoI in Molecular

communication

A nanonetworks status update system, where

nanodevices cooperate by exchanging data

over molecular communication (MC) chan-

nels.

Introduced a concept of reveal the counter-

acting mechanism when implementing an op-

timally fresh MC channel.

[143] 2022 Scheduling of wireless power trans-

fer to minimize the long-term average

weighted sum of AoI in a wireless pow-

ered NOMA communication network

with multiple source nodes.

A base station and multiple source nodes de-

ployed to extract various real-time informa-

tion.

Lyapunov optimization is exploited to dynam-

ically schedule the actions according to the

channel qualities, the energy status, and the

AoI status of source nodes.

[144] 2022 Study of extreme AoI in wireless-

powered communication systems

A source node and destination node with the

receiver equipped with multiple antennas

Extreme value theory and its corresponding

statistical features are utilized to demonstrate

that the extreme AoI

in Fig. 3, where the blue dotted lines represent the ar-

rivals of packets, while the green dotted lines indicate

the departures of packets.

The average AoI has been used in [53], where the au-

thors assume a single source status update system with

an energy harvesting server with limited battery capac-

ity. The average AoI and asymptotic average AoI ex-

pressions have been derived depending on the server’s

capability to perform simultaneous service and energy

harvesting. Further, a hybrid automatic repeat request

over an error-prone communication channel has been

investigated to minimize the average AoI by adopting an

optimal scheduling policy in [54]. Similarly, [55] delves

into the average AoI using an unmanned aerial vehicle

(UAV) as a mobile relay between a source destination

pair. An optimization problem is formulated to jointly

optimize the trajectory adopted by the UAV with energy

and service time allocation for information transmission

to maximize the information freshness. Also, [56] ana-

lyzes the average AoI in a wireless sensor network with

the capability of wireless power transfer. A dedicated

energy source is deployed, while the sensor node trans-

mits information using the energy emitted by the energy

source node. Moreover, since it is crucial to charac-

terize information freshness in vehicular networks, the

authors in [57] tackle the transmission power minimiza-

tion problem under the constraints of probabilistic AoI

with both deterministic and Markovian traﬃc arrivals.

In [58], the authors employ the average AoI to

maintain information freshness in short packet-based

machine-type communication. On the other hand, in

[59], the system throughput is optimized under the strin-

gent constraints of the average AoI and power in a fad-

ing environment assuming both channel state informa-

tion (CSI) is available and unavailable. Similarly, the

authors in [60] derive the average AoI using stochas-

tic hybrid system (SHS) technique for the packet man-

agement in a M/M/1 queuing model for a multi-source

network. Moreover, [61] analyzes the average AoI in a

multi-source update system with multiple sources sens-

ing updates to a monitor through a last-come ﬁrst-served

server. A non-linear average AoI in a ﬁrst-come ﬁrst-

served discrete-time queue is studied in [62], assum-

Age Δ(n)

Time

g1x3g5

g2g3

x1g4

V1V2V3

Age Process

Packet departure

Packet arrival

Figure 3. AoI evolution for a last-in ﬁrst-out (LIFO) system.

ing a source-destination link in a status update system

with the freshness requirement. The authors in [63] an-

alyze the average AoI in error-prone multi-source sta-

tus update systems with the round-robin and stationary

randomized scheduling policies in conjunction with re-

transmission schemes.

2.2. Peak AoI, ∆p

In some applications, it is crucial to keep the AoI be-

low a threshold with a certain probability. In this case,

the peak AoI, ∆p, is used instead of the average AoI,

∆avg, to emphasize how to treat the worst-case scenario.

In other words, the maximum AoI as it is more com-

plicated to formulate ∆avg as compared to ∆p[16]. Ac-

cording to the application requirement, when the AoI of

the whole system or a packet is required to keep below

some threshold, characterizing ∆pis more logical than

∆avg.

In [45], the authors use the peak AoI to measure the

freshness of information in a large scale wireless net-

work consisting of source-destination pairs based on

stochastic geometry with no storage, preemptive, and

non-preemptive queue disciplines. The results show

that the medium access scheme plays an important role

in characterizing the peak AoI, as compared to the ar-

rival rates. Moreover, [64] derives the minimum achiev-

able peak AoI using service preemption in single server

single source queuing system. The necessary and suﬃ-

cient conditions for preemption are used which are ben-

eﬁcial for a given service-time distribution. The authors

derive a closed-form expression for the probability of

peak AoI exceeding a certain threshold is derived for a

D/G/1 queuing system with an inﬁnite queue size and

ﬁrst come ﬁrst served policy. Following a similar theo-

retical framework, the outage probability of peak AoI is

utilized as a freshness metric in a sensor network [65].

In [66], the peak AoI is computed for a sensor net-

work consisting of various classes of packets in diﬀerent

queuing scenarios, where both ﬁrst-come ﬁrst-served

and last-come ﬁrst-served queuing disciplines are con-

sidered with ﬁnite and inﬁnite buﬀer sizes. The results

show that using inﬁnite buﬀer sizes increase the peak

AoI for both proposed queuing disciplines. [67] con-

siders a set of co-channel links, each having a pack-

ets to be transmitted, the problem of scheduling is ad-

dressed using an integer linear programming such that

the peak AoI is minimized in a wireless system. In [68]

and [68], the authors propose a computing-enabled IoT

system with multiple sensors and analyze the peak AoI.

Through the proposed system architecture, the distance

between the monitoring system and sensors is reduced,

which results in the improved communication delays

and the peak AoI. Furthermore, the peak AoI is min-

imized using through UAV-assisted relay networks by

jointly optimizing the ﬂight trajectory of the UAV, the

service time of packets and power allocation [69].

2.3. Value of Information

Value of information (VoI) is measure of the uncer-

tainty reduction by the reception of an update, indicat-

ing the degree of the importance of the received up-

date at the monitor [72]. When a packet is generated at

the source, its VoI starts decreasing, as its VoI is high-

est when it is the latest update and contains important

information about the monitoring system. Optimizing

AoI does not always improve the system performance in

many applications, because the quality of the measure-

ments in all data packets may be diﬀerent depending on

the amount of noise in the observations [70]. The infor-

mation dissemination to the monitor is governed by the

scheduler in the future vehicular networks using the VoI,

and the packet drops can be managed according to VoI

to satisfy the quality of service requirements [22]. In

[71], the authors compare AoI and VoI, where the stale

information sometimes has more VoI, as compared to

the fresh update, in case that the sources update infor-

mation with diﬀerent update rates. The VoI of diﬀer-

ent information and updates from diﬀerent sources may

have diverse VoI at the monitor independent of their

staleness; thus, [72] introduces the cost of update de-

lay at the monitor, which decreases, when an update is

received at the monitor.

2.4. Factors Aﬀecting AoI

AoI of a packet starts increasing linearly after its gen-

eration at the source until its delivery to the monitoring

system. From the generation to the reception, it has to

travel through various components in the communica-

tion system with delays. The initial parameter, which

aﬀects the AoI of a packet, is its generation rate at the

source. After the arrival at the source node, it has to

store the packets in a queue till the departure, the wait-

ing time of the packet taken in the queuing process is

also a function of AoI. The packets wait in the queue

until the scheduler processes the packet for transmission

therefore the scheduling policy and the access scheme

adopted by the system are major factors to determine

AoI. The transmission delay, the propagation delay af-

ter the departure from source node, channel conditions,

and outage probability also aﬀect the AoI of the system.

End-to-end delay is mostly used in former studies to

measure the latency from the source to the destination.

However, this metric cannot ensure the timely status up-

dates. The end-to-end delay can be reduced by decreas-

ing the update rate at the source node, but it increases

information staleness, which corresponds to higher AoI

at the monitor. Throughput, on the other hand, is the

measure of system utilization, but it can be enhanced by

increasing the update rate at the source, but owing to the

increased congestion and packet drops in the system, the

information freshness is compromised.

3. AoI with Queuing Theory

The queuing policy adopted by the system plays a

key role in the characterization of AoI, as it determines

which packet experiences how much delays [132–134].

The waiting time of a packet in the queue depends on

the queuing policy adopted by the system and it also

contributes in the staleness of information carried by the

packets. For this reason, in this section, we consider the

impacts of various queuing policies in the formulation

of AoI. Last-In First-Out (LIFO) and First-In First-Out

(FIFO) queuing are generally used in queuing theory

and the choice depends on the application requirement

in the context of AoI. LIFO is often used when every

packet carries identical information and the monitor is

interested in the freshest information extracted by the

source, therefore the eﬃcient way is to transmit the lat-

est packet stored in the queue to maintain information

Age Δ(n)

Time

g1x2

x1U2

Age Process

Packet departure

Packet arrival

g4x4

Figure 4. The AoI evolution with FIFO queuing.

fresh at the monitor. While in some applications with

multi-sensory data [137], every packet carries distinct

information and information carried by every packet is

essential to be delivered at the monitor in addition to in-

formation freshness consideration. FIFO performs bet-

ter in such systems which strives to transmit information

carried by maximum packets while preserving informa-

tion freshness. Therefore, LIFO and FIFO are adopted

intelligently according to the applications demand. The

more details of both approaches are given in the next

subsections.

3.1. Last-In First-Out (LIFO)

LIFO is most frequently adopted in the existing stud-

ies on AoI, because it transmits the latest received in-

formation ﬁrst to keep the information as fresh as possi-

ble at the monitor. This policy is useful in maintaining

the age of the system at a low level, when a source is

interested in the freshest information and every packet

carries identical information about the system. An ex-

ample of the LIFO queuing is shown in Fig. 3where

the newly arrived packet preempts the older packets in

the queue. An older packet is replaced by the fresher

update in the queue, which minimizes the AoI of the

system. In Fig.3,g1,g2, ...g5depict the generation time

of respective packet while x1,x2, ...x5are the departure

time of respective packet. Similarly, V1,V2, ...V4are

the inter-arrival time of the packets and U1,U2are the

time between the arrival and departure of the respective

packet. It can also be seen that the second and third

packets are preempt from the system on the arrival of

fourth packet which is fresher to preserve information

freshness. AoI has been studied using the LIFO queu-

ing in [27,73,74,74,76]. In addition, the authors in

[14] assume an LIFO queue and ﬁnd the average AoI of

the system with and without preemption. Similarly, [75]

computes the average and peak AoI for the LIFO queu-

ing system both in the presence and absence of the pre-

emption. The results are also extended to the case of de-

terministic service time. In addition, AoI is derived us-

ing packet management in [16], where the source node

has capability to manage the arriving samples and de-

cide which packet will be transmitted to the destination.

3.2. First-In First-Out (FIFO)

In the FIFO queuing system, the oldest packet in the

queue is transmitted ﬁrst, which is adopted by many

works as [27,67,76,77]. This policy is eﬀective when

the information carried by every packet is important

and the system strive to preserve information carried

by maximum packets. An example of the FIFO queu-

ing system is shown in Fig. 4with the queue length of

1, in which the newly arrived packet is dropped, when

there is already a packet in the queue. This dropping of

the newer packet and keeping stale information result

in the increased AoI, however, when the queue length

is greater than one and all packets carry distinct infor-

mation, FIFO performs better, because it is important to

preserve information carried by all of the packets.

Multiple packets carry diﬀerent and equally impor-

tant information about the monitoring system in some

applications [78]. However, the dropping policy is also

important, when the queue becomes overloaded; there-

fore, [79] suggests dropping of the oldest packet to min-

imize the average age of a node and, consequently, the

whole system. This policy outperforms, if the appli-

cation demands the maximum age lower than a given

threshold value. In other words, the peak AoI is kept to

be low, as the oldest packet is transmitted ﬁrst, whereas

the oldest packet is dropped in case that the queue is

full.

FIFO is introduced for AoI optimization for the ﬁrst

time in vehicular networks in [125]. The queuing pol-

icy is optimized to decrease AoI by reducing the con-

tention in the network. An optimal policy is adopted for

the heterogeneous traﬃc status update generation using

the FIFO queuing discipline [80]. In a similar way, the

FIFO queuing is adopted in a real-time streaming source

coding system, where the arrivals are random. A loss-

less coding scheme is designed to keep information of

all the status updates instead of minimizing the peak

AoI of only the latest received packet [81]. To han-

dle heterogeneous traﬃc, in case that every packet is

important and information in each packet needs to be

preserved, the FIFO queuing is a pertinent option, as in

[79]. The oldest packet dropping policy in the queue is

used to minimize AoI.

4. Modeling of AoI

The queuing policies at the source nodes and at the

information aggregation points (e.g., base stations) are

analytically modeled using various approaches. As the

model used to store the information carried by the pack-

ets in queues and their transmission to the aggrega-

tion point contribute in the delays of packets and AoI,

the model adopted must be optimized according to the

application demand to preserve information freshness

[112]. In such modeling, the main objective is to regu-

late the storage of information waiting for transmission

in queues and transmit the information to the respec-

tive monitor with the minimum delay using lower com-

plexity. AoI has been modelled using Markov chain in

various research specially in characterizing the queuing

delay and transmission scheduling [112] [136]. To op-

timize the scheduling and queuing policy, Markov de-

cision process (MDP) is used which schedules the right

node at right time to minimize AoI of the system or ac-

cording to the application demand.

4.1. Markov Chain

AoI has been modeled using Markov chain in various

literature, where a sequence of events are characterized

in a stochastic manner, where the probability of an event

is subject only to the state in the previous event. The

scheduling of packets for transmission and their stor-

age in a queue with a suitable packet dropping is essen-

tial for the AoI minimization. According to the Markov

property, the next state only depends on the current state

of queue, therefore, Markov chain can be utilized to reg-

ulate the packets’ storage in the queues [136].

Modeling the possible states of a queue as Markov

chain states, the steady-state matrix can be found, once

the state transition from one state to another is identi-

ﬁed to construct the Markov matrix. Using this steady-

state matrix, AoI and packet drops for a given schedul-

ing policy are analyzed in [79]. In this work, with the

Markov chain-based model, the optimal scheduling and

packet dropping policy are proposed to minimize AoI

and packet drops.

4.2. Markov Decision Process (MDP)

Taking advantage of the Markov property, Markov

decision process (MDP) can be exploited to schedule

the transmissions, which leads to minimizing AoI of the

system. MDP consists of a state space, action, transi-

tion probability and cost [112] [79]. An optimal action

is taken at every time step after assessment of the costs

or reward found from the Markov chain modeling to re-

duce the cost or increase the reward at the next time step.

The cost and action of MDP depends on the cost/reward

function, which is derived based on the requirement of

network model [136].

Some prior studies employ MDP to schedule the

users for the purpose of minimizing AoI of the entire

system or to prioritize a user over others. To minimize

the peak AoI, the cost function is derived to schedule

the user with the maximum AoI at the moment. Simi-

larly, to schedule a user according to any other metrics

(i.e., arrival probability or VoI), the cost metric can be

derived correspondingly.

4.3. Stochastic Hybrid System (SHS)

Recently, stochastic hybrid system (SHS) has gained

considerable interest, as noted in [82]. SHS is used

to model both continuous and discrete behaviors of the

AoI systems. The authors in [83] study the AoI mini-

mization in a large network, where transmitters harvest

energy from power beacons and then transmit informa-

tion to the receivers. The average AoI is derived using

stochastic geometry, and then stochastic optimization is

applied to derive the expression for the minimum av-

erage AoI. In [84], the authors consider a single-hop

network, where the source nodes transmit to the desti-

nation using multiple unreliable channels. The packets

are generated at the sources using a stochastic system

with known statistics, while the channel conditions vary

according to stochastic process with unknown statis-

tics. A learning algorithm is applied to understand the

channel conditions and minimize the expected total AoI.

Along the same line, the authors in [85] model AoI as a

stochastic optimization problem and analyzed the con-

trol performance of the stochastic systems with stale in-

formation. In [86], the authors consider a multiuser net-

work, where each node with stochastic arrivals strives

to transmit their information to the monitor as timely as

possible. A partially observable Markov decision pro-

cess and dynamic programming can be used to schedule

the nodes to minimize AoI of the whole network. [87]

proposed an analytical model to formulate the optimal

scheduling policy by encapsulating the queuing theory

with stochastic geometry. The authors in [88] propose

an eﬃcient scheduling algorithm to maximize infor-

mation freshness under both adversarial and stochastic

channel in a mobility-enabled cellular network. The au-

thors in [89] consider a source-destination pair, and both

stochastic and arbitrary arrivals are modeled to mini-

mize AoI. In [90], the authors analyze AoI for multiple

source-destination pairs in a large-scale wireless net-

work stochastic setting with the ﬁrst-come ﬁrst-served

queuing discipline without buﬀer. [91] considers coop-

erative communication between sources and the desig-

nated destinations. The sources randomly generate sta-

tus updates and aim to send the information to the desti-

nation as timely as possible. [92] is focused on multiple

low-power nodes that transmit timely information about

a random process to a common sink. AoI is minimized

using an optimization problem, which is solved using

the Lyapunov drift-plus-penalty method. Similarly, [93]

considers a downlink wireless system, where a base sta-

tion sends multiple stream of information to diﬀerent

destinations. The arrivals at the base station are mod-

eled as stochastic process.

5. Impact of Scheduling

AoI has been studied using diﬀerent scheduling poli-

cies, as in [86,94–96,96–111]. The scheduling is

adopted based on to the application-speciﬁc require-

ments and priorities given to the nodes determined by

system parameters. In this section, we treat four diﬀer-

ent scheduling approaches separately in the following

sub-sections in detail. The four diﬀerent scheduling ap-

proaches are compared with their key advantages and

disadvantages in Table 2.

5.1. Random Scheduling

In random scheduling, all nodes deployed in the net-

work are scheduled to transmit unbiasedly. In this

scheduling, a node is chosen randomly with the equal

probability among all of the candidate nodes. Because

it does not prioritize any node, the average or peak AoI

of the whole system cannot be optimized. In Fig. 5,

three nodes n1,n2, and n3are waiting to transmit infor-

mation to a common monitor using time division mul-

tiple access (TDMA), where the transmission policy is

governed by the scheduler. Further, λidenotes the ar-

rival rate of the ith node, while ∆ishows the AoI of the

ith node at the monitor. In Fig. 5, it schedules any node

randomly for transmission. As a random realization, in

the ﬁgure, the third node is selected for scheduling at

time nunbiasedly. This random scheduling does not

take into account any parameter or status of the system

for scheduling, which results in unoptimized AoI. When

the arrival rate on a node is low, the queue of the node

stays almost empty or very few packets exist most of the

time. In this case, therefore, giving the same priority to

the node with slower arrival rate and node with faster

update rate is not eﬃcient. In unbiased and random

scheduling, the scheduler may schedule a node without

having any packet in its queue, which results in un-used

time slots and low eﬃciency.

λ1

Scheduler

λ2

λ3

Δ1>Δ2>Δ3

λ2>λ3>λ1

Monitor

Figure 5. Random scheduling.

λ1

Scheduler

λ2

λ3

Δ1>Δ2>Δ3

λ2>λ3>λ1

Monitor

Figure 6. Arrival-based scheduling

5.2. Arrival-based Scheduling.

The arrival rate on a node controls the cumulative

AoI, as it deﬁnes the number of packets in its queue.

The node with the faster arrival rate always has the latest

information-carrying packet in its queue. In such node,

the status of queue is overloaded most of time due to the

fast update rate. In the same way, when the queue size

of a node is only 1, the maximum arrival rate stores the

latest information about the monitoring system, there-

fore, to keep average age lower in the former case and

to transmit latest information from the node, which has

highest probability of having packets in its queue in the

latter case, the scheduling according the arrival proba-

bility is beneﬁcial.

In Fig. 6, we can observe that node 2 has the fastest

arrival rate λ2and it has the maximum number of pack-

ets in the queue, where the probability of the empty

queue and going channel assigned to node 2 idle is very

low. Therefore, node 2 must be given priority based on

its arrival rate, and for the same reason node 3 must be

prioritized as compared to node 1, because λ3> λ1.

5.3. Age-based Scheduling

Apparently, scheduling of a node according to the in-

stantaneous AoI is the most eﬀective in minimizing AoI

of the system. In this type of scheduling, all nodes are

assessed at every time slot and the node with maximum

AoI is scheduled for transmission at every time instant,

as shown in Fig. 7, which decreases AoI of the entire

system. Keeping stale information in queues makes the

AoI of the node increase at every time instant. There-

fore, scheduling the node with the maximum AoI may

Table 2

Advantages and disadvantages of various queuing methods and impacts on AoI.

Advantages Disadvantages

Random Scheduling

Low Complexity

Required no information about the monitor

Unbiased

Not age optimal

No priorities

Arrival-based Scheduling Latest update is transmitted ﬁrst High peak AoI

Information about arrival rate is required at the scheduler

Age-based Scheduling Always minimizes average AoI

Minimizes peak AoI

Information about the monitor is required

Prioritizes old over latest updates

Weight-based Scheduling

Minimizes peak AoI

Balanced scheduling

Application-depedent priority

High complexity

Information about monitor is required

λ1

Scheduler

λ2

λ3

Δ1>Δ2>Δ3

λ2>λ3>λ1

Monitor

Figure 7. Age-based scheduling.

decrement the peak AoI. Fig. 7shows that the AoI of

node 1, indicated by n1, is the highest, by which the

scheduler has scheduled n1at the time instant nto keep

the peak AoI of the node low.

The arrival rate on nodes, scheduling policy, channel

conditions between nodes and the monitor may aﬀect

the AoI of the nodes in a composite fashion. Thus, in-

stead of focusing on an individual of them, to keep the

peak AoI below a certain threshold and control the aver-

age AoI, it is lucrative to schedule the nodes in order of

their instantaneous AoI. In this type of scheduling, all

nodes are assessed at every time slot and the node with

the maximum AoI is scheduled for transmission, which

decreases the AoI of the whole system.

Letting ∆ibe the instantaneous AoI of the ith node,

Fig. 7shows the instantaneous AoI of each node at

the monitor at current time instant, it can be seen that

the AoI of node 1 is largest at the moment therefore

scheduling node 1 at current time instant decreases AoI

of the system the most. Therefore, node 1 is scheduled

to minimize AoI in the next time instant as shown in

Fig. 7.

5.4. Weight-based Scheduling

A node can be prioritized according to the applica-

tion requirement or to keep a balance between the AoI

and arrival probability, a weight function is introduced

to schedule the nodes according to the value of weight

function. This kind of scheduling policy is more opti-

mized as compared to the age and arrival probability-

based scheduling, because it takes care of the applica-

tion requirements, the priority of a node and the instan-

taneous AoI as well. The weight-based scheduling is

widely employed for scheduling of nodes in prior arts

using Markov decision processes (MDP). Through the

MDP, we can capture the current state of the network

and take the optimal decision that maximizes the reward

or minimizes the cost of taking a particular decision at

current time [112]. For the arrival rate λion the ith node,

the instantaneous AoI is ∆i, while αis a weight function

used to prioritize the arrival rate and the instantaneous

AoI. Then, the value of the reward function observed

before taking scheduling decision is expressed as

ri(n)=α∆i(n)+(1 −α)λi.(2)

The value of reward function is computed using (2) for

all nodes at time n. Accordingly, the node with the max-

imum reward is scheduled at time n.

6. Impact of Access Mechanism

The access mechanism adopted in a network is one

of the driving force behind the AoI optimization. The

access schemes determine the spectral eﬃciency and the

outage probability, both of which aﬀect the AoI of the

network. The impact of using various access schemes

are explained in the subsections.

6.1. Time Division Multiple Access (TDMA)

In TDMA-based systems, the entire available band-

width is assigned to a user in a network for a time

slot. The user utilizes all available bandwidth in its

assigned time slot to send the information. The large

available bandwidth results in the lower transmission

delay, which, as a result, minimizes AoI, as compared

to the mechanism with higher transmission delay. Each

user receives a new arrival and stores such newly ar-

rived packet in a queue, until it is scheduled for trans-

mission. When the arrival rate on a user is fast, the user

should wait until his/her turn, which increases the stale-

ness of information due to the queuing delay. Therefore,

TDMA may not always be eﬀective to minimize AoI.

TDMA may outperform FDMA, when the size of

packets to transmit are huge and the arrivals on the

users are not frequent enough. A comparison between

TDMA and FDMA for the AoI minimization is stud-

ied in [113], where the results indicate that TDMA out-

performs FDMA in terms of the average AoI. On con-

trary, FDMA outperforms in minimizing the peak AoI

because of its robustness against channel variations. A

similar study in [114] also shows that FDMA is more

suitable, compared to TDMA in fading channels.

6.2. Frequency Division Multiple Access (FDMA)

While the entire bandwidth is occupied by a single

user in a time slot in TDMA, FDMA divides the avail-

able bandwidth to all users and assign a small part of

the total bandwidth to a user. As a result, an individual

user in FDMA may experience high transmission delay,

because of its lower bandwidth assigned to each user;

therefore, the average AoI becomes higher, as com-

pared to TDMA. However, when the size of transmit-

ted packets are smaller and the bandwidth requirement

is not large, FDMA can outperform TDMA. Further-

more, when the arrival rates on users are higher, FDMA

can also outperform TDMA, because every user may

have a chance of transmission in every time slot but with

smaller bandwidth.

6.3. Non-orthogonal Multiple Access (NOMA)

For the eﬃcient use of resources, non-orthogonal

multiple access (NOMA) allows multiple transmission

in a time slot, while only one node is allowed to transmit

in a time slot in TDMA. Transmitting multiple nodes

in a single time reduces the transmission delay and

eventually AoI. However, outage probability of NOMA

is greater, as compared to the conventional orthogonal

multiple access (OMA) such as TDMA and FDMA.

Hence, NOMA does not always outperform OMA.

Fig. 8shows an example illustration about the dis-

tinction between NOMA and OMA in the uplink trans-

mission network. In the ﬁgure, the deployed nodes are

clustered in pairs based on their channel conditions and

transmit to a common base station (BS). In cluster 3,

which employs the conventional OMA, only a single

node can transmit to the BS. In clusters 1, 2, and 3, on

the other hand, both nodes in each cluster transmit their

updates using the same frequency channel, simultane-

ously. In this example, we assume the power-domain

NOMA, where superposition coding is utilized. As

shown in the ﬁgure, each cluster is formed with a pair

of nodes with disparate channel conditions, where the

node with worse channel, labeled with “weak” node, is

assigned lower power, while the node with better chan-

nel, labeled with “strong” node, transmit with higher

power. When the BS receives the super-imposed sig-

nals from the two nodes, indicated by the green and

black dashed lines, it can ﬁrst decode the signal from

the strong node, then performs successive interference

cancellation (SIC) to decode the weak node’s signal.

Through NOMA, the AoI of the system can be reduced

relative to OMA, because multiple nodes transmit their

information in a single time slot and the nodes are likely

to wait less compared to OMA. However, when both

nodes in a cluster do not have packets in their queue for

transmission, it is not eﬃcient to use NOMA, a more

eﬃcient way in this case is explained in the next sub-

section.

6.4. Hybrid Scheme of OMA and NOMA

In [112], a multi-user downlink network is consid-

ered, where a BS sends status updates to the users using

the adaptive combination of NOMA and OMA. In this

scheme, either NOMA and OMA can be selectively em-

ployed. To be speciﬁc, An MDP age-based scheduling

is adopted to switch between NOMA and OMA accord-

ing to the instantaneous AoI and the downlink signal-

to-noise ratio (SNR). Further, [115] also proposes a hy-

brid NOMA and OMA-based scheduling at the BS in

updating the users in a downlink short-packet commu-

nication system. The adaptive scheduling is based on

Lyapunov optimization, whereas the power is allocated

adaptively to maintain low expected sum of AoI. The

authors in [116] study the impact of stochastic arrivals

on the peak AoI using NOMA and OMA. The results

show that NOMA outperforms OMA in case of more

frequent updates.

[117] proposes NOMA scheduling to minimize the

AoI in a vehicular communication system. They

also adaptively adjust power allocation to the nodes

to reduce the overall AoI. In [118], the age-optimal

power allocation scheme is proposed in a NOMA-based

satellite-integrated Internet-of-Things (IoT) network us-

ing Lyapunov optimization. The age performance in

a two-user downlink communication system served by

a BS using NOMA is studied in [119]. The transmis-

sion adopted is a hybrid automatic repeat request with

Arrival

Departure

Weak node

Strong node

Weak Signal

Strong Signal

ℓ1

ℓ2

Cluster 2

Cluster 1

Cluster K

Cluster 3

OMA

ℓ1

ℓ2

ℓ1

No Packets

ℓ1

ℓ2

Cluster 2

Figure 8. An IoT uplink network where groups of nodes transmit data to a central BS using NOMA principle.

chase combining (HARQ-CC) in a ﬁnite block length.

NOMA is exploited to minimize the average age of the

two users with independent information in a downlink

wireless communication [120]. The authors in [121]

propose an age-optimal protocol using NOMA in an en-

ergy harvesting IoT network, where the results show

that NOMA outperforms TDMA in terms of AoI. The

stochastic hybrid scheme (SHS) is adopted to model

AoI using NOMA, while the traditional OMA scheme

in a machine-type communication (MTC) is also con-

sidered in [122]. The results show that although NOMA

is spectral-eﬃcient, but it does not always outperform

OMA in minimizing the average AoI.

7. Applications of AoI

AoI has been used in several applications, where

status updates are used to transfer information from a

source to a monitor. Some of the applications, where

AoI is critical and noteworthy, are discussed in this sec-

tion.

7.1. Smart Agriculture

Agriculture is one of the world’s largest industries,

which employs more than one billion people. To cope

with the fact that unprecedented volume of people mov-

ing from rural to urban areas in many countries, where

agriculture is an important sector in the economy, the

concept of smart agriculture has emerged, which refers

to managing farms with the AmI technologies such as

IoT, artiﬁcial intelligence, robotics, and drones to im-

prove both quality and quantity of products.[123]. As-

suming the remote monitoring of crops, information

must to be delivered to a monitoring system from the

sensors deployed in the farms. The freshness of such

information is essential for the proper functioning of the

monitoring system. The information extracted includes

humidity, temperature, moisture, and nutrient level of

the crops. The information delivered to the monitoring

system is used to monitor the health condition of the

crops remotely. The AoI minimization in smart agricul-

ture has been studied in [78], where a vehicle is used to

collect the information from the sensors deployed in an

agricultural area.

7.2. Intelligent Transportation System (ITS)

Intelligent transportation system (ITS) is going to

be a key part of the next-generation vehicular net-

work. ITS includes vehicle-to-vehicle communication

(V2V), vehicle-to-infrastructure (V2I) communication

and vehicle-to-everything (V2X) communication [124–

127]. The V2V communication is critical, when the ve-

hicles exchange information to each other for smooth

moving on the road. Moreover, the AoI in V2V com-

munication is also important, as such information is uti-

lized by the vehicle to vary the speed and path. V2I

communication is used to monitor the vehicles by a

remote monitor through road side units (RSUs). The

vehicles update their status to the RSUs such as, the

speed of the vehicle, fuel level, trajectory path, loca-

tion, etc. This information is utilized to monitor the ve-

hicles and deliver them critical messages such as, speed

warning, online ﬁnes, fuel level warning, road condi-

tion ahead. [125] thrives to minimize the AoI in vehicu-

lar network by making variations in contention window

sizes of carrier-sense multiple access (CSMA). Schedul-

ing is optimized to minimize AoI by piggybacking in-

formation in an ad-hoc vehicular network [126]. An

AoI-based evaluation framework is proposed in [127],

in which a greedy online scheduling algorithm is used

to minimize the expected sum AoI.

7.3. Ultra-reliable and Low-latency Communication

(URLLC)

Ultra-reliable and low-latency communication

(URLLC) enables various AmI applications, which has

stringent requirements of the reliability and latency.

AoI is analyzed supporting URLLC in a multimedia

wireless network using stochastic hybrid system [128].

The average AoI is estimated in a URLLC-assisted

wireless communication [129]. Similarly, AoI is mini-

mized by controlling the tail of the AoI distribution in a

URLLC-enabled vehicular communication in [57]. AoI

has been studied using a decode-and-forward relaying

scheme under a ﬁnite block length in a relay-assisted

wireless network in [11]. The scheduling algorithm is

optimized using reinforcement learning to minimize the

AoI in a URLLC network [130].

7.4. Augmented Reality (AR) and Virtual Reality (VR)

The high quality data transformation in virtual games,

conferences and movies, remote working, enabled by

augmented reality (AR) and virtual reality (VR), has

been increasing exponentially. In such applications, AoI

should be carefully managed to improve the quality of

experience (QoE) perceived by the end users, because

a fast and large data transfer in AR and VR is highly

susceptible to various delays. Guaranteeing information

freshness and reliability in these systems is challeng-

ing to enable real-time and high-quality physical expe-

riences [131].

8. Conclusions

AoI is the key performance indicator in the data ag-

gregation and analytics for the future IoT with AmI,

which quantiﬁes the timeliness and freshness of in-

formation at the monitoring system. A wide range

of the AmI applications enabled by the advancement

in communication systems are age-sensitive, such as

smart agriculture, ITS, URLLC, AR/VR, smart health-

care systems, digital twin, and industrial automation.

These applications require the extracted information

from sources to be delivered to the monitor as soon as

possible to keep the received information fresh. To bet-

ter design and operate the AmI systems in such applica-

tions, we have provided a comprehensive review of the

diﬀerent notations of AoI and related metrics, which can

be selectively employed depending on the application

requirements and properties. Further, AoI in the AmI

applications can be easily degraded by network delays

due to transmission, queuing, propagation, and multiple

access schemes. For this reason, we have reviewed vari-

ous queuing policies to keep the information up-to-date

and have explained their impacts on the peak and aver-

age AoI. Moreover, some analytical models to optimize

AoI have discussed in detail, including Markov chain,

MDP, and SHS. In addition, as one of the key design

factors, the scheduling strategies of multiple nodes have

been presented with their pros and cons in diﬀerent sce-

narios to minimize AoI. The multiple access schemes

such as TDMA, FDMA, NOMA, and the hybrid version

of NOMA and OMA have been also compared for the

optimization of information freshness. Finally, we have

identiﬁed promising research opportunities in potential

age-sensitive applications.

References

[1] K. Lashmi, A. S. Pillai, Ambient intelligence and IoT based

decision support system for intruder detection, in: IEEE Inter-

national Conference on Electrical, Computer and Communica-

tion Technologies (ICECCT), 2019, pp. 1–4.

[2] F. Sadri, Ambient intelligence: A survey, ACM Comput. Surv.

43 (4) (Oct. 2011).

[3] S.-H. Lee, D.-W. Lee, Actual cases for smart fusion industry

based on internet of thing, Journal of the Korea Convergence

Society 7 (2) (2016) 1–6.

[4] T. M. Bojan, U. R. Kumar, V. M. Bojan, An Internet of Things

based intelligent transportation system, in: IEEE International

Conference on Vehicular Electronics and Safety, 2014, pp.

174–179.

[5] G. Elhayatmy, N. Dey, A. S. Ashour, Internet of Things based

wireless body area network in healthcare, in: Internet of Things

and big data analytics toward next-generation intelligence,

Springer, 2018, pp. 3–20.

[6] L. Li, Application of the Internet of Thing in green agricultural

products supply chain management, in: IEEE Fourth Interna-

tional Conference on Intelligent Computation Technology and

Automation, Vol. 1, 2011, pp. 1022–1025.

[7] S. Dey, A. Roy, S. Das, Home automation using Internet of

Thing, in: IEEE 7th Annual Ubiquitous Computing, Electron-

ics & Mobile Communication Conference (UEMCON), 2016,

pp. 1–6.

[8] M. A. Abd-Elmagid, N. Pappas, H. S. Dhillon, On the role of

age of information in the Internet of Things, IEEE Communi-

cations Magazine 57 (12) (2019) 72–77.

[9] S. Kaul, R. Yates, M. Gruteser, Real-time status: How often

should one update?, in: IEEE Proceedings INFOCOM, 2012,

pp. 2731–2735.

[10] A. Kosta, N. Pappas, V. Angelakis, Age of information: A new

concept, metric, and tool, Foundations and Trends®in Net-

working 12 (3) (2017) 162–259.

[11] C. M. Wijerathna Basnayaka, D. N. K. Jayakody, T. D. Pon-

nimbaduge Perera, M. Vidal Ribeiro, Age of information in

an URLLC-enabled decode-and-forward wireless communica-

tion system, in: IEEE 93rd Vehicular Technology Conference

(VTC2021-Spring), 2021, pp. 1–6.

[12] R. D. Yates, Y. Sun, D. R. Brown, S. K. Kaul, E. Modiano,

S. Ulukus, Age of information: An introduction and survey,

IEEE Journal on Selected Areas in Communications 39 (5)

(2021) 1183–1210.

[13] F. Alawad, F. A. Kraemer, Value of information in wireless sen-

sor network applications and the IoT: A review, IEEE Sensors

Journal 22 (10) (2022) 9228–9245.

[14] S. K. Kaul, R. D. Yates, M. Gruteser, Status updates through

queues, in: 46th Annual Conference on Information Sciences

and Systems (CISS), 2012, pp. 1–6.

[15] C. Kam, S. Kompella, A. Ephremides, Age of information un-

der random updates, in: IEEE International Symposium on In-

formation Theory, 2013, pp. 66–70.

[16] M. Costa, M. Codreanu, A. Ephremides, Age of information

with packet management, in: IEEE International Symposium

on Information Theory, 2014, pp. 1583–1587.

[17] N. Pappas, J. Gunnarsson, L. Kratz, M. Kountouris, V. An-

gelakis, Age of information of multiple sources with queue

management, in: IEEE international conference on commu-

nications (ICC), 2015, pp. 5935–5940.

[18] L. Huang, E. Modiano, Optimizing age-of-information in a

multi-class queueing system, in: IEEE International Sympo-

sium on Information Theory (ISIT), 2015, pp. 1681–1685.

[19] C. Kam, S. Kompella, G. D. Nguyen, A. Ephremides, Eﬀect of

message transmission path diversity on status age, IEEE Trans-

actions on Information Theory 62 (3) (2015) 1360–1374.

[20] K. Chen, L. Huang, Age-of-information in the presence of er-

ror, in: IEEE International Symposium on Information Theory

(ISIT), 2016, pp. 2579–2583.

[21] C. Kam, S. Kompella, G. D. Nguyen, J. E. Wieselthier,

A. Ephremides, Controlling the age of information: Buﬀer

size, deadline, and packet replacement, in: IEEE Military

Communications Conference (MILCOM), 2016, pp. 301–306.

[22] M. Giordani, A. Zanella, T. Higuchi, O. Altintas, M. Zorzi,

Investigating value of information in future vehicular commu-

nications, in: IEEE 2nd Connected and Automated Vehicles

Symposium (CAVS), 2019, pp. 1–5.

[23] Y. Sang, B. Li, B. Ji, The power of waiting for more than

one response in minimizing the age-of-information, in: IEEE

GLOBECOM, 2017, pp. 1–6.

[24] R. D. Yates, S. K. Kaul, Status updates over unreliable multi-

access channels, in: IEEE International Symposium on Infor-

mation Theory (ISIT), 2017, pp. 331–335.

[25] Y.-P. Hsu, E. Modiano, L. Duan, Scheduling algorithms for

minimizing age of information in wireless broadcast networks

with random arrivals, IEEE Transactions on Mobile Comput-

ing 19 (12) (2019) 2903–2915.

[26] S. Bhambay, S. Poojary, P. Parag, Fixed length diﬀerential

encoding for real-time status updates, IEEE Transactions on

Communications 67 (3) (2018) 2381–2392.

[27] C. Joo, A. Eryilmaz, Wireless scheduling for information

freshness and synchrony: Drift-based design and heavy-traﬃc

analysis, IEEE/ACM transactions on networking 26 (6) (2018)

2556–2568.

[28] B. Wang, S. Feng, J. Yang, When to preempt? Age of infor-

mation minimization under link capacity constraint, Journal of

Communications and Networks 21 (3) (2019) 220–232.

[29] H. Hu, K. Xiong, Y. Zhang, P. Fan, T. Liu, S. Kang, Age of

information in wireless powered networks in low SNR region

for future 5G, Entropy 20 (12) (2018) 948.

[30] A. Kosta, N. Pappas, A. Ephremides, V. Angelakis, Age of

information performance of multiaccess strategies with packet

management, Journal of Communications and Networks 21 (3)

(2019) 244–255.

[31] A. M. Bedewy, Y. Sun, N. B. Shroﬀ, Minimizing the age of in-

formation through queues, IEEE Transactions on Information

Theory 65 (8) (2019) 5215–5232.

[32] M. Bastopcu, S. Ulukus, Age of information for updates with

distortion, in: IEEE Information Theory Workshop (ITW),

2019, pp. 1–5.

[33] A. M. Bedewy, Y. Sun, N. B. Shroﬀ, The age of information in

multihop networks, IEEE/ACM Transactions on Networking

27 (3) (2019) 1248–1257.

[34] B. Zhou, W. Saad, Joint status sampling and updating for min-

imizing age of information in the Internet of Things, IEEE

Transactions on Communications 67 (11) (2019) 7468–7482.

[35] C. Li, S. Li, Y. T. Hou, A general model for minimizing age

of information at network edge, in: IEEE Conference on Com-

puter Communications (INFOCOM), 2019, pp. 118–126.

[36] A. Maatouk, M. Assaad, A. Ephremides, On the age of infor-

mation in a CSMA environment, IEEE/ACM Transactions on

Networking 28 (2) (2020) 818–831.

[37] R. D. Yates, The age of information in networks: Moments,

distributions, and sampling, IEEE Transactions on Information

Theory 66 (9) (2020) 5712–5728.

[38] R. D. Yates, S. K. Kaul, Age of information in uncoordinated

unslotted updating, in: IEEE International Symposium on In-

formation Theory (ISIT), 2020, pp. 1759–1764.

[39] M. Moltafet, M. Leinonen, M. Codreanu, On the age of infor-

mation in multi-source queueing models, IEEE Transactions

on Communications 68 (8) (2020) 5003–5017.

[40] S. Zhang, H. Zhang, Z. Han, H. V. Poor, L. Song, Age of infor-

mation in a cellular internet of UAVs: Sensing and communi-

cation trade-oﬀdesign, IEEE Transactions on Wireless Com-

munications 19 (10) (2020) 6578–6592.

[41] M. A. Abd-Elmagid, H. S. Dhillon, N. Pappas, A reinforce-

ment learning framework for optimizing age of information

in RF-powered communication systems, IEEE Transactions on

Communications 68 (8) (2020) 4747–4760.

[42] N. Rajaraman, R. Vaze, G. Reddy, Not just age but age and

quality of information, IEEE Journal on Selected Areas in

Communications 39 (5) (2021) 1325–1338.

[43] I. Kadota, M. S. Rahman, E. Modiano, Wifresh: Age-of-

information from theory to implementation, in: IEEE Inter-

national Conference on Computer Communications and Net-

works (ICCCN), 2021, pp. 1–11.

[44] H. H. Yang, C. Xu, X. Wang, D. Feng, T. Q. Quek, Under-

standing age of information in large-scale wireless networks,

IEEE Transactions on Wireless Communications 20 (5) (2021)

3196–3210.

[45] P. D. Mankar, M. A. Abd-Elmagid, H. S. Dhillon, Spatial dis-

tribution of the mean peak age of information in wireless net-

works, IEEE Transactions on Wireless Communications 20 (7)

(2021) 4465–4479.

[46] P. D. Mankar, Z. Chen, M. A. Abd-Elmagid, N. Pappas, H. S.

Dhillon, Throughput and age of information in a cellular-based

IoT network, IEEE Transactions on Wireless Communications

20 (12) (2021) 8248–8263.

[47] X. Chen, K. Gatsis, H. Hassani, S. S. Bidokhti, Age of in-

formation in random access channels, IEEE Transactions on

Information Theory (2022).

[48] Z. Fang, J. Wang, Y. Ren, Z. Han, H. V. Poor, L. Hanzo, Age of

information in energy harvesting aided massive multiple access

networks, IEEE Journal on Selected Areas in Communications

40 (5) (2022) 1441–1456.

[49] F. Chiariotti, J. Holm, A. E. Kalør, B. Soret, S. K. Jensen, T. B.

Pedersen, P. Popovski, Query age of information: Freshness in

pull-based communication, IEEE Transactions on Communi-

cations 70 (3) (2022) 1606–1622.

[50] M. A. Abd-Elmagid, H. S. Dhillon, Age of information in

multi-source updating systems powered by energy harvesting,

IEEE Journal on Selected Areas in Information Theory 3 (1)

(2022) 98–112.

[51] Z. Ling, F. Hu, H. Zhang, Z. Han, Age of information min-

imization in healthcare IoT using distributionally robust opti-

mization, IEEE Internet of Things Journal (2022).

[52] W. Pan, Z. Deng, X. Wang, P. Zhou, W. Wu, Optimizing the

age of information for multi-source information update in In-

ternet of Things, IEEE Transactions on Network Science and

Engineering 9 (2) (2022) 904–917.

[53] S. Farazi, A. G. Klein, D. R. Brown, Average age of infor-

mation for status update systems with an energy harvesting

server, in: IEEE Conference on Computer Communications

Workshops (INFOCOM WKSHPS), 2018, pp. 112–117.

[54] E. T. Ceran, D. G¨

und¨

uz, A. Gy¨

orgy, Average age of informa-

tion with hybrid ARQ under a resource constraint, IEEE Trans-

actions on Wireless Communications 18 (3) (2019) 1900–

1913.

[55] M. A. Abd-Elmagid, H. S. Dhillon, Average peak age-of-

information minimization in UAV-assisted IoT networks, IEEE

Transactions on Vehicular Technology 68 (2) (2019) 2003–

2008.

[56] I. Krikidis, Average age of information in wireless powered

sensor networks, IEEE Wireless Communications Letters 8 (2)

(2019) 628–631.

[57] M. K. Abdel-Aziz, S. Samarakoon, C.-F. Liu, M. Bennis,

W. Saad, Optimized age of information tail for ultra-reliable

low-latency communications in vehicular networks, IEEE

Transactions on Communications 68 (3) (2020) 1911–1924.

[58] B. Yu, Y. Cai, D. Wu, Z. Xiang, Average age of information in

short packet based machine type communication, IEEE Trans-

actions on Vehicular Technology 69 (9) (2020) 10306–10319.

[59] R. V. Bhat, R. Vaze, M. Motani, Throughput maximization

with an average age of information constraint in fading chan-

nels, IEEE Transactions on Wireless Communications 20 (1)

(2021) 481–494.

[60] M. Moltafet, M. Leinonen, M. Codreanu, Average age of infor-

mation for a multi-source M/M/1 queueing model with packet

management, in: IEEE International Symposium on Informa-

tion Theory (ISIT), 2020, pp. 1765–1769.

[61] S. Farazi, A. G. Klein, D. Richard Brown, Average age of infor-

mation in multi-source self-preemptive status update systems

with packet delivery errors, in: 53rd Asilomar Conference on

Signals, Systems, and Computers, 2019, pp. 396–400.

[62] A. Kosta, N. Pappas, A. Ephremides, V. Angelakis, Non-linear

age of information in a discrete time queue: Stationary dis-

tribution and average performance analysis, in: IEEE Interna-

tional Conference on Communications (ICC), 2020, pp. 1–6.

[63] S. Farazi, A. G. Klein, D. R. Brown, Average age of informa-

tion in update systems with active sources and packet delivery

errors, IEEE Wireless Communications Letters 9 (8) (2020)

1164–1168.

[64] J. P. Champati, R. R. Avula, T. J. Oechtering, J. Gross, Min-

imum achievable peak age of information under service pre-

emptions and request delay, IEEE Journal on Selected Areas in

Communications 39 (5) (2021) 1365–1379.

[65] J.-B. Seo, J. Choi, On the outage probability of peak age-of-

information for D/G/1 queuing systems, IEEE Communica-

tions Letters 23 (6) (2019) 1021–1024.

[66] J. Xu, N. Gautam, Peak age of information in priority queu-

ing systems, IEEE Transactions on Information Theory 67 (1)

(2021) 373–390.

[67] Q. He, D. Yuan, A. Ephremides, On optimal link scheduling

with min-max peak age of information in wireless systems,

in: IEEE International Conference on Communications (ICC),

2016, pp. 1–7.

[68] C. Xu, H. H. Yang, X. Wang, T. Q. Quek, On peak age of

information in data preprocessing enabled IoT networks, in:

IEEE Wireless Communications and Networking Conference

(WCNC), 2019, pp. 1–6.

[69] A. Cao, C. Shen, J. Zong, T.-H. Chang, Peak age-of-

information minimization of UAV-aided relay transmission,

in: IEEE International Conference on Communications Work-

shops (ICC Workshops), 2020, pp. 1–6.

[70] R. Singh, G. K. Kamath, P. Kumar, Optimal information up-

dating based on value of information, in: IEEE 57th Annual

Allerton Conference on Communication, Control, and Com-

puting (Allerton), 2019, pp. 847–854.

[71] M. Costa, M. Codreanu, A. Ephremides, On the age of in-

formation in status update systems with packet management,

IEEE Transactions on Information Theory 62 (4) (2016) 1897–

1910.

[72] A. Kosta, N. Pappas, A. Ephremides, V. Angelakis, Age and

value of information: Non-linear age case, in: IEEE Inter-

national Symposium on Information Theory (ISIT), 2017, pp.

326–330.

[73] Y.-P. Hsu, Age of information: Whittle index for scheduling

stochastic arrivals, in: IEEE International Symposium on In-

formation Theory (ISIT), 2018, pp. 2634–2638.

[74] R. Talak, S. Karaman, E. Modiano, Minimizing age-of-

information in multi-hop wireless networks, in: 55th Annual

Allerton Conference on Communication, Control, and Com-

puting (Allerton), 2017, pp. 486–493.

[75] E. Najm, R. Nasser, Age of information: The gamma awaken-

ing, in: IEEE International Symposium on Information Theory

(ISIT), 2016, pp. 2574–2578.

[76] Q. He, D. Yuan, A. Ephremides, Optimizing freshness of in-

formation: On minimum age link scheduling in wireless sys-

tems, in: 14th International Symposium on Modeling and Opti-

mization in Mobile, Ad Hoc, and Wireless Networks (WiOpt),

2016, pp. 1–8.

[77] R. Talak, S. Karaman, E. Modiano, Optimizing information

freshness in wireless networks under general interference con-

straints, IEEE/ACM Transactions on Networking 28 (1) (2020)

15–28.

[78] Q. Abbas, S. Zeb, S. A. Hassan, R. Mumtaz, S. A. R. Zaidi,

Joint optimization of age of information and energy eﬃciency

in IoT networks, in: IEEE 91st Vehicular Technology Confer-

ence (VTC2020-Spring), 2020, pp. 1–5.

[79] Q. Abbas, S. A. Hassan, H. Pervaiz, Q. Ni, A Markovian model

for the analysis of age of information in IoT networks, IEEE

Wireless Communications Letters, (2021), 10(7), 1596-1600.

[80] G. Stamatakis, N. Pappas, A. Traganitis, Optimal policies for

status update generation in an IoT device with heterogeneous

traﬃc, IEEE Internet of Things Journal 7 (6) (2020) 5315–

5328.

[81] J. Zhong, R. D. Yates, E. Soljanin, Timely lossless source cod-

ing for randomly arriving symbols, in: IEEE Information The-

ory Workshop (ITW), 2018, pp. 1–5.

[82] J. P. Hespanha, Modelling and analysis of stochastic hybrid

systems, IEEE Proceedings-Control Theory and Applications

153 (5) (2006) 520–535.

[83] O. M. Sleem, S. Leng, A. Yener, Age of information mini-

mization in wireless powered stochastic energy harvesting net-

works, in: IEEE 54th Annual Conference on Information Sci-

ences and Systems (CISS), 2020, pp. 1–6.

[84] E. U. Atay, I. Kadota, E. Modiano, Aging wireless bandits:

Regret analysis and order-optimal learning algorithm, in: IEEE

19th International Symposium on Modeling and Optimization

in Mobile, Ad hoc, and Wireless Networks (WiOpt), 2021, pp.

1–8.

[85] T. Soleymani, J. S. Baras, K. H. Johansson, Stochastic con-

trol with stale information–part i: Fully observable systems,

in: IEEE 58th Conference on Decision and Control (CDC),

2019, pp. 4178–4182.

[86] A. Gong, T. Zhang, H. Chen, Y. Zhang, Age-of-information-

based scheduling in multiuser uplinks with stochastic arrivals:

A POMDP approach, in: IEEE GLOBECOM, 2020, pp. 1–6.

[87] H. H. Yang, A. Arafa, T. Q. Quek, H. V. Poor, Optimizing in-

formation freshness in wireless networks: A stochastic geome-

try approach, IEEE Transactions on Mobile Computing 20 (6)

(2020) 2269–2280.

[88] A. Sinha, R. Bhattacharjee, Optimizing age-of-information in

adversarial and stochastic environments, IEEE Transactions on

Information Theory (2022).

[89] K. Saurav, R. Vaze, Minimizing the sum of age of infor-

mation and transmission cost under stochastic arrival model,

in: IEEE Conference on Computer Communications (INFO-

COM), 2021, pp. 1–10.

[90] P. D. Mankar, M. A. Abd-Elmagid, H. S. Dhillon, Stochas-

tic geometry-based analysis of the distribution of peak age of

information, in IEEE International Conference on Communi-

cations (ICC), 2021, pp. 1–6.

[91] B. Li, H. Chen, Y. Zhou, Y. Li, Age-oriented opportunistic re-

laying in cooperative status update systems with stochastic ar-

rivals, in: IEEE GLOBECOM, 2020, pp. 1–6.

[92] M. Moltafet, M. Leinonen, M. Codreanu, N. Pappas, Power

minimization in wireless sensor networks with constrained AoI

using stochastic optimization, in: 53rd IEEE Asilomar Confer-

ence on Signals, Systems, and Computers, 2019, pp. 406–410.

[93] I. Kadota, E. Modiano, Minimizing the age of information in

wireless networks with stochastic arrivals, IEEE Transactions

on Mobile Computing 20 (3) (2019) 1173–1185.

[94] N. Akar, E. Karasan, Scheduling algorithms for age of in-

formation diﬀerentiation with random arrivals, arXiv preprint

arXiv:2110.10992 (2021).

[95] E. Orkun Gamgam, N. Akar, Exact Analytical Model of Age

of Information in Multi-Source Status Update Systems with

Per-Source Queueing, (2022), IEEE Internet of Things Journal,

Early Access.

[96] J. Pan, A. M. Bedewy, Y. Sun, N. B. Shroﬀ, Minimizing age of

information via scheduling over heterogeneous channels, in:

Proceedings of the Twenty-second International Symposium

on Theory, Algorithmic Foundations, and Protocol Design for

Mobile Networks and Mobile Computing, 2021, pp. 111–120.

[97] B. Choudhury, V. K. Shah, A. Ferdowsi, J. H. Reed, Y. T. Hou,

AoI-minimizing scheduling in UAV-relayed IoT networks, in:

IEEE 18th International Conference on Mobile Ad Hoc and

Smart Systems (MASS), 2021, pp. 117–126.

[98] O. Ozel, P. Raﬁee, Intermittent status updating through joint

scheduling of sensing and retransmissions. In IEEE Confer-

ence on Computer Communications Workshops (INFOCOM

WKSHPS), (2021), (pp. 1-6).

[99] G. Chen, S. C. Liew, Y. Shao, Uncertainty-of-information

scheduling: A restless multi-armed bandit framework. IEEE

Transactions on Information Theory, (2022), Early Access.

[100] C. Li, Q. Liu, S. Li, Y. Chen, Y. T. Hou, W. Lou, On scheduling

with AoI violation tolerance, in: IEEE INFOCOM, 2021, pp.

1–9.

[101] K. Chen, L. Ding, F. Yang, L. Qian, MSE minimized schedul-

ing for multiple-source remote estimation with AoI constraints

in IWSN, in: IEEE International Conference on Wireless Com-

munications and Signal Processing (WCSP), 2020, pp. 1119–

1124.

[102] M. Li, C. Chen, H. Wu, X. Guan, X. Shen, Age-of-information

aware scheduling for edge-assisted industrial wireless net-

works, IEEE Transactions on Industrial Informatics (2020),

17 (8) 5562–5571.

[103] B. Han, Y. Zhu, Z. Jiang, M. Sun, H. D. Schotten, Fairness for

freshness: Optimal age of information based OFDMA schedul-

ing with minimal knowledge, IEEE Transactions on Wireless

Communications (2021), 20(12), 7903-7919.

[104] Y. Shao, Q. Cao, S. C. Liew, H. Chen, Partially observable

minimum-age scheduling: The greedy policy, IEEE Transac-

tions on Communications (2021), 70(1), 404-418.

[105] A. M. Bedewy, Y. Sun, R. Singh, N. B. Shroﬀ, Optimizing in-

formation freshness using low-power status updates via sleep-

wake scheduling, in: Proceedings of the Twenty-First Interna-

tional Symposium on Theory, Algorithmic Foundations, and

Protocol Design for Mobile Networks and Mobile Computing,

2020, pp. 51–60.

[106] A. Maatouk, S. Kriouile, M. Assaad, A. Ephremides, Asymp-

totically optimal scheduling policy for minimizing the age of

information, in: IEEE International Symposium on Informa-

tion Theory (ISIT), 2020, pp. 1747–1752.

[107] C. Li, S. Li, Y. Chen, Y. T. Hou, W. Lou, AoI scheduling with

maximum thresholds, in: IEEE INFOCOM 2020-IEEE Con-

ference on Computer Communications, 2020, pp. 436–445.

[108] J. Song, D. Gunduz, W. Choi, Optimal scheduling policy for

minimizing age of information with a relay, arXiv preprint

arXiv:2009.02716 (2020).

[109] O. Ayan, H. M. G¨

ursu, S. Hirche, W. Kellerer, AoI-based ﬁ-

nite horizon scheduling for heterogeneous networked control

systems, in: IEEE GLOBECOM, 2020, pp. 1–7.

[110] H. Ma, S. Zhou, X. Zhang, L. Xiao, Transmission scheduling

for multi-loop wireless networked control based on LQ cost

oﬀset, in: IEEE Conference on Computer Communications

Workshops (INFOCOM WKSHPS), 2020, pp. 19–25.

[111] B. Sombabu, S. Moharir, Age-of-information based schedul-

ing for multi-channel systems, IEEE Transactions on Wireless

Communications (2020), 19 (7), 4439–4448.

[112] Q. Wang, H. Chen, C. Zhao, Y. Li, P. Popovski, B. Vucetic, Op-

timizing information freshness via multiuser scheduling with

adaptive NOMA/OMA, IEEE Transactions on Wireless Com-

munications (2021), 21(3), 1766-1778.

[113] H. Pan, S. C. Liew, Information update: TDMA or FDMA?,

IEEE Wireless Communications Letters 9 (6) (2020) 856–860.

[114] N. Hirosawa, H. Iimori, G. T. F. de Abreu, K. Ishibashi, Age-

of-information minimization in two-user multiple access chan-

nel with energy harvesting, in: IEEE 8th International Work-

shop on Computational Advances in Multi-Sensor Adaptive

Processing (CAMSAP), 2019, pp. 361–365.

[115] C. Guo, S. Wu, Z. Deng, J. Jiao, N. Zhang, Q. Zhang,

Age-optimal power allocation policies for NOMA and hybrid

NOMA/OMA systems, in: IEEE International Conference on

Communications (ICC), 2021, pp. 1–6.

[116] L. Liu, H. H. Yang, C. Xu, F. Jiang, On the peak age of infor-

mation in NOMA IoT networks with stochastic arrivals, IEEE

Wireless Communications Letters 10 (12) (2021) 2757–2761.

[117] J. T. G´

omez, M. Morales-C´

espedes, A. G. Armada, G. Hirtz,

Minimizing age of information on NOMA communication

schemes for vehicular communication applications, in: 12th

IEEE International Symposium on Communication Systems,

Networks and Digital Signal Processing (CSNDSP), 2020, pp.

1–6.

[118] S. Liao, J. Jiao, S. Wu, R. Lu, Q. Zhang, Age-optimal power

allocation scheme for NOMA-based S-IoT downlink network,

in: IEEE International Conference on Communications (ICC),

2021, pp. 1–6.

[119] Z. Deng, S. Wu, C. Guo, J. Jiao, N. Zhang, Q. Zhang, Age-

optimal transmission policy for intelligent HARQ-CC aided

NOMA systems, in: IEEE International Conference on Com-

munications (ICC) 2021, pp. 1–6.

[120] J. Chen, Y. Liu, Q. Chen, X. Lan, G. Cheng, Y. Fu, Z. Zhang,

On the adaptive AoI-aware buﬀer-aided transmission scheme

for NOMA networks, in: IEEE Wireless Communications and

Networking Conference (WCNC), 2021, pp. 1–6.

[121] H. Azarhava, M. P. Abdollahi, J. M. Niya, Age of information

in wireless powered IoT networks: NOMA vs. TDMA, Ad Hoc

Networks 104 (2020) 102179.

[122] A. Maatouk, M. Assaad, A. Ephremides, Minimizing the age

of information: NOMA or OMA?, in: IEEE Conference

on Computer Communications Workshops (INFOCOM WK-

SHPS), IEEE, 2019, pp. 102–108.

[123] N. Suma, S. R. Samson, S. Saranya, G. Shanmugapriya,

R. Subhashri, IoT based smart agriculture monitoring system,

International Journal on Recent and Innovation Trends in com-

puting and communication 5 (2) (2017) 177–181.

[124] K. Ashokkumar, B. Sam, R. Arshadprabhu, et al., Cloud based

intelligent transport system, Procedia Computer Science 50

(2015) 58–63.

[125] S. Kaul, M. Gruteser, V. Rai, J. Kenney, Minimizing age of

information in vehicular networks, in: 8th Annual IEEE Com-

munications Society Conference on Sensor, Mesh and Ad Hoc

Communications and Networks, 2011, pp. 350–358.

[126] M. Patra, A. Sengupta, C. S. R. Murthy, On minimizing the

system information age in vehicular ad-hoc networks via eﬃ-

cient scheduling and piggybacking, Wireless Networks 22 (5)

(2016) 1625–1639.

[127] Y. Ni, L. Cai, Y. Bo, Vehicular beacon broadcast scheduling

based on age of information (AoI), China Communications

15 (7) (2018) 67–76.

[128] X. Zhang, Q. Zhu, H. V. Poor, Analyses for age of informa-

tion supporting URLLC over multimedia wireless networks,

in: IEEE 21st International Workshop on Signal Processing

Advances in Wireless Communications (SPAWC), 2020, pp.

1–5.

[129] C. M. W. Basnayaka, D. N. K. Jayakody, Z. Chang, Age of

information based URLLC-enabled UAV wireless communi-

cations system, IEEE Internet of Things Journal (2021) 1–1.

[130] A. Elgabli, H. Khan, M. Krouka, M. Bennis, Reinforce-

ment learning based scheduling algorithm for optimizing age

of information in ultra reliable low latency networks, in:

IEEE Symposium on Computers and Communications (ISCC),

2019, pp. 1–6.

[131] C. Chaccour, W. Saad, On the ruin of age of informa-

tion in augmented reality over wireless terahertz (THz) net-

works, in: IEEE GLOBECOM, 2020, pp. 1–6. @inpro-

ceedingssun2018information, title=Information aging through

queues: A mutual information perspective, author=Sun, Yin

and Cyr, Benjamin, booktitle=2018 IEEE 19th International

Workshop on Signal Processing Advances in Wireless Com-

munications (SPAWC), pages=1–5, year=2018, organiza-

tion=IEEE

[132] Kaul, S. & Yates, R. Timely updates by multiple sources: The

M/M/1 queue revisited. 2020 54th Annual Conference On In-

formation Sciences And Systems (CISS). pp. 1-6.

[133] Sun, Y. & Cyr, B. Information aging through queues: A mutual

information perspective. 2018 IEEE 19th International Work-

shop On Signal Processing Advances In Wireless Communica-

tions (SPAWC). pp. 1-5.

[134] Kosta, A., Pappas, N., Ephremides, A. & Angelakis, V. The age

of information in a discrete time queue: Stationary distribution

and non-linear age mean analysis. IEEE Journal On Selected

Areas In Communications.39, 1352-1364 (2021)

[135] Modina, N., El Azouzi, R., De Pellegrini, F., Menasche, D.

& Figueiredo, R. Joint traﬃc oﬄoading and aging control in

5g iot networks. IEEE Transactions On Mobile Computing.

(2022)

[136] Abbas, Q., Hassan, S., Jung, H. & Hossain, M. On Minimizing

the Age of Information in NOMA-Based Vehicular Networks

Using Markov Decision Process. IEEE Transactions On Intel-

ligent Transportation Systems. (2022)

[137] Shaﬁ, U., Mumtaz, R., Garcıa-Nieto, J., Hassan, S., Zaidi, S. &

Iqbal, N. Precision agriculture techniques and practices: From

considerations to applications. Sensors.19, 3796 (2019)

[138] He, W., Guo, C. & Wang, X. Age of Information Aware Re-

source Allocation and Packet Sampling Control in Vehicular

Networks. IEEE Wireless Communications Letters. (2022)

[139] Schoeﬀauer, R. & Wunder, G. Age-of-Information in Clocked

Networks. ArXiv Preprint ArXiv:2207.05516. (2022)

[140] Jin, W., Sun, J., Chi, K. & Zhang, S. Deep reinforcement

learning based scheduling for minimizing age of information

in wireless powered sensor networks. Computer Communica-

tions.191 pp. 1-10 (2022)

[141] Salimnejad, M. & Pappas, N. On the Age of Information in a

Two-User Multiple Access Setup. Entropy.24, 542 (2022)

[142] G´

omez, J., Pitke, K., Stratmann, L. & Dressler, F. Age of infor-

mation in molecular communication channels. Digital Signal

Processing.124 pp. 103108 (2022)

[143] Jin, W., Huang, L. & Chi, K. Age of Information Minimiza-

tion in Wireless Powered NOMA Communication Networks.

2022 IEEE 23rd International Conference On High Perfor-

mance Switching And Routing (HPSR). pp. 201-205 (2022)

[144] Miridakis, N., Shi, Z., Tsiftsis, T. & Yang, G. Extreme Age of

Information for Wireless-Powered Communication Systems.

IEEE Wireless Communications Letters.11, 826-830 (2022)

An exploratory bibliometric analysis of the literature on the age of information-aware unmanned aerial vehicles aided communication

Article

Full-text available

Aug 2023

Hybrid Grey PIPRECIA and Grey OCRA method-based dynamic multi-criteria decision-making model for mitigating non-cooperating node attacks in WSNs

Article

Full-text available

Aug 2023

Cooperation among sensor nodes is always essential for achieving reliable data dissemination in Wireless Sensor Network (WSNs). The sensor nodes can cat as a host or a router in which they establish communication between. In this context, the reliability of intermediate sensor nodes that establishes the routing path between the source and destination need to be dynamically assessed for enforcing cooperation. Thus, potential dynamic multi-criteria decision-making model for enforcing cooperation is necessary as the presence of malicious and selfish nodes in WSNs. Which severely crumbles the network performance. In this paper, Hybrid Grey PIPRECIA and Grey OCRA Method (HGP-GOCRAM) -based Dynamic Multi-Criteria Decision-Making Model is proposed for thwarting malicious and selfish nodes for achieving maximized cooperation among the sensor nodes of the routing path. This HGP-GOCRAM specifically used Grey Pivot Pairwise Relative Criteria Importance Assessment (PIPRECIA-G) for determining the weights of the criteria considered for evaluating the cooperation degree of the sensor nodes. It further used Grey Operational Competitiveness Rating (OCRA-G) for identifying the rank of the sensor nodes to isolate the worst cooperating nodes dynamically from the rouging path. The simulation results of the proposed HGP-GOCRAM achieved under different malicious and selfish node confirmed an improved packet delivery rate of 22.31%, maximized throughput of 19.78%, with minimized energy consumptions of 20.98% and end-to-end delay of 22.64%, better than the competitive cooperation enforcement approaches used for investigation. The results also proved that the proposed HGP-GOCRAM is capable in achieving rapid and accurate detection and isolation of malicious and selfish nodes.

Satisfying strict deadlines for cellular Internet of Things through hybrid multiple access

Article

Full-text available

Jul 2023

Latency‐constrained aspects of cellular Internet of Things (IoT) applications rely on Ultra‐Reliable and Low Latency Communications (URLLC), which highlight research on satisfying strict deadlines. In this study, we address the problem of latency‐constrained communications with strict deadlines under average power constraint using Hybrid Multiple Access (MA), which consists of both Orthogonal MA (OMA) and power domain Non‐Orthogonal MA (NOMA) as transmission scheme options. We aim to maximize the timely throughput, which represents the average number of successfully transmitted packets before deadline expiration, where expired packets are dropped from the buffer. We use Lyapunov stochastic optimization methods to develop a dynamic power assignment algorithm for minimizing the packet drop rate while satisfying time average power constraints. Moreover, we propose a flexible packet dropping mechanism called Early Packet Dropping (EPD) to detect likely to become expired packets and drop them immediately. Numerical results show that Hybrid MA improves the timely throughput compared to conventional OMA by up to 46%$$ 46\% $$ and on average by more than 21%$$ 21\% $$. With EPD, these timely throughput gains improve to 53%$$ 53\% $$ and 24.5%$$ 24.5\% $$, respectively.

The Age of Information in Wireless Cellular Systems: Gaps, Open Problems, and Research Challenges

Article

Full-text available

Oct 2023
SENSORS-BASEL

One of the critical use cases for prospective fifth generation (5G) cellular systems is the delivery of the state of the remote systems to the control center. Such services are relevant for both massive machine-type communications (mMTC) and ultra-reliable low-latency communications (URLLC) services that need to be supported by 5G systems. The recently introduced the age of information (AoI) metric representing the timeliness of the reception of the update at the receiver is nowadays commonly utilized to quantify the performance of such services. However, the metric itself is closely related to the queueing theory, which conventionally requires strict assumptions for analytical tractability. This review paper aims to: (i) identify the gaps between technical wireless systems and queueing models utilized for analysis of the AoI metric; (ii) provide a detailed review of studies that have addressed the AoI metric; and (iii) establish future research challenges in this area. Our major outcome is that the models proposed to date for the AoI performance evaluation and optimization deviate drastically from the technical specifics of modern and future wireless cellular systems, including those proposed for URLLC and mMTC services. Specifically, we identify that the majority of the models considered to date: (i) do not account for service processes of wireless channel that utilize orthogonal frequency division multiple access (OFDMA) technology and are able to serve more than a single packet in a time slot; (ii) neglect the specifics of the multiple access schemes utilized for mMTC communications, specifically, multi-channel random access followed by data transmission; (iii) do not consider special and temporal correlation properties in the set of end systems that may arise naturally in state monitoring applications; and finally, (iv) only few studies have assessed those practical use cases where queuing may happen at more than a single node along the route. Each of these areas requires further advances for performance optimization and integration of modern and future wireless provisioning technologies with mMTC and URLLC services.

Low-power timely random access: Packet-based or connection-based?

Article

Jun 2024
COMPUT COMMUN

This paper studies low-power random access protocols for timely status update systems with information freshness requirements, measured by age of information (AoI). A fundamental challenge for such networks is to schedule a large number of transmitters to access the wireless channel in a way that achieves a low network-wide AoI at low power consumption. Conventional packet-based random access protocols involve transmitters contending for the channel by sending their entire data packets. When packets are of long duration, the time and energy wasted due to packet collisions is considerable. In contrast, connection-based random access protocols establish connections with the receiver before transmitting data packets. From an information freshness perspective, there should be conditions that favor one approach over the other. Therefore, we conduct a comparative study of the average AoI of packet-based and connection-based random access protocols. Specifically, we consider frame slotted Aloha (FSA) as a representative of packet-based random access and design a request-then-access (RTA) protocol for connection-based random access. Our analyses indicate that the choice between packet-based and connection-based protocols depends mainly on the payload size of update packets and the transmit power budget. In particular, RTA significantly reduces AoI and saves power, especially when the payload size is large. Overall, our investigation offers insights into the practical design of random access protocols for low-power timely status update systems.

Age of information minimization in hybrid cognitive radio networks under a timely throughput constraint

Article

Feb 2024
PERFORM EVALUATION

CSformer: Enhancing deep learning efficiency for intelligent IoT

Article

Nov 2023
COMPUT COMMUN

Age of Information in Internet of Things: A Survey

Article

Jan 2023

In recent years, the increasing demand to see the status of objects over the internet leads to an increase in the number of Internet of things (IoT) applications. The unique nature of IoT, which involves potentially millions of interconnected devices with different data rate, power, bandwidth and range specifications, require different performance metrics than those conventionally employed in other communication applications. In conventional wireless communication systems such as cellular networks, performance indicators including data rate and spectral efficiency have become decisive, whereas in energy-constrained real-time IoT applications which require low data rate, the freshness of information has become a more prominent characteristic. Age of information (AoI), which is the elapsed time after the last received packet update was created at the source, has emerged as a fundamental metric for determining the freshness of information and has attracted substantial research interest. In this regard, this paper is dedicated to provide an overview of the current state-of-the-art on the use of AoI for the design and optimization of a large variety IoT applications. After a brief introduction of the IoT and AoI fundamentals, this paper presents a survey of the research works on common design issues such as AoI based optimization, scheduling for IoT networks, application of learning methods in large scale IoT systems, real life applications and experimental results together with a synopsis of potential future applications and research challenges.

AoI-Based Transmission Scheduling for Cyber-Physical Systems Over Fading Channel Against Eavesdropping

Article

Jan 2023

This paper investigates the transmission scheduling problem of cyber-physical systems (CPSs). Specifically, in a CPS, a sensor collects real-time data of the monitoring area and at each decision epoch, the system must determine whether to transmit data packets to the gateway through an unreliable wireless channel. Furthermore, we assume that the CPS is subject to an energy-harvesting (EH) eavesdropper, and the communication channel is wiretapped randomly when the harvested energy of the eavesdropper is sufficient. The objective is to obtain the optimal transmission scheduling to minimize the Age of Information (AoI) of the CPS while keeping the AoI of eavesdropper above a certain level. To achieve this, we first transform the system model into a Markov decision process (MDP). We then prove that the optimal transmission scheduling policy is a threshold behavior on the AoI of both the CPS and the eavesdropper, respectively. Based on the structural properties of the optimal policy, we have developed a new backward induction algorithm to compute the optimal AoI-based transmission scheduling and the performance index function with lower computational costs compared to the conventional induction algorithm. Finally, we verify the validity of the algorithm and the correctness of the theoretical results through simulations.

AoI-Oriented Status Updating in Large-scale Heterogeneous Multi-Channel Systems

Conference Paper

Jun 2023

Age of Information in Random Access Channels

Article

Full-text available

Oct 2022

In applications of remote sensing, estimation, and control, timely communication is critical but not always ensured by high-rate communication. This work proposes decentralized age-efficient transmission policies for random access channels with $M$ transmitters. We propose the notion of age-gain of a packet to quantify how much the packet will reduce the instantaneous age of information at the receiver side upon successful delivery. We then utilize this notion to propose a transmission policy in which transmitters act in a decentralized manner based on the age-gain of their available packets. In particular, each transmitter sends its latest packet only if its corresponding age-gain is beyond a certain threshold which could be computed adaptively using the collision feedback or found as a fixed value analytically in advance. Both methods improve age of information significantly compared to the state of the art. In the limit of large $M$ , we prove that when the arrival rate is small (below $\frac {1}{eM}$ ), slotted ALOHA-type algorithms are order optimal. As the arrival rate increases beyond $\frac {1}{eM}$ , while age increases under slotted ALOHA, it decreases significantly under the proposed age-based policies. For arrival rates $\theta $ , $\theta =\frac {1}{o(M)}$ , the proposed algorithms provide a multiplicative gain of at least two compared to the minimum age under slotted ALOHA (minimum over all arrival rates). We conclude that it is beneficial to increase the sampling rate (and hence the arrival rate) and transmit packets selectively based on their age-gain. This is surprising and contrary to common practice where the arrival rate is optimized to attain the minimum AoI. We further extend our results to other random access technologies such as Carrier-sense multiple access (CSMA).

Optimizing Age-of-Information in Adversarial and Stochastic Environments

Article

Full-text available

Jun 2022

We design efficient online scheduling policies to maximize the freshness of information delivered to the users in a cellular network under both adversarial and stochastic channel and mobility assumptions. The information freshness achieved by a policy is investigated through the lens of a recently proposed metric-Age-of-Information (AoI). We show that a natural greedy scheduling policy is competitive against any optimal offline policy in minimizing the AoI in the adversarial setting. We also derive universal lower bounds to the competitive ratio achievable by any online policy in the adversarial framework. In the stochastic setting, we show that a simple index policy is near-optimal for minimizing the average AoI in two different mobility scenarios. Further, we prove that the greedy scheduling policy minimizes the peak AoI for static users in the stochastic setting. Simulation results show that the proposed policies perform well under realistic conditions.

On Minimizing the Age of Information in NOMA-Based Vehicular Networks Using Markov Decision Process

Article

Full-text available

May 2022

Network sustainability relies on many important parameters where the timely dissemination of information has a prime role to improve network operations and henceforth the network sustainability. Age of Information is a critical metric in many applications of future networks including smart transportation systems as these networks require fresh updates from the various network entities for the successful delivery of their services. This paper considers smart vehicles in a vehicle-to-infrastructure network where each vehicle has a stream of data for transmission to the roadside unit (RSU). The information from vehicles is collected when they enter the communication range of an RSU and stay within the coverage area of that RSU for a particular time. During this time, the RSU attempts to receive information from each vehicle as timely as possible. This paper proposes a hybrid access mechanism consisting of both orthogonal and non-orthogonal multiple access that schedules the transmission of packets from vehicles to the RSU where each vehicle has a finite length queue. The transmission of the packets is modeled using a Markov decision process, where a specific cost function is optimized to collect maximum information from the vehicles in a minimum amount of time.

On the Age of Information in a Two-User Multiple Access Setup

Article

Full-text available

Apr 2022
Entropy

This work considers a two-user multiple access channel in which both users have Age of Information (AoI)-oriented traffic with different characteristics. More specifically, the first user has external traffic and cannot control the generation of status updates, and the second user monitors a sensor and transmits status updates to the receiver according to a generate-at-will policy. The receiver is equipped with multiple antennas and the transmitters have single antennas; the channels are subject to Rayleigh fading and path loss. We analyze the average AoI of the first user for a discrete-time first-come-first-served (FCFS) queue, last-come-first-served (LCFS) queue, and queue with packet replacement. We derive the AoI distribution and the average AoI of the second user for a threshold policy. Then, we formulate an optimization problem to minimize the average AoI of the first user for the FCFS and LCFS with preemption queue discipline to maintain the average AoI of the second user below a given level. The constraints of the optimization problem are shown to be convex. It is also shown that the objective function of the problem for the first-come-first-served queue policy is non-convex, and a suboptimal technique is introduced to effectively solve the problem using the algorithms developed for solving a convex optimization problem. Numerical results illustrate the performance of the considered optimization algorithm versus the different parameters of the system. Finally, we discuss how the analytical results of this work can be extended to capture larger setups with more than two users.

Value of Information in Wireless Sensor Network Applications and the IoT: A Review

Article

Full-text available

May 2022

Value of Information (VoI) is a concept to assess the usefulness of information for a specific goal, and has in the last decade experienced a growing interest also for Wireless Sensor Network (WSN) applications and the Internet of Things (IoT). By making the value of information explicit in the form of VoI, WSN and IoT applications should be able to better assess which information to spend their constrained resources on. However, the definition of VoI is highly application-dependent, which has led to a fragmented understanding of VoI, and there is a lack of a comprehensive overview. In this structured review, we first categorize application use cases and examine what VoI is used for, and explore the different approaches to defining VoI. We then provide a well-structured and comprehensive discussion of the specific approaches used in the literature to determine VoI, together with examples of use cases. We categorize the different approaches to calculating VoI, describe their properties systematically and distinguish between observed VoI and expected VoI. We also discuss adaptive VoI approaches and point towards future directions within the field.

Age of Information Aware Resource Allocation and Packet Sampling Control in Vehicular Networks

Article

Nov 2022

This letter considers a vehicular network, where multiple vehicle-to-vehicle (V2V) connections carrying status update service share spectrum with multiple capacity-hungry vehicle-to-network (V2N) links. To attain reliable data transmission, packet retransmission is performed by V2V links in case of delivery failure, which allows us to formulate each V2V’s queueing system as a D/Geo/1 queueing model. To this end, we analyze the age of information (AoI) outage probability for the D/Geo/1 queueing system. Then, we formulate a problem of resource allocation and packet sampling rate optimization to maximize the sum ergodic capacity of V2N links while guaranteeing the AoI outage probability of V2V links. For each possible V2N-V2V spectrum reusing pair, the optimal power allocation and packet sampling rate are obtained by a low-complexity algorithm. Afterwards, spectrum allocation is optimized by bipartite matching. Simulation results validate our AoI analysis and show the effectiveness of the proposed algorithm.

Age of Information Minimization in Wireless Powered NOMA Communication Networks

Conference Paper

Jun 2022

Exact Analytical Model of Age of Information in Multi-Source Status Update Systems With Per-Source Queueing

Article

Jun 2022

We study a multisource status update system with Poisson information packet arrivals and exponentially distributed service times. The server is equipped with a waiting room holding the freshest packet from each source referred to as single buffer per-source queueing (SBPSQ). The sources are assumed to be equally important, i.e., (nonweighted) average Age of Information (AoI) or average age violation probability are used as the information freshness metrics to optimize for, and subsequently, two symmetric SBPSQ-based scheduling policies are studied in this article, namely, first source first serve (FSFS) and the earliest served first serve (ESFS) policies. By employing the theory of Markov fluid queues (MFQs), an analytical model is proposed to obtain the exact distribution of the AoI for each source when the FSFS and ESFS policies are employed at the server. Additionally, a benchmark scheduling-free scheme named single buffer with replacement (SBR), which uses a single buffer to hold the freshest packet across all sources, is also studied with a similar but less complex analytical model. We comparatively study the performance of the three policies through numerical examples in terms of the average AoI and the age violation probability averaged across all sources, in a scenario of sources possessing different traffic intensities but sharing a common service time.

Uncertainty-of-Information Scheduling: A Restless Multi-armed Bandit Framework

Article

Sep 2022

This paper proposes using the uncertainty of information (UoI), measured by Shannon’s entropy, as a metric for information freshness. We consider a system in which a central monitor observes M binary Markov processes through m communication channels (m<M). The UoI of a Markov process corresponds to the monitor’s uncertainty about its state. At each time step, only m Markov processes can be selected to update its state to the monitor; hence there is a tradeoff among the UoIs of the processes that depend on the scheduling policy used to select the processes to be updated. The age of information (AoI) of a process corresponds to the time since its last update. In general, the associated UoI can be a non-increasing function, or even an oscillating function, of its AoI, making the scheduling problem particularly challenging. This paper investigates scheduling policies that aim to minimize the average sum-UoI of the processes over the infinite time horizon. We formulate the problem as a restless multi-armed bandit (RMAB) problem, and develop a Whittle index policy that is near-optimal for the RMAB after proving its indexability. We further provide an iterative algorithm to compute the Whittle index for the practical deployment of the policy. Although this paper focuses on UoI scheduling, our results apply to a general class of RMABs for which the UoI scheduling problem is a special case. Specifically, this paper’s Whittle index policy is valid for any RMAB in which the bandits are binary Markov processes and the penalty is a concave function of the belief state of the Markov process. Numerical results demonstrate the excellent performance of the Whittle index policy for this class of RMABs.

Deep reinforcement learning based scheduling for minimizing age of information in wireless powered sensor networks

Article

Apr 2022
COMPUT COMMUN

For real-time monitoring system, the age of information (AoI) is usually used to quantify the freshness of information at a monitor about some stochastic processes observed by the source node. In this paper, we consider the wireless powered sensor networks (WPSNs) where multiple sensor nodes send update packets to the base station. Time is divided into slots of equal duration and at each slot either wireless energy transfer or packet transmission is conducted. We aim to minimize the long-term average weighted sum-AoI of different processes at the base station. Specifically, we first formulate the average weighted sum-AoI minimization problem as a multi-stage stochastic non-linear integer programming (NLP) subject to the energy causality constraints. Second, we design an algorithm which first applies Lyapunov optimization to decouple the multi-stage stochastic NLP into per-frame deterministic NLP problems. Then in each frame, our algorithm utilizes the model-free DRL to solve the per-frame NLP problem with very low computational complexity where one exploration policy is designed to obtain multiple one-hot candidate actions based on single real-number output of neural network. We demonstrate through simulations that, our proposed algorithm can achieve greatly smaller average weighted sum-AoI than the available DQN-based algorithm and also alleviate the problem that some source nodes may have large instantaneous AoIs.

A Comprehensive Survey on Age of Information in Massive IoT Networks

Abstract

Recommended publications

On Minimizing the Age of Information in NOMA-Based Vehicular Networks Using Markov Decision Process

Q2A-NOMA: A Q-Learning-based QoS-Aware NOMA System Design for Diverse Data Rate Requirements

Scheduling to Minimize Age of Information with Multiple Sources

Optimizing Age of Information in RIS-Empowered Uplink Cooperative NOMA Networks