Content uploaded by Sahra Sedigh Sarvestani
Author content
All content in this area was uploaded by Sahra Sedigh Sarvestani on Sep 03, 2014
Content may be subject to copyright.
SmartBrick wireless sensor node for high-resolution
structural health monitoring
A. Gunasekaran, P. Chulick, and S. Sedigh
Department of Electrical and Computer Engineering
Missouri University of Science and Technology, USA
ABSTRACT
This paper introduces a wireless sensor node for the SmartBrick platform, which provides a low-cost, autonomous
method for structural health monitoring. Design and testing of the SmartBrick base station have been described
in previous publications. The SmartBrick sensor node presented in this paper leverages Zigbee short-range
communication to communicate with the base station. This facilitates an increase in the monitoring range of the
system. The primary function of the node is to measure humidity, temperature, tilt, strain, and vibration; then
transmit these values to the base station via Zigbee. The GSM modem, which is included in the base station, has
been omitted from the sensor node. This reduces cost, size, and power consumption. Long-range communication
of data and alerts is carried out by the base station. It serves as the gateway to the outside world and relays
remote configuration and maintenance commands to the sensor nodes. In this paper we discuss the integration
of the Zigbee daughterboard with the SmartBrick, and the development and testing of highly-capable sensor
nodes with extensible features.
Keywords: structural health monitoring, sensor nodes, low power sensing, Zigbee, distributed sensor network,
embedded sensors, remote monitoring
1. INTRODUCTION
Aging and degradation of critical infrastructure, in particular bridges, is rapidly increasing the need for high
resolution structural health monitoring (SHM). The SmartBrick SHM platform, and its predecessor, the Flood
Frog, have proven to be accurate and cost-effective solutions for remote bridge monitoring, as articulated in our
previous publications on the topic, e.g.,.1, 2 Both devices were capable of collecting and reporting data from
a single strain gauge; however, they were fell short of delivering the high resolution required for accurate and
precise monitoring of strain from multiple locations on a bridge.
Our first attempt at addressing this challenge involved the development of Zigbee-enabled sensor nodes,
based on off-the-shelf hardware (Texas Instruments ez430-rf2480 modules).3–5 Lack of customizability led to
problems in integrating these nodes with other sensors. Their high power consumption was also a problem, as
long (multi-year) unattended field life is a major design objective of the SmartBrick platform.
This paper articulates our most recent attempt at facilitating high-resolution SHM with the SmartBrick. We
have designed a daughterboard using the CC2480 Zigbee network processor shown in Figure 1. The daughter-
board was designed as a reusable plugin module that can be easily integrated with current and future SmartBrick
modules. The daughterboard was interfaced with the SmartBrick base station and sensor nodes for testing and
development. A star network comprised of two sensor nodes with a SmartBrick as the coordinator was used
for testing. Temperature data from the sensor nodes was collected and transmitted to the SmartBrick, which
in turn delivered the data to the serial port of a computer. The coordinator and sensor nodes were subjected
to thorough laboratory testing, and the network setup was similarly validated in laboratory conditions. The
sensor nodes were tested at two different voltage levels, to ensure low power consumption. Results are reported
in Section 4.1.4.
The purpose-built sensor nodes described in this paper further enrich the suite of sensing options provided
by the SmartBrick platform, and facilitates use of the platform in applications requiring high-resolution sensing.
Further author information, please contact S. Sedigh: E-mail: sedighs@mst.edu, Telephone: +1 573 341 7505
Figure 1 shows the block diagram of the information flow in SmartBrick, which presents a holistic view of
the entire network. The sensor nodes collect the data from onboard and external sensors and send them to the
base station at hourly intervals. The base station can send the data via SMS, Email or FTP to the background
IT infrastructure. It is also capable of generating and transmitting alerts, in case the data collected by any of
the sensors exceeds the threshold value. A web-based application with a graphical user interface is used to view
and download the data.
One of the primary motivations behind development of the sensor node is high-resolution monitoring of
strain. Each wireless sensor node will be able to measure strain from 16 different locations on a structure, by
multiplexing these gauges to the same signal conditioning circuit, which drastically reduces the number of nodes
required for monitoring an area. Data collection can take place at regular intervals or when triggered by events
of interest.
We discuss the hardware and software implementation details in Section 2. The tests performed on the system
are discussed in Section 3.2.
Figure 1. CC2480 Daughterboard for Zigbee connection
2. RELATED WORK
SHM is a broad field and several existing projects use wireless communication to facilitate coordination of devices
for more effective measurement of a structure, whether the goal is improved spatial resolution, network resilience,
or advanced in-situ analysis. These systems can be broadly divided into two types: those that use commercial
motes as hardware and those that use custom hardware for implementation. The majority of these systems6–14
use commercial wireless motes. These motes are great for quick prototyping and development but, they are
unsuitable for long term installation in civil structures due to high power consumption and inadequate power
supplies. Another drawback of these devices is the lack of I/O ports which reduces their expandability. They
also use base stations and laptops for aggregating their data which rely on mains power supply. These drawbacks
severely limit the sensing and long term installation capabilities in the remote areas where a power supply is not
available or unreliable.
Several projects described in the literature15–18 utilize hardware specifically developed for SHM. The authors
of these studies have summarily demonstrated the effectiveness of wireless sensor networks for SHM, but the
systems are not described as being designed for extended operations in the field. Power consumption is still high,
but not as high as systems using motes. With the exception of 17 which is reported to yield a battery life of five
years, none of the aforementioned systems are designed for continuous operation of more than a year.
Qu Tiezhu et. al. developed a CC2480 based control and monitoring solutions.19 They provide detailed speci-
fications on hardware and system software design for communicating with CC2480 using low power MSP430F149
microcontroller from TI. Their experimental and test results show that the modules are reliable, low-power, self-
forming, auto-routing and self-healing.
Zigbee has several benefits over other choice for communication with nodes. It has a fast wake up time, low
power consumption, low memory requirements.20,21 Bluetooth and Wi-Fi, which work in the same frequency
range as Zigbee, offer higher transmission rates at the cost of higher power consumption.22, 23
3. IMPLEMENTATION DETAILS
3.1 Hardware Implementation
3.1.1 Base station (Coordinator) Hardware
The implementation of the base station has been discussed in our previous publications.1, 24–26 Short range
wireless capability using TI’s ez430-rf2480 has been presented in the publications.3, 5, 27 In this paper we mainly
talk developing the sensor nodes and integrating the custom made CC2480 Daughterboard to the SmartBrick
platform. Initially we tried interfacing the Daughterboard to SmartBrick 2.1A board, due to the complex
SPI bus the implementation was found to be very complex and was discontinued. We choose to integrate the
daughterboard with Flood Frog 2.0 A board1, 25, 28 which had a less complicated SPI bus. After some attempts
we were able to communicate to the CC2480 daughterboard successfully. The dsPIC is configured as master and
CC2480 acts as a slave. Figure 3 shows the Zigbee capable SmartBrick connected to the CC2480 Daughterboard.
The SmartBrick acts as a coordinator in the network. All the other devices like routers and end devices connect
to the coordinator. The hardware architecture of the Zigbee enabled SmartBrick is shown in Figure 3.1.1. The
Daughterboard is interfaced to the SmartBrick board via SPI.
Figure 2. Architecture of the SmartBrick network
Figure 3. Flood Frog 2.0A board connected with the Zigbee Daughterboard.
Figure 4. Overall architecture of the Zigbee-capable SmartBrick
3.1.2 Sensor Node(End Device) Hardware
The Sensor Node has a similar architecture like the SmartBrick the only main difference being it does not have
a GSM modem for long range communication capability. The sensor node is expandable through the expansion
headers and can be interfaced to multiple sensors. One main capability of it is the provision to multiplex a single
signal conditioning circuit to multiple strain gauges for high resolution strain monitoring. The architecture
of the SmartBrick is shown in Figure 3.1.2 below. The initial prototype was built with bread boarding the
DSPIC30F6014A breakout board with the daughterboard. It was extensively tested for self forming and self
healing.
Figure 5. Overall architecture of the sensor node
3.2 SOFTWARE IMPLEMENTATION
In this section we discuss the procedure to set up of the Zigbee network and also discuss the software state
machine of the Sensor node. The flow chart given in the figure 3.2 shows the procedure to startup and configure
the CC2480 for connecting it to the Zigbee network. Once the DSPIC boots and initializes all the hardware it
resets the CC2480 and waits for successful reset indication from the CC2480. After successfully resetting it the
dsPIC sends the configuration parameters to the chip and then verifies it for any errors. After these procedures
CC2480 will be ready for Zigbee connection.
The software in the sensor node has been implemented as a state machine. In order to conserve power the
sensor node spends most of the time in the sleep state. It then wakes up by the interrupts generated by Timer1.
Then it checks for the data present in CC2480. If there is any data to be sent it sends it to processing else it
checks for any task scheduled. In case there is any task scheduled it executes the task and then goes to sleep
again. It also checks for acknowledgement while sending the data and retries if the sending fails. The state
machine diagram is given in Figure 3.2.
4. TESTING AND EVALUATION
Laboratory testing of the sensor nodes were carried out. Tests were mainly carried out for self forming and self
healing. The following conditions were performed on the networked devices and the network was found to be
self forming and self healing. The test conditions simulated power and location disturbances.
4.1 Tests on the network
4.1.1 End device Started with coordinator already on
This is the scenario typical to the network operation. In this setup the coordinator is powered on in working
state and then the end device is either reset, powered or wakes up and switches on the radio. The end device
issues a request to join the network to the coordinator. The coordinator accepts the request and the network
was established successfully.
4.1.2 End device started without coordinator
When the End device is started it scans all the available channels for coordinator. It goes to sleep and wakes up
periodically and checks for available network.
Figure 6. Starting, configuring, and setting up the Zigbee network using the CC2480
Figure 7. Sensor node state machine
4.1.3 Start with coordinator already on and resetting the power source of the coordinator
This setup simulates the disturbance of the power supply to the coordinator. This was simulated by disconnecting
and reconnecting the power supply to the coordinator. When there is no coordinator the end device broadcasts
an orphan notification. The coordinator on encountering this will send a request to join its network.
4.1.4 Started with coordinator off and on later
In this setup the end device is first switched on and it scans periodically for a coordinator. When the coordinator
arrives it broadcasts a beacon and this initiates the end device to send a joining request. The coordinator accepts
it and the network formed successfully.
4.2 Power consumption
Initially the sensor nodes were operated on 5 V with divider resistors connecting the daughterboard. The power
consumption was 50maA which is much higher than the evaluation modules during transmit and receive cycles.
Then changes were made in the circuitry to make it operate at 3.3 V which significantly brought down the power
consumption. The power consumption was found to be almost equal to the evaluation modules27 with 35 mA
during receive and transmit cycles.
4.3 Antenna Range and Tuning
The range with the custom developed daughterboard was found to significantly lower than company manufactured
evaluation modules. The effective range inside laboratory conditions was found to be 10 feet without obstructions.
The range can be improved by improving the antenna design and tuning the antenna. Tuning the antenna is a
costly affair which requires high end equipment.
5. CONCLUSION AND FUTURE WORK.
The Daughterboard was successfully integrated with the SmartBrick and the network was successfully established
with the sensor nodes. The SmartBrick network was tested for self forming and healing and the network was
found to be resilient. However the power consumption was found to be almost equal to the TI ez430-rf2480
evaluation modules which opens the scope for more improvement. The range of the modules were found to be
low due to the lack of antenna tuning which requires sophisticated equipment.
Development efforts are undergoing to replace the Texas Instrument CC2480 Zigbee transceiver module with
the more capable MRF24J40MA from Microchip. This will result in lower power consumption and a longer
communication range. Another important reason for replacing is its ease of programming, which decreases
the likelihood of errors and facilitates maintenance and troubleshooting. Laboratory and field tests, including
underwater tests are scheduled for validation of reliable operation and performance of the system.
The SmartBrick project has evolved a lot since its inception. It has more sensors, longer battery life and
more resilient network. We will continue to refine the design so that the number of applications and potential
uses increase.
REFERENCES
[1] Bastianini, F., Plessi, V., Sedigh Sarvestani, S., Galati, N., and Nanni, A., “A low-cost wireless system for
real-time structural health monitoring,” in [Proc. of the 6th Int’l. Workshop on Structural Health Monitoring
(IWSHM ’07)], (2007).
[2] Harms, T., Bastianini, F., and Sedigh, S., “Recent enhancements to the SmartBrick structural health
monitoring platform,” in [Proceedings of 12th IEEE International Conference on Intelligent Transportation
Systems, St], 358–363 (Oct. 2009).
[3] Harms, T., Shah, P., Sedigh, S., Bourque, Z., and Bastianini, F., “Zigbee-enabled structural health monitor-
ing with the SmartBrick network,” in [Proceedings of the 7th International Workshop on Structural Health
Monitoring, Stanford, CA], (Sept. 2009).
[4] Banks, B., Harms, T., Bastianini, F., and Sedigh, S., “A low-cost wireless system for autonomous generation
of road safety alerts,” in [Proc. of the 16th Int’l SPIE Symp. on Smart Structures and Materials and
Nondestructive Evaluation and Health Monitoring], (March 2009).
[5] Harms, T., Banks, B., Bastianini, F., and Sarvestani, S. S., “Enhancement of autonomous real-time SHM
with short-range wireless communication,” in [the 4th European Workshop on Structural Health Monitoring
(EWSHM 2008)], (July 2008).
[6] Mechitov, K., Kim, W., Agha, G., and Nagayama, T., “High-frequency distributed sensing for structure
monitoring,” in [Trans. of the Soc. of Instrument and Control Engineers (SICE) ], E-S-1(1), 109–114 (2006).
[7] Olund, J. K., Cardini, A. J., Troiano, G. P. J., Liu, C., Feldblum, E., Attilio, P. D., and DeWolf, J. T.,
“Development and implementation of a solar powered wireless monitoring system on a truss bridge in
connecticut,” in [6th Int’l Workshop on Structural Health Monitoring ], 1174–1181 (September 2007).
[8] Kim, S., Pakzad, S., Culler, D., Demmel, J., Fenves, G., Glaser, S. D., and Turon, M., “Health monitoring
of civil infrastructures using wireless sensor networks,” Information Processing in Sensor Networks, 2007.
IPSN 2007. 6th International Symp. , 254–263 (April 2007).
[9] Arms, S. W., Townsend, C. P., Galbreath, J. H., and Newhard, A. T., “Wireless strain sensing networks,”
in [Proc. of the European Workshop on Structural Health Monitoring], (July 2004).
[10] Glaser, S. D., “Some real-world applications of wireless sensor nodes,” in [Proc. of the Int’l. SPIE Symp.
on Smart Structures and Materials], 344–355 (March 2004).
[11] Joshi, H., Burlacu, B., Prabhu, M., Yoon, S., Sichitiu, M., Dutta, R., and Rizkalla, S., “Wireless structural
health monitoring system design, implementation and validation,” Research Report RD-06-02, Constructed
Facilities Lab (CFL), North Carolina State University (NCSU), Raleigh, North Carolina (January 2006).
[12] Kling, R. M., “Intel mote: an enhanced sensor network node,” in [Proc. of Int’l. Workshop on Advanced
Sensors, Structural Health Monitoring, and Smart Structures], (2003).
[13] Deix, S. and Ralbovsky, M., “Intelligent sensor networks-the future of SHM,” in [Proc. of the 6h Int’l.
Workshop on Structural Health Monitoring ], 1239–1244 (September 2007).
[14] Choi, H., Choi, S., and Cha, H., “Structural health monitoring system based on strain gauge enabled wireless
sensor nodes,” in [Proc. of the 5th Int’l. Conf. on Networked Sensing Systems (INSS ’08) ], 211–214 (June
2008).
[15] Wang, Y., Lynch, J. P., and Law, K. H., “A wireless structural health monitoring system with multi-
threaded sensing devices: Design and validation,” Structure and Infrastructure Engineering - Maintenance,
Management and Life-Cycle Design & Performance 3(2), 103–120 (2007).
[16] Masquelier, M. P., Ball, S., Allen, D. W., Farrar, C. R., and Park, G., “A wireless data acquisition network
designed for structural health monitoring software algorithm and active sensing,” in [Proc. of the 4th Int’l.
Conf. on Information Processing in Sensor Networks (IPSN ’05)], (2005).
[17] Galbreath, J. H., Townsend, C. P., Mundell, S. W., Hamel, M. J., Esser, B., Hutson, D., and Arms, S. W.,
“Civil structure strain monitoring with power-efficient, high-speed wireless sensor networks,” in [Proc. of
the 4th Int’l. Workshop on Structural Health Monitoring ], 1215–1222 (September 2003).
[18] Uhl, T., Hanc, A., Peeters, B., Moya, E., and Van Der Auweraer, H., “Bridge monitoring system using
wireless sensor network-hardware solution and preliminary tests,” in [Proc. of the 6th Int’l. Workshop on
Structural Health Monitoring ], 1231–1238 (September 2007).
[19] Qu, T., Chen, X., Zhang, L., Xia, C., and Dai, S., “Zigbee wirelessly network module design based on
CC2480,” International Conference on Measuring Technology and Mechatronics Automatio 3, 1103–1106
(2010).
[20] Jiang, X.-D., Tang, Y.-L., and Lei, Y., “Wireless sensor networks in structural health monitoring based on
zigbee technology,” in [Proceedings of the 3rd international conference on Anti-Counterfeiting, security, and
identification in communication], ASID’09, 449–452, IEEE Press, Piscataway, NJ, USA (2009).
[21] Morello, R., De Capua, C., and Meduri, A., “Remote monitoring of building structural integrity by a
smart wireless sensor network,” in [Instrumentation and Measurement Technology Conference (I2MTC),
2010 IEEE], 1150 –1154 (May 2010).
[22] Mehta, V. and Zarki, M. E., “A Bluetooth based sensor network for civil infrastructure health monitoring,”
in [Proc. of the European Workshop on Wireless Sensor Networks], (2004).
[23] Park, C., Xie, Q., Chou, P., and Shinozuka, M., “Duranode: wireless networked sensor for structural health
monitoring,” in [Sensors, 2005 IEEE], 4 pp. (November 2005).
[24] Plessi, V., Bastianini, F., and Sedigh Sarvestani, S., “An autonomous and adaptable wireless device for
flood monitoring (short paper),” in [Proc. of the 30th Annual Int’l. Computer Software and Applications
Conference (COMPSAC ’06) ], 378–379 (2006).
[25] Plessi, V., Bastianini, F., and Sedigh Sarvestani, S., “A wireless embedded system for real-time environ-
mental and structural monitoring,” in [Proc. of the 5th IFIP Workshop on Software Technologies for Future
Embedded & Ubiquitous Systems (SEUS ’07)], 456–465 (May 2007).
[26] Harms, T., Bastianini, F., and Sedigh Sarvestani, S., “An embedded wireless system for remote monitoring
of bridges,” in [Proceedings of the 15th Int’l SPIE Symp. Smart Structures and Materials and Nondestructive
Evaluation and Health Monitoring], (March 2008).
[27] Harms, T., Bastianini, F., and Sedigh, S., “Design and testing of a low-power wireless sensor network for
structural health monitoring of bridges,” in [Proc. of the 16th Int’l SPIE Symp. on Smart Structures and
Materials and Nondestructive Evaluation and Health Monitoring], (March 2009).
[28] Plessi, V., Bastianini, F., Sedigh Sarvestani, S., Galati, N., and Nanni, A., “An autonomous networked
wireless device for structural health monitoring,” in [The 3rd Int’l Conf. Structural Health Monitoring of
Intelligent Infrastructure (SHMII-3)], (November 2007).