Figure - available from: International Journal of Advanced Robotic Systems
This content is subject to copyright.
Source publication
Most of the reported three-dimensional chemical plume tracing methods use stereo sensing method to determine the next tracing step direction. For example, multiple sensors are used for detection in the left, right, up and down directions. Left and right detections are feasible for stereo sniffing; unfortunately, the same approach is infeasible for...
Citations
... An oft-cited concern for use of multicopter UASs in making air quality measurements is the potential effect of the rotor downwash on the sensor readings and sampling system [4]. Rotor downwash has been visually demonstrated by Crazzolara et al. [6] and has been the subject of numerous computational studies [7][8][9][10][11], as cited in Burgués and Marco [4]. While air disturbances are generally considered to be limited above the rotors (~50 cm above the UAS, [12]), alongside the UAS (see references in Burgués and Marco [4]), and especially in front of a moving UAS, they can extend multiple UAS diameters below the rotors [13]. ...
Laboratory and field tests examined the potential for unmanned aircraft system (UAS) rotor wash effects on gas and particle measurements from a biomass combustion source. Tests compared simultaneous placement of two sets of CO and CO2 gas sensors and PM2.5 instruments on a UAS body and on a vertical or horizontal extension arm beyond the rotors. For 1 Hz temporal concentration comparisons, correlations of body versus arm placement for the PM2.5 particle sensors yielded R2 = 0.85, and for both gas sensor pairs, exceeded an R2 of 0.90. Increasing the timestep to 10 s average concentrations throughout the burns improved the R2 value for the PM2.5 to 0.95 from 0.85. Finally, comparison of the whole-test average concentrations further increased the correlations between body- and arm-mounted sensors, exceeding an R2 of 0.98 for both gases and particle measurements. Evaluation of PM2.5 emission factors with single-factor ANOVA analyses showed no significant differences between the values derived from the arm, either vertical or horizontal, and those from the body. These results suggest that rotor wash effects on body- and arm-mounted sensors are minimal in scenarios where short-duration, time-averaged concentrations are used to calculate emission factors and whole-area flux values.
... The sensitivities of the PAN/PPy nanofiberscoated chemical sensor against organic vapors are obtained for dichloromethane, chloroform, carbon tetrachloride, benzene, toluene and m-xylene, as 6.17, 5.75, 5.59, 4.96, 4.91, and 4.71 (Hz ppm −1 ) × 10 −4 , respectively. In this study, the rise time was defined as the time taken by the sensor to achieve 90% of the maximum frequency shift (Δf max ) in the case of gas adsorption [45,46]. When surface adsorption effect took place, a sudden increase can be seen in Δf. ...
In this study, electrospun polyacrylonitrile (PAN)/polypyrrole (PPy) nanofibers (NFs) coated quartz crystal microbalance (QCM) were investigated for their sensing characteristics against six different volatile organic compounds (VOCs): chloroform, dichloromethane, carbon tetrachloride, benzene, toluene and xylene. SEM, TEM, FT-IR and TGA analysis were carried out for the characterization of PAN/PPy nanofibers and characterization results of PAN/PPy NFs showed that these nanofibers were morphologically well-arranged and straightforward with a cylindrical shape with the average fiber diameter of 253.17 ± 27 nm. Among all the gas measurement tests, dichloromethane displayed the highest response values for PAN/PPy coated QCM sensors. When the reproducibility of kinetic studies for PAN/PPy NFs coated QCM sensors were examined, the most repetitive results were obtained by this QCM sensor during dichloromethane investigation and the diffusion coefficients of VOCs for the first and second regions increased with the order of xylene < toluene < benzene < carbontetrachloride < chloroform < dichloromethane. The sensitivities of the PAN/PPy nanofibers-coated QCM sensor against organic vapors are determined between 4.71 and 6.17 (Hz ppm⁻¹) × 10–4. As a result, PAN/PPy nanofibers exhibited high sensitivity and selectivity for VOCs sensor applications, especially for dichloromethane.
... This method can respond quickly to hazardous material leakage and deploy urgently, localizing the plume source efficiently and accurately. Recently, researchers have proposed various plume source localization methods based on active olfaction, mainly including the concentration gradient-based algorithms [6][7][8][9], bionicbased algorithms [10][11][12], and multi-robot-based algorithms [13][14][15][16][17][18][19][20][21][22][23]. Compared with single robot-based active olfaction algorithms, multi-robot-based algorithms show high flexibility and adaptability in the source localization process by sharing chemical plume cues among the robots, having less search time, wider search area, and higher search efficiency. ...
... At this point, the Gaussian parameter of the type label belonging to m , i can be represented as θ i = {μ c i , c i }. Since the mixing proportions π cannot be observed directly, the setting θ i is introduced by G. Thus, the following DPGMM can be obtained: (10) In the DPGMM, the base distribution G 0 is a conjugate Normal-Inverse Wishart (NIW) prior defined as follows: ...
Accidents of leaks and emissions of flammable, explosive, and toxic substances severely threaten people’s health and public safety. Traditional heuristic algorithms for source localization that utilize wind information to guide the robots to search for plumes in the airflow environment, substantial potential plume information is not fully considered, reducing the success rate and precision of source localization. To solve this problem, in this study, we propose a novel strategy for plume source localization using information gain. This strategy is inspired by the navigation and foraging behaviors of the salps in the ocean, and the robots use a mutation random salp swarm algorithm to track the plume. This strategy utilizes the Dirichlet process Gaussian mixture model to color the plume information and achieve a dynamic update of the plume information distribution. Therefore, this developed strategy provides potential source location clues for the robots, dynamically adjusting the robots’ search behaviors and enhancing the source-seeking success rate and efficiency. This strategy is verified by experimental validations. The test results show that with two methods that use wind information, namely, the improved whale optimization algorithm and wind utilization II particle swarm optimization, and two approaches independent of wind information, namely, the fixed-step fruit fly optimization algorithm and improved particle swarm optimization, the proposed strategy improves the source localization success rate by 5% to 33% and accelerates the efficiency by 0.1 to 2.6 times in the obstacle-free scenario, and improves the source localization success rate by 4% to 100% and accelerates the efficiency by 0.4 to 2.2 times in the obstacle environment.
... On flying robots, Gao et al. [34] presented a multirotor UAS-based application to find an odor source in an indoor environment. In Eu and Yap's work [35], a quadrotor drone is controlled to vary its height to search odor plumes. Luo et al. [36] presented a flying odor compass to detect odor plumes in a three dimensional space. ...
... Hence, 3-D-CPT has typically been verified in ideal environments, and the system utilized must be expanded to enable 3-D-CPT in real-world environments with significant changes in the airflow direction. In recent years, drones have garnered attention as unmanned aerial vehicles (UAVs) that traverse three dimensions, and researchers have attempted to impart 3-D-CPT capabilities to drones [18], [19], [20], [21]. In this study, we employed a quadcopter comprising of four propellers. ...
... However, Luo et al. [22] utilized this phenomenon and proposed the "aero-olfactory effect," which attracts chemicals. In a previous study, a sensor arrangement that uses the aero-olfactory effect of a quadcopter was proposed, and it was reported that 3-D-CPT could be performed using the proposed sensor arrangement [18], [20], [21]. Luo et al. [18] implemented a conventional odor compass [13] on a drone and showed that it was effective for 3-D odor tracking. ...
... Ercolani and Martinoli [19] installed an odor sensor in the center of a quadcopter and attempted to identify an odor source in a 3-D space by switching between searching in the vertical and horizontal directions based on the odor sensor value. Eu and Yap [20] performed an experiment using an actual quadcopter with odor sensors placed under the propeller and verified the performance of the 3-D spiral casting and 3-D spiral-zigzag casting algorithms, which are extensions of the bio-inspired search algorithm. The proposed algorithms require many simultaneous height and horizontal movements during the odor search, which increases the risk of overlooking an area of odor source and significantly increases the required time to identify the odor source. ...
In this study, we designed and experimentally verified the placement of odor sensors and an algorithm using the aero-olfactory effect of a palm-sized quadcopter to solve the three-dimensional chemical plume tracking (3D-CPT) problem. Solving 3D-CPT is important in engineering as it helps perform rescue operations during disasters and identify sources of harmful substances. Moreover, the odor sensors must be properly located and a CPT algorithm be applied to improve the tracking performance of a chemical. However, studies regarding the use of quadcopters for solving the 3D-CPT problem are scarce, and the relationship between the odor sensor location and algorithm is debatable. Hence, we utilized particle image velocimetry, an airflow visualization technology, to evaluate the arrival direction of chemicals at different heights. The results showed that odor sensors must be placed on the upper and front surfaces of a quadcopter to monitor the chemicals three-dimensionally. Additionally, we designed a 3D surge-casting algorithm, which is an extension of the CPT strategy of a flying moth, that is, surge casting, to accommodate the proposed odor sensor placement. By conducting 3D-CPT experiments based on different heights of odor sources using the proposed system, we discovered that even in an environment with significant changes in the wind direction the CPT performance is better than that of the conventional 3D-CPT algorithm. Thus, 3D-CPT should be further improved to enable its application in unknown and cluttered environments. In this study, we improved the 3D-CPT performance of a palm-sized quadcopter by designing an appropriate sensor arrangement and algorithm balance.
... An interesting study is to evaluate both single algorithms in case of different configurations, sampling displacements and mixed approaches when multiple drones could define at the same time different information. For example, a spiral approach has been tested in single and multiple sampling [33,52,53], but some optimization could be considered for example by taking in account the change in wind direction. In a swarm displacement that cover a large area, spiral algorithms could be early stopped thanks to the information of the other team members. ...
The advancements in the field of robotics, specifically in the aerial robotics, combined with technological improvements of the capability of drones, have increased dramatically the use of these devices as a valuable tool in a wide range of applications. From civil to commercial and military area, the requirements in the emerging application for monitoring complex scenarios that are potentially dangerous for operators give rise to the need of a more powerful and sophisticated approach. This work aims at proposing the use of swarm drones to increase plume detection, tracking and source declaration for chemical releases. The several advantages which this technology may lead to this research and application fields are investigated, as well as the research and technological activities to be performed to make swarm drones efficient, reliable, and accurate.
... In order to improve performance by taking distributed synchronized measurements of the field while also improving resiliency under failure scenarios, multiple vehicle systems have been proposed. The use of multiple, spatially distributed vehicles overcomes the primary drawback to many single vehicle AN control approaches, which is the need to execute time-consuming spatial dithering maneuvers in order to obtain local field information upon which navigation decisions are made [9][10] [11] [12]. However, these multivehicle systems require increased system sophistication due to the additional complexity associated with group control. ...
... Single vehicle simulations assuming simplified vehicle kinematics or point mass models have demonstrated extrema seeking with strategies utilizing gradient following [36] [37], hybrid-sliding mode control [38], and forward and angular velocity regulation as the robot approaches the source [39]. Experimentally, [12] demonstrated source seeking of a dynamic gas plume with a single UAV under fuzzy-logic control-based spatial dithering, in a small indoor testbed (~9 m 2 ). ...
... Equations (12) and (13) specify the cluster translational velocity commands to implement this motion control strategy. In (12), the first term directs the cluster in a direction that is tangential to the desired surface given the cross product with the gradient vector ; the specific direction is dictated by the choice of the navigation reference vector ̂ and dorbit which is set to +1 and -1 for counterclockwise and clockwise travel relative to ̂, respectively. ...
Adaptive Navigation (AN) control strategies allow an agent to autonomously alter its trajectory based on realtime measurements of the environment. Compared to conventional navigation methods, these techniques can reduce required time and energy to explore scalar characteristics of unknown and dynamic regions of interest (e.g., temperature, concentration level). Multiple Uncrewed Aerial Vehicle (UAV) approaches to AN can improve performance by exploiting synchronized spatially-dispersed measurements to generate realtime information regarding the structure of the local scalar field, which is then used to inform navigation decisions. This article presents initial results of a comprehensive program to develop, verify, and experimentally implement mission-level AN capabilities in three-dimensional (3D) space using our unique multilayer control architecture for groups of vehicles. Using our flexible formation control system, we build upon our prior 2D AN work and provide new contributions to 3D scalar field AN by a) demonstrating a wide range of 3D AN capabilities using a unified, multilayer control architecture, b) extending multivehicle 2D AN control primitives to navigation in 3D scalar fields, and c) introducing state-based sequencing of these primitive AN functions to execute 3D mission-level capabilities such as isosurface mapping and plume following. We verify functionality using high-fidelity simulations of multicopter drone clusters, accounting for vehicle dynamics, outdoor wind gust disturbances, position sensor inaccuracy, and scalar field sensor noise. This paper presents the multilayer architecture for multivehicle formation control, the 3D AN control primitives, the sequencing approaches for specific mission-level capabilities, and simulation results that demonstrate these functions.
... The downwash disturbs the local air distribution around the drone (especially below it) and can have negative consequences for the utility of the on-board sensor data. The aerodynamic characteristics of the downwash generated by rotorcrafts have been simulated using Computer Fluid Dynamics technology (CFD) (Eu et al., 2014;Eu and Yap, 2018;Koziar et al., 2019;Kuantama et al., 2019;Luo et al., 2016;McKinney et al., 2019;Roldán et al., 2015;Sanchez-Cuevas et al., 2017), experimentally measured with anemometers (Prudden et al., 2016;Sjöholm et al., 2014;Wolf et al., 2017), or particle tracking velocimetry (PTV) systems (Shigaki et al., 2018;Shukla and Komerath, 2018), and visualized using smoke experiments (Hollenbeck et al., 2019b;Hutchinson et al., 2019;Kang et al., 2018;Luo et al., 2017;Neumann et al., 2012;Prudden et al., 2016;Smith et al., 2016). While airflow is typically negligible at 40-50 cm above the drone (Alvarado Fig. 2. Downwash of a hovering DJI Matrice 600 drone visualized using colored smoke (Crazzolara et al., 2019(Crazzolara et al., ). ...
... Top-and frontmounting ( Fig. 8i-k) is very popular for GSL applications (see Section 4.2), in which it is more important to rapidly detect the gas plume than to accurately quantify the gas concentration. For this purpose, pairs of replicate MOX sensors are often placed on the front of small quadrotors (Koval et al., 2017;Kuantama et al., 2019;Letheren et al., 2016;Shigaki et al., 2018;Takei et al., 2019) (Fig. 8k), or in more complex configurations such as three sensors placed in a triangle arrangement (Luo et al., 2017) or four sensors below the propellers (Eu and Yap, 2018). ...
Recent advances in miniaturization of chemical instrumentation and in low-cost small drones are catalyzing exponential growth in the use of such platforms for environmental chemical sensing applications. The versatility of chemically sensitive drones is reflected by their rapid adoption in scientific, industrial, and regulatory domains, such as in atmospheric research studies, industrial emission monitoring, and in enforcement of environmental regulations. As a result of this interdisciplinarity, progress to date has been reported across a broad spread of scientific and non-scientific databases, including scientific journals, press releases, company websites, and field reports. The aim of this paper is to assemble all of these pieces of information into a comprehensive, structured and updated review of the field of chemical sensing using small drones. We exhaustively review current and emerging applications of this technology, as well as sensing platforms and algorithms developed by research groups and companies for tasks such as gas concentration mapping, source localization, and flux estimation. We conclude with a discussion of the most pressing technological and regulatory limitations in current practice, and how these could be addressed by future research.
... Experimental observations have confirmed these effects under the actual flight conditions of a single copter-type propeller [38,39]. The results of CFD investigations have likewise been confirmed by experimental method measurements, indicating that BVI phenomena occur around single copter-type propellers, double coaxial copter-type propellers, multi-rotors, and quadtilt rotors [38][39][40][41][42]. Kok et al. suggested that the turbulence that occurs around the strong down flow under a rotor interferes with a chemical plume tracing (CPT) algorithm, obstructing its function in the detection of chemical agents [22,43]. For these reasons, the aerodynamics around the drone must be considered for quadrotor drone chemical detection. ...
... The PIV experiment was conducted with the assumption that the size of the sensor was sufficiently small to negligibly affect the overall air flow. The location of a sensor on a quadrotor drone is not a trivial issue and its effect requires aerodynamic analysis [43]. After sufficient filming of 300-ms movements of the tracer particles by the PIV camera, an analysis software (DynamicStudio (Dantec, Inc., Skovlunde, Denmark) which can measure and post-process analysis of the PIV data) was used to plot the average velocity vectors, which were used to roughly estimate the wind strength during the flight of a drone. ...
The determination of a suitable sensor location on quadrotor drones is a very important issue for chemical reconnaissance platforms because the magnitude and direction of air velocity is different for each location. In this study, we investigated a customized chemical reconnaissance system consisting of a quadrotor drone and a chip-sized chemical sensor for detecting dimethyl-methylphosphonate (DMMP; a Sarin simulant) and investigated the chemical detection properties with respect to the sensor position through indoor experiments and particle image velocimetry (PIV) analysis of the system. The PIV results revealed an area free of vortex–vortex interaction between the drone rotors, where there was distinctly stable and uniform chemical detection of DMMP. The proposed chemical reconnaissance system was found to be realistic for practical application.
... A 3D simulation framework making use of data from 3D CFD results was then developed by Awadalla, Lu, Tian and Dally [28] to create a more realistic environment for testing and training plume-tracing robots. Subsequent research concerning 3D scenarios can be found in the studies conducted by Soares, Marjovi, Giezendanner, Kodiyan, Aguiar, Pascoal and Martinoli [29]; Neumann, Kohlhoff, Hüllmann, Lilienthal and Kluge [30]; Eu and Yap [31]. ...
Bio-inspired chemical plume-tracing methods have been applied in robots to detect chemical emissions and to localise the plume sources in both indoor and outdoor environments. In the first part of this study, a comparison of performance of several widely used plume-tracing algorithms was conducted. A plume-tracing algorithm can be divided into three stages for analysis: plume sensing (PS), plume tracking (PT) and source localisation (SL).¹ These algorithms, which had been previously presented and tested in either simulation framework in 2D scenarios or experiments, were tested and compared in two different 3D scenarios in this study. In one scenario, a chemical source is located away from walls in a channel and in the other scenario, the chemical source is located near a wall. This is the first time that the performance of different plume-tracing algorithms in wall plumes has been tested and assessed and included in the literature. Sixteen different algorithms were tested and compared and the algorithm constituted by normal casting, surge anemotaxis and normal stepsize performed the best among all. In the second part of the study, this algorithm was further optimised by an ‘along-wall’ obstacle avoidance method and finally a novel algorithm, named vallumtaxis, was proposed and shown to achieve higher efficiency.
