Name
Abstract | ||
|---|---|---|
At present, odor sensors and anemometers are simply used by most chemical plume tracing algorithms to measure the concentration and wind speed/direction, respectively. To make full use of the information of concentration and wind for chemical plume tracing, the concept of virtual odor-source-probability sensor (VOSPS) is put forward. The VOSPS adopts the data of odor sensor and anemometer as input and outputs local odor source probability distribution through Bayesian estimation and fuzzy inference. The global odor source probability distribution is constructed by the method of log odds-ratio. The expectation of odor source probability distribution is used to express the fitness function of the PSO algorithm which is used to coordinate the multi-robot system. To validate the tracing strategy, the plume model corresponding to the actual boundary condition of an indoor ventilated environment is set up. Considering the slow response and recovery time of most real odor sensors, a second-order sensor model is built. Simulation results demonstrate that, compared with the existing PSO based plume tracing algorithms, the proposed algorithm has quick search speed and high success rate. |
Year | DOI | Venue |
|---|---|---|
2009 | 10.1109/FSKD.2009.533 | FSKD (1) |
Keywords | Field | DocType |
global odor source probability distribution,fuzzy set theory,outputs local odor source,log odds-ratio,wind speed-direction,multi-robot,bayesian inference,chemical plume tracing,bayes methods,fuzzy inference,inference mechanisms,particle swarm optimisation,virtual odor-source-probability sensor,real odor sensor,bayesian estimation,multi-robot systems,chemical plume,odor source probability distribution,second-order sensor model,anemometers,odor source search,plume model,probability distribution,probability-fitness-function based particle swarm optimization (p-pso),electronic noses,global odor source probability,multirobot based chemical plume tracing,ventilated indoor environments,odor sensor,atmospheric modeling,odd ratio,robot kinematics,fitness function,chemicals,log odds ratio,wind speed,bayesian methods,boundary condition,second order | Wind speed,Bayesian inference,Computer science,Odor,Anemometer,Fitness function,Probability distribution,Artificial intelligence,Machine learning,Tracing,Bayesian probability | Conference |
Volume | ISBN | Citations |
1 | 978-0-7695-3735-1 | 1 |
PageRank | References | Authors |
0.37 | 5 | 5 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Fei Li | 1 | 69 | 16.39 |
| Qing-Hao Meng | 2 | 144 | 18.73 |
| Ji-Gong Li | 3 | 60 | 4.85 |
| Shuang Bai | 4 | 45 | 8.01 |
| Ming Zeng | 5 | 29 | 4.97 |