Searching the diffusive source in an unknown obstructed environment by cognitive strategies with forbidden areas

Yong Zhao,Bin Chen,Zhengqiu Zhu,Feiran Chen,Yiduo Wang,Yatai Ji

DOI: https://doi.org/10.1016/j.buildenv.2020.107349

IF: 7.093

2020-12-01

Building and Environment

Abstract:<p>Using sensing robots (searchers) to search for the diffusive source in indoor environments is a common and vital task in human society. Such environments, however, are always obstructed by some obstacles, and sometimes the prior information of obstacles is unknown. To solve the source searching problem in the random obstructed environment, this paper formalizes a general model (i.e. the structured map), in which random obstacles are denoted by obstructed cells. Then, we employ cognitive strategies to guide the searcher to perform the search, which are only employed in the environment without obstacles before. To prevent the searcher from being trapped by obstacles, we devise a passive escaping mechanism that marks some explored areas as forbidden areas. Once an area is marked, it cannot be reached by the searcher anymore. Two up-to-date cognitive strategies, i.e. Infotaxis and Entrotaxis are combined with the proposed mechanism to form the hybrid strategies: IWFA (Infotaxis with forbidden areas) and EWFA (Entrotaxis with forbidden areas). Each of them has four variants. The experimental results reveal the superiority of the hybrid strategies: IWFA-II and EWFA-II (two variants) achieve an improvement by a maximum of 28% and 27% in success rate, respectively, compared to their original strategies. The effectiveness of the hybrid strategies is also verified in the obstructed environment generated by the computational fluid dynamics (CFD) software, FLUENT, which considers the effect of obstacles on the diffusion of leaked substances. These findings show important practicality in our proposed hybrid strategies for actual search missions.</p>

engineering, environmental,construction & building technology, civil