A Coordinated Wilderness Search and Rescue Technique Inspired by Bacterial Foraging Behavior

In recent years, multi-agent path planning is a topic of interest to robotics. This paper studies the problem of multirobot path planning to search for lost people in a natural environment; known as WiSAR (Wilderness Search and Rescue). WiSAR is an optimization problem which requires: i) a fast and effective coverage of the area, and ii) to minimize the time, and distance traveled for finding the target. Heuristic algorithms find approximate solutions at an acceptable complexity both in time and space. This paper presents new adaptations to the Bacterial Foraging Optimization Algorithm (BFOA) that allow: 1) Semi-guided navigation and 2) Taking decisions based on a probabilistic model that represent the dynamic environment. The algorithm was simulated in the Matlab platform by using a 2-D plane. Experimental results show that the proposed technique reduces 60.7% of the time and 70.6 % of the distance traveled (in average) with respect to the Grid Search technique when five agents are used in a search of a child in outdoors. Results show that this reduction can be much greater when fewer agents are deployed or the person moves at a higher rate.