A decentralized collision-free search algorithm in 3D area for multiple robots

摘要

This paper proposed a decentralized random algorithm for a group of robots to search targets in 3D areas through a grid without collisions. Both the area and obstacles are unknown a priori. Considering the search for MH370 with Wi-Fi communication, the cube grid pattern is chosen to ensure that the least vertices are used. Robots start from a corner of the grid and move straightly between neighbor vertices. They choose the next steps in sequence based on their explored area and local information. This paper considers the physical characters and constraints of real robots. A rigorous mathematical proof of the convergence with probability 1 is given and the algorithm is compared with others in simulations to demonstrate the effectiveness.