Distributed Behavior Control Method for Indoor Multi-robot Collaborative Odor Source Localization

The paper proposes a distributed behavior control method for indoor multi-robot collaborative odor source localization, aiming to improve the efficiency of odor source search while avoiding collisions. This problem is crucial in scenarios such as fire rescue, drug inspection and so on. Multiple robots search for the plume in a spiral fashion. They share the information of the location with the highest detected odor concentration among them and guide robots that do not detect gas to move towards it, which can accelerate the plume discovery speed of the whole system. Multiple robots use particle filter to estimate the location of the odor source. They identify the odor source by setting particle convergence and distance convergence thresholds. A state machine is used to achieve effective odor source search by flexibly selecting the appropriate method according to the current environment and task requirements. Monte Carlo simulation results demonstrate that the proposed algorithm outperforms multi-robot infotaxis method 50 %.