Adaptive Path Planning for Effective Information Collection

We consider the problem of information collection from an environment by a multi-robot system, where the locations to sample information from are known with some amount of uncertainty by the robots. Different robots are equipped with different types of sensors and the quality of information collected can be improved if information is sampled from the environment with different sensors, wherever possible. The problem facing the robots is to determine their paths through information sampling locations, so that the overlap between paths of robots with identical sensors is reduced, while the robots with complementary sensors are sllowed to have overlapping paths. To address this problem, we describe a distributed path planning algorithm that comprises of two steps path merging and path decoupling so that the robots are able to collect more information from the environment. We have shown analytically that our proposed algorithm has polynomial time complexity and our experimental results, in simulation, show that by dynamically adapting paths, the robots are able to collect higher information than by following initially generated individual paths.