Distributed Strategies for Local Minima Escape in Motion Planning for Mobile Networks

This paper studies the problem of controlling networked mobile agents while maintaining connectedness among them. Our previous work in [13] and other existing distributed methods use potential field based techniques and hence suffer from local minimum problems. In this paper we propose a preliminary categorization of different types of local minima that can arise and distributed strategies to deal with these local minima, based on our previous backbone based connectivity control (BBCC) framework in [13], where a communication backbone was used to maintain system connectivity. The local minima in our problem can arise either due to obstacles or due to connectivity constraints, or a combination of the two. Our categorization and the corresponding strategies to escape local minima are an initial attempt to deal with these issues in a systematic manner. As backbone is an effective and efficient representation of the formation topology, it provides a good leverage to exploit all members in the mobile network to gain knowledge of the environment and make decisions, and our simulations show that these backbone based strategies are very effective in escaping these local minima.