Multiple-Robot Motion Planning in an Unknown and Uncertain Environment

When a Robot finds itself in an unknown and uncertain environment it needs to retrieve information from its surroundings so it can perform a task in the appropriate way. A task may demand far much more capabilities than the ones a robot can perform by itself, a Multiple-Robot System can accomplish lots of tasks, however, working with them implies more uncertainty while executing actions. The desired task in this case is to move a huge object from an initial position to a final one. There are two main solutions to this problem, the first one is to employ a big and expensive robot that is capable of moving the object. The second choice is to employ a Multiple-Robot System composed by several low-cost robots. The second one was chosen because of its economic advantages, although it needs a more complex way of communication between the agents. This paper presents a novel form of deploying a coupled Multiple-Robot System based in a Master-Slave configuration in an Unknown and Uncertain Environment, where initially only the goal is known, no a priori environment information is given. This way, the Multiple-Robot System will need to retrieve data from its sensors so it can spot itself and build a map while moving through the space so Classic Path Planning techniques can be used. The employed algorithms were Approximate Cell Decomposition, A star, Kalman Filter and Particle Filter; they all run in MATLAB 2012a. The used robots were Parallax Boe-Bots.