Multi-robot based implementation for a sample gathering problem

For a class of multi-robot task allocation problems, a theoretical solution and its real time implementation are discussed in this work. The basic problem requires a team of mobile robots to gather randomly spread samples from an environment and deposit them to a storage facility. The proposed solution considers the environment as being abstracted to a weighted graph and randomly distributed samples in the nodes of this graph. Using a setup consisting of a platform on which mobile robots can move, a computer that controls the robots and an overhead camera that closes the loop between these devices, the real time implementation of the theoretical solution is tested. Experimental results of various sample distributions are conducted and the performance of the entire method is discussed.