Role assignment for multi-robot system in dynamic and uncertain environments

The growing interest in developing colonies of robots engaged in complex tasks has caused an increasing interest in coordination approaches. In particular coalitions can perform tasks better than single robots and the quality of the execution of a specific task depends both on the individual capabilities of each robot and on how they can be combined together. We propose a hybrid framework which use the Max-sum algorithm applied over a factor graph system representation to guarantee robot coalitions to positively handle both task allocation and collision avoidance problems