Robotics and Artificial Intelligence

Multi-robot systems are increasingly employed in various fields such as search and rescue, space and underwater exploration, surveillance and target tracking and logistics. Distributed task allocation is one of the key topics in such multi-robot systems. This presentation will introduce our recent work on developing novel algorithms for distributed task allocation under different constraints from real applications, in particular in search and rescue scenario. The objectives of the problem are analyzed, including maximizing the number of assignments (people rescued) and minimizing the average start times of tasks with deadlines. Then its mathematical formulation is developed considering only the local communication for connected robots. A novel task swapping mechanism is proposed to exchange tasks between vehicles through local networked communication in order to optimize the performance of the whole team. This mechanism is further supported by the concept of task performance impact which is measured by the contribution to the local cost generated by a vehicle. Two performance impacts are considered, Removal Performance Impact (RPI) and Inclusion Performance Impact (IPI) of a task assignment. Different communication topologies for multi-robot systems have been considered to investigate the proposed distributed task allocation algorithms. Experimental results demonstrate that the proposed algorithms can achieve better performance in comparison with other algorithms in this area.