Immune Based Task Allocation Method for Multi-Robot System

摘要

To efficiently accomplish unknown tasks in the multi-robot multi-task system, an efficient task allocation method and autonomous cooperation among robots are required. This paper fully takes advantage of the interactions among antibodies and antigen stimulus of immune system to solve the problem. Firstly, a new artificial immune network (AIN) model is proposed for the multi-robot system based on the principles of the biological immune system. Based on AIN model, the multi-robot task allocation algorithm is designed by utilizing the interactions among the antibodies, and the event-triggered task reallocation is adopted to realize the dynamic task allocation. Then the dynamic task allocation method is developed and extended by integrating the cooperative idea into the antigen stimulus. By the self-reinforcement learning of the antigen stimulus, the autonomous cooperation among robots is realized and deadlock situation is avoided. Based on the committed/opportunistic attribute of the robots, three different methods are proposed to implement the autonomous cooperation among robots. In the simulation and discussion, the immune based allocation method is further analyzed from the communication and computation aspects and is verified. And in the experiment of autonomous emergency handling, the three integration methods are validated, studied and compared.