Soccer Robot Cooperative Control Research Based On Immune Mechanism

Robot cooperative control is one of the core issues of multi-robot systems research, which relates to many aspects such as the robot architecture, perception, communication, algorithms, collaborative mechanisms and so on. As a typical multi-robot system, the soccer robot system provides a good experimental platform for the intelligence theory studies and a variety of integrated application.The subject is based on the soccer robot system research platform of Xi’an University of Science and Technology. Aiming at the dynamic environment of soccer robot competition, the robot tracking accuracy is increased by improving the target tracking algorithms, and the autonomous collaboration of the soccer robot is advanced through using immune mechanism and artificial immune algorithm, to solve the congestion and resource waste problems which often happens in the actual course of the game, and improve the intelligence of the system finally.The soccer robot system, hardware structure composition and working principle are introduced and analyzed in the paper. Focusing on the problems of mobile robot target tracking, the gaussian sum particle filter(GSPF) improved algorithm is provided for soccer robot target tracking, parallel extended kalman filtering is replaced by parallel gaussian particle filter of gaussian sum filtering in the GSPF, and full consideration is given to the influence of the current observed quantity to the state distribution, which improves the target tracking precision of system. The simulation results show that this algorithm can effectively apply to soccer robot target tracking in the dynamic and complex environment.According to the robot architecture, the basis of the collaborative behavior for robot soccer, considering the incentive and the inhibition between antigen and different antibody, and the immune mechanism such as the communication between different immune cells with each other, the football robot individual hybrid architecture and group hierarchical structure are build in the paper. The stimulation of environment, the sharing of information between robots, and the incentive and inhibition between different soccer robots, which have an affect on the behaviors of Individual robot and cooperation strategy between the overall system. The architecture makes sure the soccer robot system with strong real-time, dynamic adaptability and robustness.The Soccer Robot Immune(SRI) system is built which relate to the Interaction between the different antibody created by different roles and the outside world antigen. The soccer robot cooperation control method is put forward based on immune algorithm. By achieving the antigen and antibody concentration Information of other robots, combing the antigenic information of the local environment, the incentive value of antigen can be calculated and the corresponding antibody can be produced to choose the correct robot to implement the corresponding strategy behavior. When the robot can not complete the task alone, it can send requests to other robot cooperation, and make the response of the robot as a reward signal of antigen stimulation value to complete the antigen self-reinforcement adjustment, which can finish the task as soon as possible.The simulation and the actual football environment experiment show that the algorithm can be used in the cooperative control of soccer robot in dynamic complex adversarial game environment, improve the conflict chaos of multi-robot, enhance the effectiveness of soccer robot strategy, increase the number of cooperation attack during the game and increase the intelligent collaboration of the soccer robot system, success in achieving the design goal of paper. Because the improvements in both algorithm and structure of the soccer robot system, our university’s soccer robot team in the national competition over the years has made excellent achievements.