On Modular Design And Motion Control Of Small Size Soccer Robot

Soccer robot is a kind of mobile robot which works under highly dynamic environment through a friendly competition, and it is a high-tech platform focusing on the problem of intelligent multi-agent cooperation and control. Research on soccer robot is benefit to promote as well as enhance the development and application of robotics, artificial intelligence and other related fields, which is meaningful to both academic and engineering application.The small size soccer robot is investigated in this thesis, the motion control and modular design of the small size soccer robot are studied. The main work and its achievements are summarized as follows:1. Numerical method of modular division is studied. A method based on the most optimal constrained value of the cohesion is presented, it converts the conventional modular division problem which is high-dimensional nonlinear combinatorial optimization problem into the one-dimensional module division problem. Improved quantum particle swarm algorithm applied to the optimization of the constrained value of the cohesion is given in this paper. With the improved quantum particle swarm algorithm, many modules of the small size soccer robot called ICrobot are obtained quickly and efficiently, and the modules which are low coupling and high cohesion can meet actual demand.2. The dribbling module of the small size soccer robot is designed. By the force analysis on the dribbling mechanism, the factors which impact the robot soccer’s stability of dribble are identified. The innovative design for mechanism of the dribbling module and related parameters are implemented. And the design result shows that, compared with the conventional mechanism, the new one has a better stability.3. The combination of the small size soccer robot’s motion module and its motion control are analyzed. The motion module combination regimen is given according to the functional requirements of the role of the soccer robots. The motion control strategy under non-fault state is put forward base on PID control. A fault diagnosis method based on probabilistic neural network for motion module is presented owing to the failures of the soccer robot’s motion module when in the game. At last, the soccer robot’s fault tolerant control strategy under fault state is proposed. The simulation results show that the proposed fault diagnosis method has higher accuracy rate overall, the motion control strategy considering the fault state has good real-time performance and accuracy.Kinematics and mechanical analysis in this thesis are completed by means of the software ADAMS, the processing and analysis of the data and simulation from related algorithms are completed by utilizing the software MATLAB.