This project is sponsored by National Natural Science Fund (60574023) andQingdao Technology Division Technology Development Plan (05-1-jc-88). It doessome research on the problems of robot soccer's path planning based on improved GAIt is the first time to put the CONSENSUS algorithm into the multi-agent pathplanning by using the coordination control theory.Based on the research platform of Robocup, this paper does profound research inpath-planning problem, and tries to find effective schemes to solve the path-planningproblems in practice. Firstly, the architecture of soccer robot system and its keytechnologies are reviewed. By studying some typical path-planning methods, wepropose GA algorithm, CONSENSUS algorithm and their applications in soccer robotpath planning.In order to avoid getting into local optimum and improve the global optimizationability and speed, a Niche Genetic Algorithm based on the robot soccer's pathplanning algorithm is proposed. GA and Ant Group algorithms are combined toimprove the algorithm solution validity and local searching efficiency, and meanwhileavoid the premature stagnancy. Besides, an improved CONSENSUS based artificialpotential field algorithm is proposed to realize more' real-time algorithm and moreeffective navigation and obstacle-avoidance performance.In the study of multi-robot path-planning problem, every robot is treated as anode of a small world modal, by using a small-world network based CONSENSUSalgorithm, we solve the traditional problems of low convergence speed, huge agentscommunication amount and high system resource cost. Finally, dynamic link library(DLL) technology is used to build the algorithm modules of digital image processingalgorithms and some path-planning algorithms, which help to maintain and update thesystem in future.Practical erigineering tests show that the algorithm proposed in this paper canrealize path-planning more rapidly, accurately and effectively, under the preconditionsof guaranteeing real-time and reliable system operation. |