论文部分内容阅读
以未知环境下多机器人学习为研究平台,因案例推理方法可存储以前的问题和解信息,用该方法的长期记忆特性可帮助粒子群优化算法更好地解决新的问题。在特定的仿真环境里,粒子群优化算法可训练机器人的几个基本行为,经过学习使机器人具有更好的鲁棒性和自适应学习能力。根据机器人不同行为在复杂环境下的性能指标,CBR可从案例库中选择特定的行为,并将其参数传送到粒子群优化算法的初始解库,从而加速整体的学习过程。利用机器人仿真软件MissionLab,采用基于行为的多机器人编队任务,用来测试该算法的有效性。仿真和实验结果表明,案例推理方法和粒子群优化算法相结合,使机器人获得更优的控制参数,同时在未知环境下的多机器人编队具有更好的性能。
Taking the multi-robot learning under unknown circumstances as the research platform, the case-based reasoning method can store the previous problem and solution information, and the long-term memory characteristics of this method can help the particle swarm optimization algorithm to better solve new problems. In a specific simulation environment, Particle Swarm Optimization (PSO) can train several basic behaviors of robots, which make the robot have better robustness and adaptive learning ability through learning. According to the performance of robots in complex environment under different circumstances, CBR can select specific behaviors from the case base and transfer their parameters to the initial solution library of particle swarm optimization algorithm, thus accelerating the overall learning process. Using the robot simulation software MissionLab, a behavioral-based multi-robot formation task is used to test the effectiveness of the algorithm. The simulation and experimental results show that the combination of case-based reasoning method and particle swarm optimization algorithm can make robots obtain better control parameters and have better performance in multi-robot formation in unknown environment.