基于人工神经网络的多机器人协作学习研究

Cooperative learning based on artificial neuron network in multi-robot system

机器人足球比赛是一个有趣并且复杂的新兴的人工智能研究领域,它是一个典型的多智能体系统。文中主要研究机器人足球比赛中的协作行为的学习问题,采用人工神经网络算法实现了两个足球机器人的传球学习,实验结果表明了该方法的有效性。最后讨论了对ＢＰ算法的诸多改进方法。

Robot Soccer is an interesting and complex new area of artificial intelligence, and it is a typical multi-agent system. This paper focuses on cooperative learning in robot soccer. Artificial neuron network is applied to learn passing ball for two soccer robots. The experiment results show this method is very effective. Finally many improvements are discussed for BP algorithm.