论文部分内容阅读
为了实现两轮机器人的自平衡控制,利用Skinner操作条件反射机理,以概率自动机为平台,融入模糊推理,构造了模糊操作条件概率自动机(OCPA)仿生自主学习系统.该学习系统是一个从状态集合到操作行为集合的随机映射,采用操作条件反射学习机制,从操作行为集合中随机学习作为控制系统控制信号的最优行为,并利用学习到的操作行为取向值信息,调整操作条件反射学习算法.此外,学习系统还引入行为熵,以验证其自学习和自组织能力.应用于两轮机器人自平衡控制的仿真结果,验证了模糊OCPA学习系统的可行性.
In order to realize the self-balancing control of two-wheeled robot, this paper constructs the fuzzy operating condition probability automata (OCPA) bionic autonomous learning system by using Skinner operating condition reflex mechanism and probabilistic automaton as the platform, State set to the set of operational behavior of the random mapping, the use of operating conditions reflection learning mechanism, from the set of operational behavior of random control as the control system of the optimal behavior of the control signal and the use of learning behavior behavior orientation value information to adjust the operating conditions of reflection learning Algorithm.In addition, the learning system also introduces behavioral entropy to verify its self-learning and self-organizing ability.The simulation results of two-wheeled robot self-balancing control verify the feasibility of fuzzy OCPA learning system.