论文部分内容阅读
针对本质上非线性、强耦合的两轮自平衡机器人复杂动态系统,构造操作条件反射概率自动机(OCPA)仿生自主学习系统.OCPA仿生自主学习系统是一个基于Skinner操作条件反射的概率自动机,主要特征在于模拟生物的操作条件反射机制,具有仿生的自组织功能,包括自学习和自适应功能,可用于描述、模拟、设计各种自组织系统.从理论上分析OCPA学习系统的操作条件反射学习机制的收敛性.应用于两轮机器人姿态平衡控制的仿真和实验结果均表明,设计的OCPA仿生自主学习系统不需要系统的模型,通过模拟生物的操作条件反射机制,自组织地渐进形成、发展和完善其姿态平衡控制技能.
Aiming at the intrinsically nonlinear and strongly coupled two-wheeled self-balancing robot complex dynamic system, an operating condition reflective probability automaton (OCPA) is proposed. The OCPA bionic autonomous learning system is a probabilistic automaton based on Skinner operating conditions, The main features of the simulation system are the operating conditions of the simulated biological reflection mechanism, with bionic self-organizing functions, including self-learning and self-adaptive function, can be used to describe, simulate, design a variety of self-organizing system.Study theoretically OCPA learning system operating conditions reflect The convergence of learning mechanism.The simulation and experimental results applied to attitude balance control of two-wheeled robot show that the designed OCPA system does not need a systematic model and can be self-organized and gradually formed by simulating the operating conditions of the biological system, Develop and improve their attitude control skills balance.