一种不确定对象的自适应智能PID控制系统

Study on self-adaptive intelligent PID control system for uncertain objects

针对不确定非线性、时滞对象,提出了一种自适应智能PID控制系统。将模糊神经网络和PID神经网络相结合,构成一种智能型PID控制器;控制器参数采用混沌策略与粒子群算法结合的混沌粒子群离线优化和误差反传算法在线调整相结合的方法获得;通过引入最小二乘支持向量机用作辩识器,使控制系统能处理具有未知特性的不确定对象的控制问题。仿真结果表明:通过辩识器的良好非线性映射能力和控制器及其优化算法的有效结合,系统响应速度快、平稳、超调小,且具有一定的鲁棒性,验证了该系统的可行性和有效性。

Aiming at nonlinear and time lag objects, a control system is presented, which consists of a novel self-adaptive intelligent PID controller. LS-SVM model is used as the identifier for uncertain objects. The novel controller combines fuzzy neural network and PID neural network to overcome the problems of conventional fuzzy method. The parameters of the controller are optimized by the mixed learning method integrating offline particle swarm optimization algorithm, which combines chaos strategies with global searching ability and online BP algorithm with local searching ability. Simulation results show that with the fine nonlinear mapping ability of LS-SVM and the controller with effective training algorithm, good control performance is obtained, and the feasibility and validity of the control system are proved.