免疫克隆算法调节参数的非线性控制器设计

A nonlinear controller design using an immune clonal selection algorithm

以模糊神经网络(FNN)为基础,结合误差线性反馈构造了一种新型的非线性控制器.非线性控制器的设计难点在于参数的确定问题,用传统的算法对控制器参数寻优时容易陷入局部收敛,难于取得可靠的参数,因此提出一种改进的免疫克隆选择算法,用于确定非线性控制器的最优参数.倒立摆的仿真实验表明改进的免疫克隆算法在控制器参数寻优中取得良好的效果,所设计的控制器具有很强的非线性适应能力.

A nonlinear controller based on a fuzzy neural network （FNN） and linear error feedback was constructed. But the difficulty in nonlinear controller design lies in identification of the parameters being controlled. Optimization of the parameters may lead to a local convergence if traditional methods are used, resulting in unreliable operation. We therefore developed a modified immune clonal selection algorithm （m-ICSA） to optimize parameters of the nonlinear controller. A simulation with an inverted pendulum demonstrated that the m-ICSA effectively optimizes controller parameters and the design has strong nonlinear adaptive ability.