基于RBF-ARX模型的四旋翼飞行器的预测控制

摘要

四旋翼飞行器运行中具有不稳定、非线性和强耦合特性,较难建立其准确的数学模型,针对这个问题,提出了基于RBF-ARX模型的预测控制设计方法;RBF-ARX模型是线性自回归(Auto-Regressive eXogenous,ARX)模型和高斯径向基函数(RadialBasis Function,RBF)神经网络相结合设计的模型,可用于建立非线性系统的全局模型,描述非线性系统的非线性特征.预测控制算法根据系统输入、输出信号预测对象未来输出变化趋势,并将其与系统实际输出的误差反馈校正,使误差最小;该法首先建立四旋翼飞行器的RBF-ARX模型结构,就模型参数的辨识、优化给出了详细分析;并基于该模型设计了系统预测控制器,最后通过仿真和实时控制效果证实了该方法的可行性和有效性.%The Quad--Rotor Helicopter which is unsteady、nonlinear and coupled with each other makes it difficult to establish accurate mathematical model.for the problem,the design of predictive controller based on RBF-ARX mode were discussed.The RBF-ARX model is a combination model of linear autoregressive (Auto-Regressive eXogenous,ARX) model and Gauss radial basis function (Radial Basis Function) neural network (RBF),It can be used to establish the global model of nonlinear system and describe the nonlinear characteristics of nonlinear system.Predictive control strategy predicts the trend of the system's output according to the input and output of the system,and the error feedback of the system's actual output is adjusted to minimize the error.The structure of the RBF-ARX model was built firstly,the parameters' identification,optimization were analyzed in detail.The predictive controller is designed based on this model Finally,the simulation results and real time control verify the feasibility and validity of the method.