针对工业过程中被控对象往往具有复杂、强非线性、多变量的特性,提出一种基于多变量径向基函数(radial base function,RBF)神经网络的非线性内部迭代预测控制方法.采用多个RBF神经网络在线逼近多输入多输出的非线性系统,得到一个近似模...针对工业过程中被控对象往往具有复杂、强非线性、多变量的特性,提出一种基于多变量径向基函数(radial base function,RBF)神经网络的非线性内部迭代预测控制方法.采用多个RBF神经网络在线逼近多输入多输出的非线性系统,得到一个近似模型作为预测模型.同时为了减少求解系统控制律的计算量,将每个输出预测值沿着输入轨迹展开,从而把求解复杂非线性优化问题转化为求解简单的二次规划问题,解决了在线实时递推控制律时求解非线性微分方程的困难.最后通过t次内部迭代直至满足迭代条件,得到了最优的控制律.p H中和过程的仿真结果表明了该算法是有效而可行的.展开更多
To overcome nonlinear and 6-DOF(degrees of freedom)under-actuated problems for the attitude and position of quadrotor UAVs,an adaptive backstepping sliding mode method for flight attitude of quadrotor UAVs is proposed...To overcome nonlinear and 6-DOF(degrees of freedom)under-actuated problems for the attitude and position of quadrotor UAVs,an adaptive backstepping sliding mode method for flight attitude of quadrotor UAVs is proposed,in which an adaptive law is designed to online estimate the parameter variations and the upper bound of external disturbances and the assessments is utilized to compensate the backstepping sliding mode control.In addition,the tracking error of the design method is shown to asymptotically converge to zero by using Lyapunov theory.Finally,based on the numerical simulation of quadrotor UAVs using the setting parameters,the results show that the proposed control approach can stabilize the attitude and has hover flight capabilities under the parameter perturbations and external disturbances.展开更多
文摘针对工业过程中被控对象往往具有复杂、强非线性、多变量的特性,提出一种基于多变量径向基函数(radial base function,RBF)神经网络的非线性内部迭代预测控制方法.采用多个RBF神经网络在线逼近多输入多输出的非线性系统,得到一个近似模型作为预测模型.同时为了减少求解系统控制律的计算量,将每个输出预测值沿着输入轨迹展开,从而把求解复杂非线性优化问题转化为求解简单的二次规划问题,解决了在线实时递推控制律时求解非线性微分方程的困难.最后通过t次内部迭代直至满足迭代条件,得到了最优的控制律.p H中和过程的仿真结果表明了该算法是有效而可行的.
基金Project(61203021)supported by the National Natural Science Foundation of ChinaProject(2011216011)supported by the Scientific and Technological Project of Liaoning Province,China+1 种基金Project(2013020024)supported by the Natural Science Foundation of Liaoning Province,ChinaProjects(LJQ2015061,LR2015034)supported by the Program for Liaoning Excellent Talents in University,China
文摘To overcome nonlinear and 6-DOF(degrees of freedom)under-actuated problems for the attitude and position of quadrotor UAVs,an adaptive backstepping sliding mode method for flight attitude of quadrotor UAVs is proposed,in which an adaptive law is designed to online estimate the parameter variations and the upper bound of external disturbances and the assessments is utilized to compensate the backstepping sliding mode control.In addition,the tracking error of the design method is shown to asymptotically converge to zero by using Lyapunov theory.Finally,based on the numerical simulation of quadrotor UAVs using the setting parameters,the results show that the proposed control approach can stabilize the attitude and has hover flight capabilities under the parameter perturbations and external disturbances.
