基于RBF神经网络的三轴云台自适应滑模控制

Trajectory Tracking Control of Three-axis Gimbal Based on RBF Neural Network

针对云台系统在日常运转过程中,易受到外界非线性干扰、参数时变等不确定因素影响的问题,提出了一种基于径向基函数神经网络(RBFNN)的自适应滑模控制方法。根据对云台系统工作时性能状态进行分析,建立了动力学模型;利用扩张状态观测器实现对内外部环境扰动量的实时预测,从而获取到神经网络自学习的信息;通过Lyapunov分析法推导出云台系统的滑模控制率,并采用一种新型饱和函数消除了滑模抖振对系统带来的影响,也使得控制量切换时更加具有连续性。仿真结果表明,所提三轴云台控制策略与普通滑模控制方法相比,控制精度更高且抗干扰能力更强。

Aiming at the problem of impact brought by uncertain factors such as external non-linear interference and parameter time-varying during the daily operation of the Gimbal system,an adaptive sliding control system based on the radial base function neural network(RBFNN)has been proposed.Firstly,a dynamic model is established based on the analysis of the performance status of the Gimbal system at work.Secondly,the information about self-learning of the neural networks is obtained with real-time prediction of the internal and external environment disturbance by using the expansion state observer.Lastly,the sliding control rate of the Gimbal system is derived by using the Lyapunov analysis method,and a new saturation function is adopted to eliminate the impact of the system brought by the shake of sliding mode,which also makes the control volume switch more continuous.The simulation results demonstrates that the trajectory of the three-axis gimbal control strategy shows higher control accuracy and stronger anti-interference ability compared with traditional sliding mode control methods.