基于干扰补偿的全地域发射平台高性能控制方法研究

【作者】 刘雷；

【导师】 胡健；

【作者基本信息】 南京理工大学 ， 机械工程（专业学位）， 2018， 硕士

【摘要】 随着全地域发射平台的应用越来越广泛,对其控制性能的要求也随之越来越高。但由于全地域发射平台具有参数不确定性和其他不确定非线性的特点,这给控制性能的提高带来了难度,因此如何消除这些非线性因素的影响是提高控制性能的关键。针对如何消除非线性因素的影响,本文在以下几个方面展开了工作:1、介绍全地域发射平台的系统构成,阐述和推导电机伺服系统的三环传递函数。对全地域发射平台干扰特性进行分析,并建立电机伺服系统两轴耦合动力学模型。2、结合BP神经网络与PID控制,设计神经网络监督控制器。利用神经网络的自学习能力去弥补PID控制无法自适应的缺陷,有效提高了系统的自适应能力和稳定跟踪精度。仿真结果表明神经网络监督控制不仅能够保证系统的快速响应能力,同时提高了控制器的自适应能力和稳定跟踪精度。3、基于全地域发射平台两轴耦合的动力学模型,设计基于RBF神经网络的自适应鲁棒控制器。利用参数自适应率可以在线更新系统参数,RBF神经网络可以用来对系统不确定非线性进行估计,并将估计值加入到前馈补偿项中进行前馈补偿,提高了系统的抗干扰能力与稳定跟踪精度。仿真结果表明基于RBF神经网络的自适应鲁棒控制器能够有效削弱系统不确定性对系统造成的影响。4、以DSPF2812为核心设计位置控制器的软硬件,并将控制算法移植到位置控制器中进行实验。实验结果验证了神经网络监督控制策略和基于神经网络的自适应鲁棒控制策略均能获得优异的控制性能,为提高全地域发射平台的控制性能提供了参考依据。更多还原

【Abstract】 As the application of the all-terrain launching platform becomes more and more widespread,the requirements for the control performance of the all-terrain launching platform are also getting higher and higher.Since the all-terrain launching platform has the characteristics of parameter uncertainty and other uncertain nonlinearities,how to eliminate the impact of these non-linear factors is the key to improving control performance.Aiming at how to eliminate the influence of non-linear factors,the main works of this paper are focused as follows:1.The structure features of the all-terrain launching platform are introduced,and the tricyclic transfer function of the motor servo system is analyzed and established.In addition,the coupling disturbance characteristics of the all-terrain launching platform are analyzed,and then establish the two-axis coupling cdynamics model.2.Combined with BP neural network and PID control,the neural network supervisory controller is designed.Utilize the self-learning ability of neural network to improve the self-adaptive ability and the stable tracking accuracy of the controller.Simulation results show that neural network supervisory control could not only obtain the rapid response capability of the system,but also improve the adaptability and stability of the controller.3.Based on the dynamic model of the two-axis coupling of the all-terrain launching platform,the RBF neural network based adaptive robust controller is designed.The system parameters can be adjusted online using the parameter adaptation rate,the RBF neural network can be used to estimate the uncertain nonlinearity of the system and compensated in and feed-forward term,which improves the anti-interference ability and stability tracking accuracy of the system.The simulation results show that the RBF neural network based adaptive robust controller can reduce the influence of system uncertainty on the system effectively.4.The hardware circuit and software programming of position controller based on DSP F2812 are achieved,and the control algorithm is transplanted to the position controller for experiment.The experimental results verify that neural network supervisory controller and neural network-based adaptive robust controller could obtain good control performance,providing the reference for improving the performance of the launching platform.更多还原