【作者】 李磊；

【导师】 朱齐丹；

【作者基本信息】 哈尔滨工程大学 ， 控制理论与控制工程， 2008， 博士

【摘要】 六自由度并联平台具有承载能力强、刚度大、精度高、动态响应快和累积误差小等特点，因此在机器人、运动模拟器、新型机床和飞船对接器等领域获得了越来越广泛的应用。本论文所研究的六自由度并联平台是用来模拟舰船在海洋中运动的设备，与舰船进行实际航行试验相比，它具有可控性、无破坏性、经济性和可靠性等优点，因此其应用前景非常广阔。本论文针对提高六自由度并联平台的运动性能，主要围绕位置正解和控制策略两方面开展相关的研究工作。首先，详细介绍了整个六自由度并联平台的结构，通过拉格朗日法建立了并联平台的动力学模型，分析了液压缸和电液伺服阀的数学模型，并建立了液压伺服系统的数学模型，为实现六自由度并联平台的控制器设计奠定了基础。其次，利用神经网络法无需初值选取和Newton-Raphson迭代法求解精度高的优点，提出了神经网络法和Newton-Raphson迭代法相结合的六自由度并联平台位置正解方法。此方法先利用神经网络正解法得到位置正解的粗解，再以此为初值利用Newton-Raphson迭代法通过较少次的迭代获得正解的精确解，这使得两种方法得到了很好的互补。同时为了提高神经网络正解法的学习能力，结合不完全演化策略与空间收缩思想，提出了基于比例再生空间收缩的粒子群优化算法，详细分析了收缩系数和收缩周期对粒子群算法搜索能力的影响，通过基于比例再生空间收缩粒子群优化算法优化神经网络的权值来提高位置正解的精度。再次，针对液压伺服系统具有复杂外干扰和系统参数摄动的特点，提出了一种鲁棒复合控制结构，该结构由PD控制器、鲁棒内回路控制器、零相位误差跟踪控制器和动态模糊神经网络补偿器等四部分组成。其中，PD控制器实现系统的反馈控制，以保证整个系统的稳定性；鲁棒内回路控制器能够抑制外界不确定性干扰和系统参数摄动对系统的影响；零相位误差跟踪控制器能够提高系统响应的快速性；动态模糊神经网络补偿器实现对PD控制器的补偿控制，进一步提高了系统对外界不确定性干扰和系统参数摄动影响的抑制能力。最后，为了降低六自由度并联平台建模误差和外界干扰对系统性能的影响，利用动力学模型中M(q)-2C(q，(?))的斜对称性，通过Lyapunov直接法得到了鲁棒控制律，设计了一种基于逆向力补偿的鲁棒跟踪控制策略，其内回路采用逆动力学进行补偿，外回路为依据耗散性理论设计的鲁棒控制器，保证了系统跟踪误差的一致终值有界，提高了系统的鲁棒性。仿真结果验证了本论文提出的六自由度并联平台位置正解和控制方法的有效性。更多还原

【Abstract】 The 6-DOF parallel platform has many advantages of large bearing capacity,good rigidity,high accuracy,fast dynamic response and without accumulativeerror,so it is widely used in many fields,such as robot,motion simulator,newtype machine and spacecraft docking.In this paper,the 6-DOF parallel platform isused as an equipment to simulate the ship moving in the sea,and it hascontrollable,non-destructive,economical and reliable abilities compared with theactual trial,so it has very broad application prospects.In order to improve theperformance of the 6-DOF parallel platform,this paper mainly focuses on forwardkinematics and control strategy.Firstly,this paper introduces the structure of 6-DOF parallel platform indetail,establishes the platform dynamics model with Lagrange method,analyzesthe mathematical model of hydraulic cylinders and servo valve,and establishesthe hydraulic servo model.So these establish the foundation for the realization of6-DOF parallel platform controller.Secondly,this paper uses the advantages of neural network method with nolimit of initial value and Newton-Raphson method with high accuracy,andpresents a forward kinematics method which is put the neural network methodand the Newton-Raphson method together.It uses the neural network method tosolve the imprecise solution of forward kinematics,and then uses theNewton-Raphson method to solve the precise solution after a few iterations withthe imprecise solution as initial iteration,so it has well complementary with thetwo methods.In order to improve the learning ability of neural network method,itpresents renewable proportion space contracting particle swarm optimizationalgorithm with non-complete evolution and space contracting,and analyzesparticle swarm optimization algorithm with contraction factor and contractioncycle in detail.It uses renewable proportion space contracting particle swarm optimization algorithm to optimize the weights of neural network to improveaccuracy of the forward kinematics.Thirdly,this paper presents a robust composite control structure withhydraulic servo systems of complex outside interference and system parametersperturbation,and it is composed of PD controller,robust inner loop controller,zero phase error tracking controller and dynamic fuzzy neural network controller.In this composite control structure,PD controller realizes feedback control toensure the stability of the whole system;robust inner loop controller can inhibitthe influence of uncertainty outside interference and system parametersperturbation;zero phase error tracking controller can improve the rapid response;dynamic fuzzy neural network controller realizes the compensation of PDcontroller to further improve the suppression of uncertainty outside interferenceand system parameters perturbation.Finally,in order to reduce the modeling error and the external disturbance ofthe 6-DOF parallel platform,it designs a robust tracking control strategy based oninverse dynamics compensation,which is used Lyapunov direct method to obtainthe robust control law with the skewed symmetry of (?)(q)-2C(q,(?)) indynamics model.The control strategy uses inverse dynamics to compensate theinner loop and designs the robust controller based on dissipation theory in outerloop,and it can guarantee the tracking errors uniformly and ultimately boundedand enhance the robustness of the system.Simulation results show that the forward kinematics and the control methodsof 6-DOF parallel platform are all effective.更多还原