空间机器人自主接管非合作目标的轨迹规划与控制研究

发布时间：2018-03-09 10:05

本文选题：空间机器人　切入点：非合作目标　出处：《哈尔滨工业大学》2017年博士论文　论文类型：学位论文

【摘要】：利用空间机器人对高价值航天器开展在轨维修维护是合理使用轨道资源、恢复故障航天器的功能、提升航天器的经济效益、降低航天产业成本、保障空间资产安全的重要手段之一,具有显著的政治、经济和社会意义。空间机器人接管非合作目标后形成复合体系统,由于该系统的非线性、耦合性强,而且初始动量未知,操作过程基座姿态约束条件多,给系统的构型设计、参数辨识、轨迹规划、基座姿态稳定等带来一系列挑战。为了深入理解空间机器人与目标组成的复合体系统的动力学特性,推导了复合体系统的动力学方程,并针对典型运动轨迹分析了不同质量、惯量目标与基座之间的耦合情况,基于此提出了空间机器人系统的控制策略。进一步地,分析了单臂及多臂空间机器人系统的工作空间及可操作度。基于可操作度指标,对采用多臂的空间机器人系统的构型进行了优化。针对复合体系统初始动量未知时非合作目标动力学参数实时辨识过程中基座姿态扰动较大的问题,提出一种基于多约束条件下零反作用空间自适应轨迹规划的参数辨识方法。以多维关节角度势函数的梯度为约束,利用空间机器人的零反作用空间特性,采用带遗忘因子的最小二乘自适应方法在关节角度约束的条件下实现对基座姿态较小的扰动,建立系统动量的差分式增量方程,精确辨识出目标的动力学参数。针对空间机器人操作自旋目标过程中基座姿态失稳的难题,提出基于自适应滑模控制的基座与机械臂协同稳定的方法。该方法采用了延时参数估计来降低对系统动力学精确模型的依赖,设计一种增益导数与滑模变量成比例的自适应切换律,实现了操作目标过程中良好的控制性能和较小的颤振效应。通过空间机器人基座和机械臂的协调运动,有效减小了目标操控过程中产生的基座姿态扰动,实现了复合体系统的稳定控制。为了对本文提出的方法进行验证和评估,研制了一套半物理实验验证系统,该系统由空间机器人的数学模型和模块化可重构机械臂组成,主要思想是将动力学模型与实物结合起来。基于该系统,开展了空间机器人零反作用空间自适应轨迹规划的实验,验证了算法的有效性,也说明了空间机器人半物理仿真实验系统具备开展实验验证的能力。该实验系统具有较好的可扩展性,可在后续的研究中继续发挥作用。从在轨服务和空间安全的发展趋势来看,空间机器人接管控制非合作目标是未来的研究应用热点。本文的研究成果为空间机器人技术的实用化提供参考。
[Abstract]:Using space robots to maintain and maintain high value spacecraft in orbit is to make rational use of orbital resources, restore the functions of the faulty spacecraft, enhance the economic benefits of the spacecraft, and reduce the cost of the space industry. One of the important means of ensuring the safety of space assets has significant political, economic and social significance.Space robots take over non-cooperative targets and form a complex system because of its nonlinearity, strong coupling, and unknown initial momentum. In order to understand the dynamic characteristics of the complex system composed of space robot and target, there are a series of challenges to the configuration design, parameter identification, trajectory planning and attitude stability of the base during the operation. The dynamic equation of the complex system is derived, and the coupling between the target of inertia and the base is analyzed according to the typical motion trajectory. Based on this, the control strategy of the space robot system is proposed. The workspace and operability of single-arm and multi-arm space robot systems are analyzed. The configuration of a space robot system with a Dobby is optimized. In view of the large disturbance of the base attitude in the real-time identification process of the dynamic parameters of the non-cooperative target when the initial momentum of the complex system is unknown, the structure of the space robot system is optimized. In this paper, a parameter identification method for adaptive trajectory planning in zero-reaction space under multi-constraint condition is proposed. Taking the gradient of multi-dimensional joint angle potential function as constraint, the spatial characteristics of zero-reaction space robot are used. The least square adaptive method with forgetting factor is used to realize the small disturbance to the base attitude under the condition of joint angle constraint, and the differential increment equation of the momentum of the system is established. The dynamic parameters of the target are accurately identified. In view of the difficult problem of the attitude instability of the base during the operation of the spin target by the space robot, A cooperative stabilization method based on adaptive sliding mode control for pedestal and manipulator is proposed, in which delay parameter estimation is used to reduce the dependence on precise dynamic model of the system. An adaptive switching law with gain derivative proportional to sliding mode variables is designed to achieve good control performance and small flutter effect in the process of operation. In order to verify and evaluate the method proposed in this paper, a semi-physical experimental verification system is developed for effectively reducing the base attitude disturbance in the course of target control, and realizing the stability control of the complex system. The system is composed of the mathematical model of space robot and the modular reconfigurable manipulator. The main idea is to combine the dynamic model with the real object. Based on the system, the experiment of space robot zero-reaction space adaptive trajectory planning is carried out. The validity of the algorithm is verified, and the capability of the experimental system of semi-physical simulation of space robot is demonstrated. Could continue to play a role in subsequent research. In view of trends in in-orbit services and space security, The non-cooperative target of space robot take-over control is a hotspot in future research and application. The research results in this paper provide a reference for the practical application of space robot technology.
【学位授予单位】：哈尔滨工业大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】