SIA OpenIR  > 空间自动化技术研究室
面向在轨捕获任务的空间机器人运动规划与解耦控制研究
Alternative TitleMotion Planning and Decoupling Control of Space Robots for On-Orbit Capture Tasks
张鑫1,2
Department空间自动化技术研究室
Thesis Advisor刘金国
Keyword空间机器人 复合刚体动力学建模 动力学奇异性 运动规划 抓捕策略
Pages146页
Degree Discipline机械电子工程
Degree Name博士
2019-05-19
Degree Grantor中国科学院沈阳自动化研究所
Place of Conferral沈阳
Abstract面向在轨服务的空间机器人工作模式主要有两种:自由飞行模式和自由漂浮模式。其中,自由飞行模式下的空间机器人,其基体的位置和姿态是完全可控的;而自由漂浮模式下的空间机器人,其基体的位置和姿态是不受控制的。由于,不同工作模式下的系统具有不同运动学和动力学特性,导致不同工作模式下捕获任务中难点与侧重不同。从当前的文献来看,大多数学者的研究集中于自由漂浮空间机器人,这主要由于其欠驱动多体动力学系统具有丰富的理论研究意义;而自由飞行空间机器人,更偏重于实际工程应用。从空间机器人技术的发展趋势来看,其构型逐渐从单臂式向多臂式发展,其服务对象逐渐从面向合作目标向面向非合作目标方向发展。基于上述背景,本文的研究对象将不局限于单一构型的空间机器人,将面向不同工作模式下空间机器人目标捕获任务中的难点和瓶颈问题,开展如下几个方面的研究工作:(1)自由漂浮模式下的空间机器人目标捕获策略与运动规划研究。自由漂浮空间机器人系统仅机械臂受控,系统运动学方程的雅可比矩阵为广义雅可比矩阵。由于广义雅可比矩阵具有动力学奇异性,它的奇异性不仅与系统的尺寸参数有关,而且与系统的惯性参数有关,因此无法预先判断系统的奇异构型,给空间机器人的运动规划带来巨大的挑战。本文基于遥编程框架,提出一种运动参考变量与时间无关的自主规划方法。首先,通过分析捕获任务的抓捕约束,将轨迹规划问题转化为轨迹优化问题;然后,采用正向微分运动学、数值积分以及优化算法,设计一套可有效地回避广义雅可比矩阵的动力学奇异的离线规划流程;其次,为进一步为提高离线规划效率,提出一种考虑基体停靠位置和抓取域的新型抓捕策略。此外,考虑到基体姿态可以通过控制动量轮进行补偿,推导了基体姿态可控模式下漂浮基空间机器人的速度级运动方程,并基于奇异值滤波方法,研究了空间机器人在操作空间的在线规划问题。(2)自由飞行模式下的空间机器人解耦控制方法研究。空间机器人由基体卫星和机械臂两部分组成,二者之间存在动力学耦合作用。由于动力学耦合作用,机械臂的运动会引起基体的位姿改变,进而,基体位姿的改变也会影响机械臂的操作精度,因此,主动补偿耦合对空间机器人的精准操作具有重要意义。本文基于系统整体模型出发,分析了系统的耦合特性,并提出两种鲁棒解耦控制方法:基于模型(基于计算力矩的滑模控制)和无模型方法(基于时延预测的滑模控制),通过定性分析和定量分析对比两种控制方法的效果。(3)面向不具备抓捕特征目标的捕获方法研究。本文将非合作目标进一步细分为:可抓捕目标和不具备抓捕特征的目标。可抓捕目标带有可抓捕特征,采用常规机械手爪即可完成抓捕任务;而不具备抓捕特征的目标不带有抓捕特征,无法通过常规机械手爪抓捕。针对不具备抓捕特征的目标,本文提出一种锁笼对捕获方法,并在机械臂速度操作性和力操作性的基础上,提出锁困操作性概念以及其性能指标,用于定量地描述所提出方法的捕获能力。然后,基于该性能指标,提出一种规划预捕获构型的规划算法以获得最优捕获能力。最后,搭建二维气浮式空间机器人实验平台,验证所提捕获方法的有效性。以上三方面研究,可为后续我国在轨服务空间机器人的目标捕获任务提供一定的理论基础与技术支撑。
Other AbstractThere are two main working modes for OOS space robots, i.e., free-flying mode and free-floating mode. For free-flying space robots, the position and attitude of their bases are actively controlled; for free-floating space robots, neither of them is controlled. Usually, space robots in different working modes have different kinematic and dynamic characteristics, which leads space robots in different working modes to face different problems in the capture tasks. From the current literature, most researchers focus on the research of the free-floating space robots because of their multi-body underactuated system properties, which includes lots of theoretical research significance. However, for free-flying space robots, they are apt to the practical applications. From the development trend of space robots, the configuration of space robots is gradually developing from single arm to multi-arms, and the serviced target is gradually extending from the cooperative objects to uncooperative objects. In summary, the research object in this paper is not limited to some space robot system, and we aim at the tough issues and bottleneck problems of capturing objects by single/dual-arm space robots working in different working modes. (1) The strategy and motion planning for space robots capturing objects in free-floating mode. When the space is working in free-floating mode, only its manipulator is under control. The Jacobian matrix in the kinematic equation is called Generalized Jacobian Matrix (GJM), and CJM has the dynamic singularity characteristics. As the dynamic singularity is related to not only the kinematic parameters but also the dynamic parameters, the singular configuration cannot be predicted in joint space, which brings a huge challenge to the motion planning system. This paper proposed an autonomous off-line motion planning approach based on the tele-programming frame, where the planned motion reference variables have nothing to do with time. First, we analyze the constraints of the capture task and convert the planning problem into an optimization problem. Then, we use the forward differential kinematic equation, the numerical integration, and particle swarm optimization (PSO) to formulate a motion planning procedure, which can avoid the dynamic singularity. Furthermore, for improving the efficiency of the off-line planning, we proposed a novel motion planning strategy with considering the base berth position and grasping area. Besides, considering that the base attitude can be compensated by momentum wheels, we deduce the velocity-level motion equation of the floating space robot with an attitude-controlled base. Base on the singularity filtering technique, we study the on-line planning problems in the operational space. (2) The robust decoupling control of space robots in free-flying mode. The entire space robot system consists of a satellite and one or more manipulators, and there exists a dynamic coupling effect between the base and manipulators. Due to the dynamic coupling, the motion of the manipulator will induce the motion of the base, and thereby the coupling motion of the base will deteriorate the operational accuracy of the end-effector. Thus, the active compensation of the dynamic coupling has a significant impact on the accurate operation missions for space robot systems. Based on the viewpoint of the entire system model, we analyzed the coupling property and proposed two kinds of robust control schemes, i.e., the model-based method (Computed-torque based sliding mode control) and model-free method (Time-delay based sliding mode control). Through the qualitative analysis and quantitative analysis, we will compare the control performance of two proposed controllers. (3) Research on capturing objects without graspable features for space robots. In this paper, we further classified the uncooperative objects into graspable objects and objects without graspable features. Here, the graspable objects owned some graspable handles which can be captured by the conventional mechanical grippers. On the contrary, the other one does not own the graspable handles, and the mechanical grippers are inapplicable. For capturing objects without graspable features, we proposed a caging-pair capturing method which use a series of hollow-shaped end-effector pairs to cage objects without graspable features. Based on the velocity and force manipulability of robotic arms, we proposed the conception of caging manipulability and its evaluation index to quantitatively describe the capturing performance. Furthermore, based on the caging manipulability, an effective algorithm by the optimization of the desired caging manipulability index is proposed to the near-optimal configurations for pre-grasping cages. Finally, we establish a two-dimensional (2D) air-bearing test platform to verify the robustness and superior capturing performance of the proposed method by experiments. The above three research aspects can provide some theoretical foundation and technical support for Chinese OOS space robots in tasks of capturing targets.
Language中文
Contribution Rank1
Document Type学位论文
Identifierhttp://ir.sia.cn/handle/173321/25158
Collection空间自动化技术研究室
Affiliation1.中国科学院沈阳自动化研究所
2.中国科学院大学
Recommended Citation
GB/T 7714
张鑫. 面向在轨捕获任务的空间机器人运动规划与解耦控制研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2019.
Files in This Item:
File Name/Size DocType Version Access License
面向在轨捕获任务的空间机器人运动规划与解(11893KB)学位论文 开放获取CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[张鑫]'s Articles
Baidu academic
Similar articles in Baidu academic
[张鑫]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[张鑫]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.