«上一篇
文章快速检索     高级检索
下一篇»
  哈尔滨工程大学学报  2020, Vol. 41 Issue (12): 1847-1853  DOI: 10.11990/jheu.201905014
0

引用本文  

艾海平, 陈力. 空间机器人双臂捕获航天器操作的力/位置控制[J]. 哈尔滨工程大学学报, 2020, 41(12): 1847-1853. DOI: 10.11990/jheu.201905014.
AI Haiping, CHEN Li. Force/position fuzzy control of space robot capturing spacecraft by dual-arm clamping[J]. Journal of Harbin Engineering University, 2020, 41(12): 1847-1853. DOI: 10.11990/jheu.201905014.

基金项目

国家自然科学基金项目(11372073,11072061);江西理工大学博士科研启动项目(205200100514)

通信作者

艾海平, E-mail:ahpwuhan@163.com

作者简介

艾海平, 男, 博士研究生;
陈力, 男, 教授, 博士生导师

文章历史

收稿日期:2019-05-06
网络出版日期:2020-11-30
空间机器人双臂捕获航天器操作的力/位置控制
艾海平 1,2, 陈力 2     
1. 江西理工大学 能源与机械工程学院, 江西 南昌 330013;
2. 福州大学 机械工程及自动化学院, 福建 福州 350116
摘要:针对漂浮基空间机器人双臂夹持捕获航天器过程的冲击效应及其后镇定运动的力/位置控制问题,本文提出了一种基于无源性理论的模糊控制方案。依据多刚体系统动力学理论,获得了捕获前双臂空间机器人及目标航天器的动力学模型;应用系统动量守恒关系,捕获后闭链系统运动几何关系及力的传递规律,分析了捕获接触、碰撞过程对闭链系统产生的冲击效应,并进而建立了捕获操作完成后两者形成的闭链混合体系统动力学方程。以此为基础,针对非合作航天器带来的系统建模误差及扰动,基于无源性理论并结合速度观测器,设计了捕获操作后闭链混合体系统夹持操作的镇定运动力/位置模糊滑模控制方案。结果表明:提出的方案无需实际测量相关速度信号,并具有基于无源性理论所设计控制方案本身秉承的良好动态特性及较强的鲁棒性;可以有效抵御捕获操作引起的较强冲击效应对控制系统的干扰,快速实现闭链混合体系统的运动镇定。系统数值模拟仿真结果验证了所提控制方案的正确性。
关键词漂浮基空间机器人    双臂夹持捕获航天器    闭链混合体系统    无源性理论    冲击效应    速度观测器    力/位置模糊滑模控制    镇定运动    
Force/position fuzzy control of space robot capturing spacecraft by dual-arm clamping
AI Haiping 1,2, CHEN Li 2     
1. School of Energy and Mechanical Engineering, Jiangxi University of Science and Technology, Nanchang 330013, China;
2. School of Mechanical Engineering and Automation, Fuzhou University, Fuzhou 350116, China
Abstract: A fuzzy control scheme based on the passivity theory is proposed to solve the impact effect of floating-based space robots in the process of capturing spacecraft by dual-arm clamping and the force/position control of its post-stabilization movement. Dynamic models of the double-arm space robot and the target spacecraft before capture are obtained based on the multi-rigid system dynamics theory. Afterward, based on the law of the conservation of momentum and the motion geometry relation and force transfer law of the closed-chain system after the capture, the impact effect of the capture contact and collision process on the closed-chain system after the collision is analyzed. Then, the dynamic equation of the closed-chain mixture system formed by them after the capture operation is established. Moreover, aiming at the system modeling errors and disturbances caused by non-cooperative spacecraft, based on passivity theory and speed observer, a stabilization force/position fuzzy sliding mode control scheme for the clamping operation of the closed-chain hybrid system after capture operation is designed. The proposed scheme based on the passivity theory does not need to actually measure relevant speed signals and has good dynamic characteristics and strong robustness. Moreover, it can effectively resist the interference of the strong impact effect caused by the capture operation on the control system and quickly realize motion stabilization of the closed-chain hybrid system. The system numerical simulation results verify the correctness of the proposed control scheme.
Keywords: free-floating space robot    capturing spacecraft by dual-arm clamping    closed-chain composite system    passivity theory    impact effect    velocity observer    force/position fuzzy sliding mode control    stabilization control    

随着国内外空间技术的不断发展,空间任务日益增多,考虑到太空环境的恶劣性,宇航员出舱进行操作任务会面临很大的危险,利用空间机器人系统来替代宇航员完成空间操作任务成为较佳的选择[1-5]。值得注意的是,太空中现存的航天器有不少因为燃料耗尽、轨道偏移等原因失去控制,若对其进行回收接管,这些航天器仍能进行工作。为了实现上述任务,空间机器人对目标航天器的捕获能力成为其中必不可少的关键技术。考虑到捕获操作的复杂性,较单臂空间机器人系统[6-9]具有更大的负载能力、以及较为灵活的多任务协作能力的双臂及多臂空间机器人系统已成为新的研究趋势[10-15]。目前,双臂及多臂空间机器人系统进行捕获任务主要针对的是合作目标,这是因为合作目标上具备便于空间机器人抓持的把手,但针对非合作目标,抓持捕获操作难以进行;夹持捕获具有不需要目标协作把手,且无需考虑被捕获目标构型的优点,因而其对非合作目标更具适用性。空间机器人对非合作目标航天器完成夹持捕获操作后,二者形成了闭链混合体系统,这使得控制时既要考虑控制系统的位置,又要考虑夹持内力的协调分配,加大了控制的难度。同时,由于处于太空微重力环境下,闭链混合体系统载体与各构件之间存在强烈的动力学耦合作用[16],因此空间机器人双臂捕获航天器相关碰撞动力学与控制特点体现为:捕获操作前,两者构成的系统存在非完整动力学约束;捕获操作过程系统存在内部动量、动量矩及能量的传递变化;捕获前、后系统存在结构开、闭环结构变拓扑与闭环运动学、位置约束问题。上述多重问题的共存,使得双臂捕获航天器的控制问题远复杂于单臂捕获操作或地面机器人的控制。程靖等[16]提出了一种双臂空间机器人系统捕获目标后的分块滑模自适应神经网络控制。Nguyen等[17]基于零反应空间的操作理念,结合动量守恒定律实现对旋转目标的捕获。Bandyopadhyay等[18]提出了用于运载大型目标航天器的姿态控制策略及一种非线性跟踪控制器,所提控制器保证了跟踪误差的全局指数收敛性。Abiko等[19]提出了自由漂浮空间机器人抓取模型不确定的翻滚目标时的阻抗控制方案。但是上述研究多考虑的是抓持捕获控制问题。

本文针对载体位置不受控的空间机器人双臂夹持捕获非合作目标航天器的力/位置镇定控制,提出了基于无源性理论的模糊滑模控制方案。考虑到速度信号会因为噪声等原因无法精确测量,设计了结合速度观测器的控制器。针对由捕获非合作航天器所带来的建模误差及扰动项,设计了滑模补偿项以消除其影响。同时,鉴于模糊控制具备模型依赖性低、容错性强,且可处理具备不确定性信息复杂非线性系统的优点,结合模糊控制[20-22],可有效地消除滑模补偿项引起的抖振现象。所设计基于无源性理论的控制方案,具有鲁棒性强[23-24],对捕获冲击抗扰能力强的优点,可实现力/位置的协调控制。

1 双臂空间机器人夹持捕获目标航天器动力学建模

以空间机器人系统双臂夹持捕获目标航天器操作过程为例,其结构如图 1所示,建立其动力学模型。双臂空间机器人系统由漂浮刚性基座B0,刚性左臂及刚性右臂组成,左臂、右臂分别由刚性连杆Bi(i=1, 2, 3)及Bj(j=4, 5, 6)组成,被捕获目标航天器为Bt。任取一点O为惯性坐标系原点,建立系统惯性坐标系XOY。同时取双臂空间机器人各分体的连体坐标系为xiOiyi(i=0, 1, …, 6),其中O0为载体质心,Oi(i=1, 2, …, 6)为各关节相应转动铰中心,并设各臂杆长度为li(i=1, 2, …, 6),O0O1O0O4长度均为d0。系统总质心C及载体、左右臂各个臂杆的质心在惯性坐标系下位置矢量分别为rcr0ri(i=1, 2, …, 6),被捕获航天器质心在惯性坐标系下的位置矢量为rt

Download:
图 1 漂浮双臂空间机器人系统及目标航天器系统 Fig. 1 Dual-arm free-floating space robot system and target spacecraft system

针对图 1中未发生捕获碰撞的空间机器人系统及被捕获目标航天器系统,根据拉格朗日第二类方程及牛顿-欧拉法分别获得其动力学方程为:

$ \mathit{\boldsymbol{D}}(\mathit{\boldsymbol{q}})\mathit{\boldsymbol{\ddot q + H}}(\mathit{\boldsymbol{q}},\mathit{\boldsymbol{\dot q}})\mathit{\boldsymbol{\dot q}} = {[{{\bf{0}}_{1 \times 2}}\quad {\mathit{\boldsymbol{\tau }}^{\rm{T}}}]^{\rm{T}}} + {\mathit{\boldsymbol{J}}^{\rm{T}}}\mathit{\boldsymbol{F}} $ (1)
$ {\mathit{\boldsymbol{D}}_{\rm{t}}}{\mathit{\boldsymbol{\ddot q}}_{\rm{t}}} = \mathit{\boldsymbol{J}}_{\rm{t}}^{\rm{T}}{\mathit{\boldsymbol{F}}^\prime } $ (2)

式中:D(q)∈R9×9DtR3×3分别为空间机器人系统及被捕获航天器系统具有对称、正定性的惯量阵;H(q, $\dot{\boldsymbol{q}}$)$\dot{\boldsymbol{q}}$∈R9×1为空间机器人系统包含科氏力、离心力项;q=[q0T  θmT]T为空间机器人系统广义坐标列向量,其中q0=[x0  y0]T为载体质心位置坐标,θm=[θ0  θLT  θRT]T,其中θ0为载体姿态广义坐标,θL=[θ1  θ2  θ3]T为左臂广义坐标列向量,θR=[θ4  θ5  θ6]T为右臂广义坐标列向量;qt=[xt  yt  θt]T为被捕获航天器的广义坐标列向量;τ=[τ0  τLT  τRT]Tτ0为空间机器人系统载体姿态控制力矩,τLτRR3×1为左右臂各关节控制力矩。JR6×9为两机械臂末端对应的运动Jacobian矩阵,JtR6×3为被捕获航天器与机械臂末端对应2个接触点的运动Jacobian矩阵;FR6×1为2个机械臂末端所受力,F′为被捕获航天器所受作用力。FF′是一对相互作用力,满足F=-F′

由式(2)可得:

$ {\mathit{\boldsymbol{F}}^\prime } = {(\mathit{\boldsymbol{J}}_{\rm{t}}^{\rm{T}})^ + }{\mathit{\boldsymbol{D}}_{\rm{t}}}{\mathit{\boldsymbol{\ddot q}}_{\rm{t}}} + {\mathit{\boldsymbol{F}}_{\rm{I}}} $ (3)

式中:(JtT)+JtT的伪逆;(JtT)+Dt$\ddot{\boldsymbol{q}}$t为操作力项;FI为定义在JtT零空间的内力项,且有JtTFI=0。

根据牛顿第三定律,结合式(1)~(3),可得:

$ \begin{array}{*{20}{l}} {\mathit{\boldsymbol{D}}(\mathit{\boldsymbol{q}})\mathit{\boldsymbol{\ddot q}} + \mathit{\boldsymbol{H}}(\mathit{\boldsymbol{q}},\mathit{\boldsymbol{\dot q}})\mathit{\boldsymbol{\dot q}} = {{[{0_{1 \times 2}}\quad {\mathit{\boldsymbol{\tau }}^{\rm{T}}}]}^{\rm{T}}} - }\\ {{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\mathit{\boldsymbol{J}}^{\rm{T}}}{{(\mathit{\boldsymbol{J}}_{\rm{t}}^{\rm{T}})}^ + }{\mathit{\boldsymbol{D}}_{\rm{t}}}{{\mathit{\boldsymbol{\ddot q}}}_{\rm{t}}} - {\mathit{\boldsymbol{J}}^{\rm{T}}}{\mathit{\boldsymbol{F}}_{\rm{I}}}} \end{array} $ (4)

假设空间机器人系统在t0时刻对目标航天器进行夹持捕获,经过极小的时间Δt完成捕获操作。根据冲量定理,同时参考文献[15]相关推导,对式(4)两边进行积分,则可化简为:

$ \begin{array}{*{20}{l}} {{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} \mathit{\boldsymbol{D}}(\mathit{\boldsymbol{q}})(\mathit{\boldsymbol{\dot q}}({t_0} + \Delta t) - \mathit{\boldsymbol{\dot q}}({t_0}) + }\\ {{\mathit{\boldsymbol{J}}^{\rm{T}}}{{(\mathit{\boldsymbol{J}}_t^\mathit{\boldsymbol{T}})}^ + }{\mathit{\boldsymbol{D}}_t}({{\mathit{\boldsymbol{\dot q}}}_{\rm{t}}}({t_0} + \Delta t) - {{\mathit{\boldsymbol{\dot q}}}_{\rm{t}}}({t_0})) = 0} \end{array} $ (5)

对于捕获完成后空间机器人与目标航天器组成的混合体系统,空间机器人机械臂末端与目标航天器两端接触点速度相同,即从t0t时刻皆有:

$ \mathit{\boldsymbol{J\dot q}} = {\mathit{\boldsymbol{J}}_{\rm{t}}}{\mathit{\boldsymbol{\dot q}}_t} $ (6)

结合式(5)、(6),可得:

$ \mathit{\boldsymbol{\dot q}}({t_0} + \Delta t) = {\mathit{\boldsymbol{U}}^{ - 1}}[\mathit{\boldsymbol{D}}(\mathit{\boldsymbol{q}})\mathit{\boldsymbol{\dot q}}({t_0}) + {\mathit{\boldsymbol{J}}^{\rm{T}}}{(\mathit{\boldsymbol{J}}_{\rm{t}}^{\rm{T}})^ + }{\mathit{\boldsymbol{D}}_{\rm{t}}}{\mathit{\boldsymbol{\dot q}}_t}({t_0})] $ (7)

式中:U=D(q)+JT(JtT)+DtJt+JJt+Jt的伪逆。由式(7)可得夹持捕获对空间机器人系统产生的冲击效应。为了实现对碰撞失稳后的混合体系统的镇定控制,需要设计控制算法对其进行主动控制,以实现系统的稳定。

为实现对捕获后闭链混合体系统的控制方案的设计,需要获得捕获完成后空间机器人与目标航天器组成闭链混合体的综合动力学方程。对式(6)求导,并整理得:

$ {\mathit{\boldsymbol{\ddot q}}_{\rm{t}}} = \mathit{\boldsymbol{J}}_{\rm{t}}^ + [\mathit{\boldsymbol{J\ddot q}} + (\mathit{\boldsymbol{\dot J}} - {\mathit{\boldsymbol{\dot J}}_{\rm{t}}}\mathit{\boldsymbol{J}}_{\rm{t}}^ + \mathit{\boldsymbol{J}})\mathit{\boldsymbol{\dot q}}] $ (8)

结合式(4)、(8),可得到闭链混合体的综合动力学方程:

$ {\mathit{\boldsymbol{D}}_1}(\mathit{\boldsymbol{q}})\mathit{\boldsymbol{\ddot q}} + {\mathit{\boldsymbol{H}}_1}(\mathit{\boldsymbol{q}},\mathit{\boldsymbol{\dot q}})\mathit{\boldsymbol{\dot q}} = {[{{\bf{0}}_{1 \times 2}}\quad {\mathit{\boldsymbol{\tau }}^{\rm{T}}}]^{\rm{T}}} - {\mathit{\boldsymbol{J}}^{\rm{T}}}{\mathit{\boldsymbol{F}}_{\rm{I}}} $ (9)

其中:

$ \begin{array}{l} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\mathit{\boldsymbol{D}}_1}(\mathit{\boldsymbol{q}}) = \mathit{\boldsymbol{D}}(\mathit{\boldsymbol{q}}) + {\mathit{\boldsymbol{J}}^{\rm{T}}}{(\mathit{\boldsymbol{J}}_{\rm{t}}^{\rm{T}})^ + }{\kern 1pt} {\kern 1pt} {\mathit{\boldsymbol{D}}_{\rm{t}}}\mathit{\boldsymbol{J}}_{\rm{t}}^ + \mathit{\boldsymbol{J}},{\mathit{\boldsymbol{H}}_1}(\mathit{\boldsymbol{q}},\mathit{\boldsymbol{\dot q}}) = \\ \mathit{\boldsymbol{H}}(\mathit{\boldsymbol{q}},\mathit{\boldsymbol{\dot q}}) + {\mathit{\boldsymbol{J}}^{\rm{T}}}{(\mathit{\boldsymbol{J}}_{\rm{t}}^{\rm{T}})^ + }{\kern 1pt} {\kern 1pt} {\mathit{\boldsymbol{D}}_{\rm{t}}}\mathit{\boldsymbol{J}}_{\rm{t}}^ + (\mathit{\boldsymbol{\dot J}} - {{\mathit{\boldsymbol{\dot J}}}_{\rm{t}}}\mathit{\boldsymbol{J}}_{\rm{t}}^ + \mathit{\boldsymbol{J}})。\end{array} $

考虑到闭链混合体系统载体位置不受控,所得到的式(9)为欠驱动形式的动力学方程,不利于控制的设计。为将式(9)化为全驱动形式动力学方程,将其写成分块子矩阵形式:

$ \begin{array}{l} \left[ {\begin{array}{*{20}{l}} {{\mathit{\boldsymbol{D}}_{111}}}&{{\mathit{\boldsymbol{D}}_{112}}}\\ {{\mathit{\boldsymbol{D}}_{121}}}&{{\mathit{\boldsymbol{D}}_{122}}} \end{array}} \right]\left[ {\begin{array}{*{20}{l}} {{{\mathit{\boldsymbol{\ddot q}}}_0}}\\ {{{\mathit{\boldsymbol{\ddot \theta }}}_m}} \end{array}} \right] + \left[ {\begin{array}{*{20}{l}} {{\mathit{\boldsymbol{H}}_{111}}}&{{\mathit{\boldsymbol{H}}_{112}}}\\ {{\mathit{\boldsymbol{H}}_{112}}}&{{\mathit{\boldsymbol{H}}_{122}}} \end{array}} \right]\left[ {\begin{array}{*{20}{l}} {{{\mathit{\boldsymbol{\dot q}}}_0}}\\ {{{\mathit{\boldsymbol{\dot \theta }}}_m}} \end{array}} \right] = \\ {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} \left[ {\begin{array}{*{20}{c}} {{{\bf{0}}_{2 \times 1}}}\\ \mathit{\boldsymbol{\tau }} \end{array}} \right] - \left[ {\begin{array}{*{20}{c}} {\mathit{\boldsymbol{J}}_1^{\rm{T}}}\\ {\mathit{\boldsymbol{J}}_2^{\rm{T}}} \end{array}} \right]{\mathit{\boldsymbol{F}}_{\rm{I}}} \end{array} $ (10)

式中:Dl11R2×2Dl12R2×7Dl21R7×2Dl22R7×7Dl相应的子矩阵,且Hl11Hl21都为零矩阵[8],由式(10)第1行解得$\ddot{\boldsymbol{q}}$0表达式,并代入第2行,可得完全驱动形式动力学方程:

$ {\mathit{\boldsymbol{D}}_{\rm{C}}}{\mathit{\boldsymbol{\ddot \theta }}_{\rm{m}}} + {\mathit{\boldsymbol{H}}_{\rm{C}}}{\mathit{\boldsymbol{\dot \theta }}_{\rm{m}}} = \mathit{\boldsymbol{\tau }} - {\mathit{\boldsymbol{J}}_{\rm{C}}}{\mathit{\boldsymbol{F}}_{\rm{I}}} $ (11)

式中:DC=Dl22-Dl21Dl11-1Dl12HC=Hl22-Dl21Dl11-1Hl12JC=Dl21Dl11-1J1T-J2Tτ=τ1+JCFId,其中τ1为轨迹运动控制项;JCFId为内力控制项,其对系统运动不产生影响,FId为内力期望值,且为定义在JtT零空间项,有JtTFId=0。

同时,混合体系统完全驱动动力学方程满足如下结构特性:

特性1  DCHC满足一致有界性,即有:

$ 0{\rm{ }} < {\rm{ }}{D_{{\rm{cm}}}} \le \left\| {{\mathit{\boldsymbol{D}}_{\rm{C}}}} \right\| \le {D_{{\rm{cM}}}} $
$ \left\| {{\mathit{\boldsymbol{H}}_{\rm{C}}}({\mathit{\boldsymbol{\theta }}_{\rm{m}}},{{\mathit{\boldsymbol{\dot \theta }}}_{\rm{m}}})} \right\| \le {H_{{\rm{cM}}}}\left\| {\mathit{\boldsymbol{\dot \theta }}} \right\| $

式中DcmDcMHcM为已知正常数。

特性2  对于任意选取的zR7×1,满足:

$ \frac{1}{2}{\mathit{\boldsymbol{z}}^{\rm{T}}}{\mathit{\boldsymbol{\dot D}}_{\rm{C}}}z - {\mathit{\boldsymbol{z}}^{\rm{T}}}{\mathit{\boldsymbol{H}}_{\rm{C}}}z = 0 $

$\frac{1}{2} \dot{\boldsymbol{D}}_{\mathrm{C}}-\boldsymbol{H}_{\mathrm{C}}$满足斜对称性。

特性3  HC满足互换可加性,即任意x, y, zR7×1有:

$ {\mathit{\boldsymbol{H}}_{\rm{C}}}({\theta _m},x)y = {\mathit{\boldsymbol{H}}_{\rm{C}}}({\mathit{\boldsymbol{\theta }}_m},y)x $
$ {\mathit{\boldsymbol{H}}_{\rm{C}}}({\theta _m},z + ax)y = {\mathit{\boldsymbol{H}}_{\rm{C}}}({\theta _m},z)y + a{\mathit{\boldsymbol{H}}_{\rm{C}}}({\mathit{\boldsymbol{\theta }}_m},x)y $

式中a为常数。

2 基于无源性理论及速度观测器的模糊滑模控制方案设计

针对捕获完成后的闭链混合体系统,定义其轨迹及速度跟踪误差为:

$ \mathit{\boldsymbol{e}} = {\mathit{\boldsymbol{\theta }}_{\rm{m}}} - {\mathit{\boldsymbol{\theta }}_{{\rm{md}}}},\mathit{\boldsymbol{\dot e}} = {\mathit{\boldsymbol{\dot \theta }}_{\rm{m}}} - {\mathit{\boldsymbol{\dot \theta }}_{{\rm{md}}}} $ (12)

式中θmd$\dot{\boldsymbol{\theta}}_{\mathrm{md}}$为镇定运动期望轨迹及期望速度项。

定义滑模切换函数:

$ {\mathit{\boldsymbol{s}}_1} = \mathit{\boldsymbol{\dot e}} + \lambda \mathit{\boldsymbol{e}} $ (13)

式中λ=diag(λ1, λ2, λ3, λ4, λ5, λ6, λ7)。

因为捕获目标为非合作航天器,所以捕获后混合体系统的惯性参数无法精确获得,且参数的摄动是难以避免的。同时,考虑系统存在由捕获产生的有界扰动项τd,则式(11)可表示为:

$ {\mathit{\boldsymbol{\hat D}}_{\rm{C}}}{\mathit{\boldsymbol{\ddot \theta }}_{\rm{m}}} + {\mathit{\boldsymbol{\hat H}}_{\rm{C}}}{\mathit{\boldsymbol{\dot \theta }}_{\rm{m}}} = \mathit{\boldsymbol{\tau }} + \mathit{\boldsymbol{d}} - {\mathit{\boldsymbol{J}}_{\rm{C}}}{\mathit{\boldsymbol{F}}_{\rm{I}}} $ (14)

式中:$\hat{\boldsymbol{D}}_{\mathrm{C}}$$\hat{\boldsymbol{H}}_{\mathrm{C}}$为系统标称模型;ΔDC,ΔHC为系统建模误差。d=-(ΔDC$\ddot{\boldsymbol{\theta}}_{\mathrm{m}}$HC$\dot{\boldsymbol{\theta}}_{\mathrm{m}}$+τd),且满足‖d‖≤dd为总扰动项的上确界。

针对标称模型,设计控制器为:

$ \mathit{\boldsymbol{\tau }} = {\mathit{\boldsymbol{\hat D}}_{\rm{C}}}{\mathit{\boldsymbol{\ddot \theta }}_{{\rm{mr}}}} + {\mathit{\boldsymbol{\hat H}}_{\rm{C}}}{\mathit{\boldsymbol{\dot \theta }}_{{\rm{mr}}}} - {\mathit{\boldsymbol{K}}_{\rm{p}}}\mathit{\boldsymbol{e}} + \mathit{\boldsymbol{\nu + }}{\mathit{\boldsymbol{J}}_{\rm{C}}}{\mathit{\boldsymbol{F}}_{{\rm{Id}}}} $ (15)

式中$\dot{\boldsymbol{\theta}}_{\mathrm{mr}}=\dot{\boldsymbol{\theta}}_{\mathrm{md}}-\lambda \boldsymbol{e}$Kp为对角、正定矩阵。

将式(15)代入式(11),可得:

$ {\mathit{\boldsymbol{\hat D}}_{\rm{C}}}{\mathit{\boldsymbol{\dot s}}_1} + {\mathit{\boldsymbol{\hat H}}_C}{\mathit{\boldsymbol{s}}_1} + {\mathit{\boldsymbol{K}}_{\rm{p}}}\mathit{\boldsymbol{e}} = \mathit{\boldsymbol{\nu }} + \mathit{\boldsymbol{d}} + {\mathit{\boldsymbol{J}}_{\rm{C}}}{\mathit{\boldsymbol{F}}_{{\rm{Id}}}} - {\mathit{\boldsymbol{J}}_{\rm{C}}}{\mathit{\boldsymbol{F}}_{\rm{I}}} $ (16)

将式(16)两端左乘JtTJC+JC+JC伪逆,可得:

$ {\mathit{\boldsymbol{D}}_{\rm{X}}}{\mathit{\boldsymbol{\dot s}}_1} + {\mathit{\boldsymbol{H}}_{\rm{X}}}{\mathit{\boldsymbol{s}}_1} + {\mathit{\boldsymbol{K}}_{\rm{X}}}\mathit{\boldsymbol{e}} = \mathit{\boldsymbol{J}}_{\rm{t}}^{\rm{T}}\mathit{\boldsymbol{J}}_{\rm{C}}^ + (\mathit{\boldsymbol{\nu }} + \mathit{\boldsymbol{d}}) $ (17)

式中:DX=JtTJC+$\hat{\boldsymbol{D}}_{\mathrm{C}}$HX=JtTJC+$\hat{\boldsymbol{H}}_{\mathrm{C}}$KX=JtTJC+Kp

定义系统输出为:

$ \mathit{\boldsymbol{y}} = {\mathit{\boldsymbol{s}}_1} $ (18)

定义能量函数为:

$ {V_0} = \frac{1}{2}\mathit{\boldsymbol{s}}_1^{\rm{T}}{\mathit{\boldsymbol{D}}_{\rm{X}}}{\mathit{\boldsymbol{s}}_1} + \frac{1}{2}{\mathit{\boldsymbol{e}}^{\rm{T}}}{\mathit{\boldsymbol{K}}_{\rm{X}}}\mathit{\boldsymbol{e}} $ (19)

对式(19)求导:

$ {\dot V_0} = \mathit{\boldsymbol{s}}_1^T(\mathit{\boldsymbol{\nu }} + \mathit{\boldsymbol{d}}) - {\mathit{\boldsymbol{e}}^{\rm{T}}}\lambda {\mathit{\boldsymbol{K}}_{\rm{p}}}\mathit{\boldsymbol{e}} \le \mathit{\boldsymbol{s}}_1^{\rm{T}}(\mathit{\boldsymbol{\nu }} + \mathit{\boldsymbol{d}}) $ (20)

ν+d视为系统输入,则该系统满足传统的无源性控制理论,因而输入ν+d到输出θm是无源的。

在实际控制中,速度信号因为噪声干扰等原因,无法获得其精确值。因而,可设计速度观测器对其进行动态估计:

$ \left\{ \begin{array}{l} {{\mathit{\boldsymbol{\dot {\hat \theta }}}}_{\rm{m}}} = \mathit{\boldsymbol{Z}} + {\mathit{\boldsymbol{K}}_{\rm{d}}}{{\mathit{\boldsymbol{\tilde \theta }}}_{\rm{m}}}\\ \mathit{\boldsymbol{\dot Z}} = \mathit{\boldsymbol{\hat D}}_{\rm{C}}^{ - 1}(\mathit{\boldsymbol{\tau }} - {\mathit{\boldsymbol{J}}_{\rm{C}}}{\mathit{\boldsymbol{F}}_{\rm{I}}} - {{\mathit{\boldsymbol{\hat H}}}_{\rm{C}}}({\mathit{\boldsymbol{\theta }}_{\rm{m}}},{{\mathit{\boldsymbol{\dot \theta }}}_{{\rm{mo}}}}){{\mathit{\boldsymbol{\dot \theta }}}_{{\rm{mo}}}} + \\ {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\mathit{\boldsymbol{K}}_{{\rm{pl}}}}({{\mathit{\boldsymbol{\tilde \theta }}}_{\rm{m}}}) + {\mathit{\boldsymbol{K}}_{{\rm{p2}}}}{{\mathit{\boldsymbol{\tilde \theta }}}_{\rm{m}}} \end{array} \right. $ (21)

式中:$\dot{\hat{\boldsymbol{\theta}}}_{\mathrm{m}}$为速度估计项;$\tilde{\boldsymbol{\theta}}_{\mathrm{m}}=\boldsymbol{\theta}_{\mathrm{m}}-\hat{\boldsymbol{\theta}}_{\mathrm{m}}$KdR7×1Kp1R7×1Kp2R7×1为控制增益,并有:

$ {{\mathit{\boldsymbol{\dot \theta }}}_{{\rm{mo}}}} = {{\mathit{\boldsymbol{\dot {\hat \theta }}}}_{\rm{m}}} - {\mathit{\boldsymbol{\lambda }}_0}{{\mathit{\boldsymbol{\tilde \theta }}}_{\rm{m}}} $ (22)

式中λ0R7×1为对称、正定控制增益。

对式(21)进行如下假设:

假设1  KdKp1Kp2为正定、对角常值矩阵,且KdKp2满足:

$ {\mathit{\boldsymbol{K}}_{\rm{d}}} = {k_{\rm{d}}}\mathit{\boldsymbol{I}} + {\lambda _0},{\mathit{\boldsymbol{K}}_{{\rm{p2}}}} = {k_{\rm{d}}}{\lambda _0} $

式中kd为正常数。

假设2  速度信号是有界的,且满足:

$ {K_{\rm{Z}}} = \mathop {{\rm{sup}}}\limits_t \left\| {{{\mathit{\boldsymbol{\dot \theta }}}_{\rm{m}}}(t)} \right\| $

结合式(14)及式(21),可有:

$ \begin{array}{l} {{\mathit{\boldsymbol{\hat D}}}_{\rm{C}}}{{\mathit{\boldsymbol{\dot s}}}_2} + {{\mathit{\boldsymbol{\hat H}}}_{\rm{C}}}({\mathit{\boldsymbol{\theta }}_{\rm{m}}},{{\mathit{\boldsymbol{\dot \theta }}}_{\rm{m}}}){\mathit{\boldsymbol{s}}_2} + {k_{\rm{d}}}{{\mathit{\boldsymbol{\hat D}}}_{\rm{C}}}{\mathit{\boldsymbol{s}}_2} + {\mathit{\boldsymbol{K}}_{{\rm{p1}}}}{{\mathit{\boldsymbol{\tilde \theta }}}_{\rm{m}}} = \\ {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {{\mathit{\boldsymbol{\hat H}}}_{\rm{C}}}({\mathit{\boldsymbol{\theta }}_{\rm{m}}},{\mathit{\boldsymbol{s}}_2})({\mathit{\boldsymbol{s}}_2} - {{\mathit{\boldsymbol{\tilde \theta }}}_{\rm{m}}}) \end{array} $ (23)

式中$\boldsymbol{s}_{2}=\dot{\boldsymbol{\theta}}_{\mathrm{m}}-\dot{\boldsymbol{\theta}}_{\mathrm{mo}}=\dot{\tilde{\boldsymbol{\theta}}}_{\mathrm{m}}+\lambda_{0} \tilde{\boldsymbol{\theta}}_{\mathrm{m}}$

根据所设速度观测器,控制器改写为:

$ \mathit{\boldsymbol{\tau }} = {\mathit{\boldsymbol{\hat D}}_{\rm{C}}}{\mathit{\boldsymbol{\ddot \theta }}_{{\rm{mro}}}} + {\mathit{\boldsymbol{\hat H}}_{\rm{C}}}({\mathit{\boldsymbol{\theta }}_{\rm{m}}},{\mathit{\boldsymbol{\dot {\hat \theta }}}_{\rm{m}}}){\mathit{\boldsymbol{\dot \theta }}_{{\rm{mr}}}} - {\mathit{\boldsymbol{K}}_{\rm{p}}}\mathit{\boldsymbol{e}} + \mathit{\boldsymbol{\nu }} + {\mathit{\boldsymbol{J}}_{\rm{C}}}{\mathit{\boldsymbol{F}}_{{\rm{Id}}}} $ (24)

式中$\ddot{\boldsymbol{\theta}}_{\text {mro }}=\ddot{\boldsymbol{\theta}}_{\mathrm{md}}-\lambda\left(\dot{\hat{\boldsymbol{\theta}}}_{\mathrm{m}}-\dot{\boldsymbol{\theta}}_{\mathrm{md}}\right)$

定理1  存在kd>Dcm-1HcMKZ,使得观测器误差是局部指数收敛的,即存在a, b>0,使得:

$ {\left\| {\mathit{\boldsymbol{x}}(t)} \right\|^2} \le a{\mathit{\boldsymbol{e}}^{ - bt}}{\left\| {\mathit{\boldsymbol{x}}(0)} \right\|^2},t \ge 0 $ (25)

式中xT=[s2T  $\tilde{\boldsymbol{\theta}}_{\mathrm{m}}^{\mathrm{T}}$]。

此外,给出收敛吸引域:

$ A = \left\{ {x \in {{\bf{R}}^{14 \times 1}}|\left\| \mathit{\boldsymbol{x}} \right\|{\rm{ }} < ({k_{\rm{d}}}{D_{{\rm{cm}}}}H_{{\rm{cM}}}^{ - 1} - {K_{\rm{Z}}})\sqrt {\frac{{{B_{\rm{m}}}}}{{{B_{\rm{M}}}}}} } \right\} $

式中:Bm=min{Dcm, Kp1m};BM=max{DcM, Kp1M}。

证明  定义能量函数:

$ {V_{\rm{m}}}({\mathit{\boldsymbol{\tilde \theta }}_{\rm{m}}},{\mathit{\boldsymbol{s}}_2}) = \frac{1}{2}\mathit{\boldsymbol{s}}_2^{\rm{T}}{\mathit{\boldsymbol{\hat D}}_{\rm{C}}}{\mathit{\boldsymbol{s}}_2} + \frac{1}{2}\mathit{\boldsymbol{\tilde \theta }}_{\rm{m}}^{\rm{T}}{\mathit{\boldsymbol{K}}_{{\rm{p1}}}}{\mathit{\boldsymbol{\tilde \theta }}_{\rm{m}}} $ (26)

对式(27)求导,根据特性1及假设2,可有:

$ \begin{array}{l} {{\dot V}_{\rm{m}}}({{\mathit{\boldsymbol{\tilde \theta }}}_{\rm{m}}},{\mathit{\boldsymbol{s}}_2}) \le ({H_{{\rm{cM}}}}(\left\| {{\mathit{\boldsymbol{s}}_2}} \right\| + {K_Z}) - {k_{\rm{d}}}{D_{{\rm{cm}}}}){\left\| {{\mathit{\boldsymbol{s}}_2}} \right\|^2} - \\ {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\lambda _{{\rm{0m}}}}{{\rm{K}}_{{\rm{p1m}}}}{\left\| {{{\mathit{\boldsymbol{\tilde \theta }}}_{\rm{m}}}} \right\|^2} \end{array} $ (27)

如果符合条件:

$ |\left\| {{\mathit{\boldsymbol{s}}_2}} \right\|{\rm{ }} < {k_{\rm{d}}}{D_{{\rm{cm}}}}H_{{\rm{cM}}}^{ - 1} - {\mathit{\boldsymbol{K}}_{\rm{Z}}} $ (28)

则存在一个常数ε>0,使得式(27)满足:

$ {\dot V_{\rm{m}}}({\mathit{\boldsymbol{\tilde \theta }}_{\rm{m}}},{\mathit{\boldsymbol{s}}_2}) \le - \varepsilon {\left\| \mathit{\boldsymbol{x}} \right\|^2} $ (29)

结合式(28)及式(29),可证明(25)的正确性。

即表示如果有:

$ \left\| {\mathit{\boldsymbol{x}}(0)} \right\|{\rm{ }} < ({k_{\rm{d}}}{\mathit{\boldsymbol{D}}_{{\rm{cm}}}}\mathit{\boldsymbol{H}}_{{\rm{cM}}}^{ - 1} - {\mathit{\boldsymbol{K}}_{\rm{Z}}})\sqrt {\frac{{{B_{\rm{m}}}}}{{{B_{\rm{M}}}}}} $ (30)

同时,结合式(30)、(31),可得:

$ {{V_{\rm{m}}}(\mathit{\boldsymbol{x}}(t)) \le {V_{\rm{m}}}(\mathit{\boldsymbol{x}}(0))\quad t \ge 0} $ (31)
$ {{{\dot V}_{\rm{m}}}(\mathit{\boldsymbol{x}}(t)) \le - e{{\left\| {\mathit{\boldsymbol{x}}(t)} \right\|}^2}\quad t \ge 0} $ (32)

对定理1证毕。

显然,x的指数收敛性意味着观测器状态误差$\mathit{\boldsymbol{y}}_0^{\rm{T}} = \left[ {\mathit{\boldsymbol{\dot {\tilde \theta} }}_{\rm{m}}^{\rm{T}}\;\;\;\;\;\;\mathit{\boldsymbol{\tilde \theta }}_{\rm{m}}^{\rm{T}}} \right]$具有指数收敛性,这是因为xy0是线性相关的。因而,证明了速度观测器的稳定性。

由式(24)可知,所设计控制器轨迹运动控制项τ1包含轨迹误差反馈控制部分及总扰动项补偿部分,所设计滑模变结构控制项ν即为总扰动补偿项,其设计为:

$ \mathit{\boldsymbol{\nu }} = - k{\rm{sgn}} ({\mathit{\boldsymbol{s}}_1}) $ (33)

式中:增益k=d;sgn(s1)={sgn(s1(1));sgn(s1(2));sgn(s1(3));sgn(s1(4));sgn(s1(5));sgn(s1(6));sgn(s1(7))}。为保证滑模控制条件成立,需要选取较大的控制增益,但这会使得系统控制产生抖振,不利于控制方案在捕获后镇定控制的应用。因此,引入模糊控制系统来权衡补偿项增益k的系数,以实现有效抑制抖振的目的。

根据变结构控制原理,对式(24)所设计控制器采用如下规则:

$ {\rm{if }}({\mathit{\boldsymbol{s}}_1}(t){\rm{is}}{\kern 1pt} {\kern 1pt} {Z_0}){\kern 1pt} {\kern 1pt} {\kern 1pt} {\rm{then}}{\kern 1pt} {\kern 1pt} {\kern 1pt} (\mathit{\boldsymbol{\tau }}{\kern 1pt} {\kern 1pt} {\rm{is}}{\kern 1pt} {\kern 1pt} {\kern 1pt} {\mathit{\boldsymbol{\tau }}_0}) $
$ {\rm{if }}({\mathit{\boldsymbol{s}}_1}(t){\rm{is}}{\kern 1pt} {\kern 1pt} {N_Z}){\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\rm{then}}{\kern 1pt} {\kern 1pt} {\kern 1pt} (\mathit{\boldsymbol{\tau }}{\kern 1pt} {\kern 1pt} {\rm{is}}{\kern 1pt} {\kern 1pt} {\kern 1pt} {\mathit{\boldsymbol{\tau }}_0} + \mathit{\boldsymbol{\nu }}) $

式中:$\boldsymbol{\tau}_{0}=\hat{\boldsymbol{D}}_{\mathrm{C}} \ddot{\boldsymbol{\theta}}_{\mathrm{mro}}+\hat{\boldsymbol{H}}_{\mathrm{C}}\left(\boldsymbol{\theta}_{\mathrm{m}}, \dot{\hat{\boldsymbol{\theta}}}_{\mathrm{m}}\right) \dot{\boldsymbol{\theta}}_{\mathrm{mr}}-\boldsymbol{K}_{\mathrm{p}} \boldsymbol{e}+\boldsymbol{J}_{\mathrm{C}} \boldsymbol{F}_{\mathrm{Id}}$Z0NZ各表示“零”与“非零”。

上述规则意味着,如果s1(t)=0时,控制器输出为τ0;若s1(t)≠0时,控制器输出为τ0+ν

根据以上设计,控制器输出可表示为:

$ \mathit{\boldsymbol{\tau }} = \frac{{{k_{{{\rm{Z}}_0}}}({s_1}){\mathit{\boldsymbol{\tau }}_0} + {k_{{N_Z}}}({s_1})({\mathit{\boldsymbol{\tau }}_0} + \mathit{\boldsymbol{\nu }})}}{{{k_{{Z_0}}}({\mathit{\boldsymbol{s}}_1}) + {k_{{N_Z}}}({\mathit{\boldsymbol{s}}_1})}} = {\mathit{\boldsymbol{\tau }}_0} + {k_{{N_Z}}}({\mathit{\boldsymbol{s}}_1})\mathit{\boldsymbol{\nu }} $ (34)
$ {{k_{Z0}}({\mathit{\boldsymbol{s}}_1}) + {k_{{N_Z}}}({\mathit{\boldsymbol{s}}_1}) = 1} $ (35)

式中:kZ0(s1)及kNZ(s1)为对应隶属度函数,通过调节kNZ(s1)的大小,以抑制系统的抖振。

将模糊控制器设计为以s1(t)为输入,增益k的系数kNZ(s1)为输出的控制单元。定义3个模糊规则语言词集分别为{PZN}={正,零,负},利用图 2所示模糊规则对输入、输出进行模糊化处理,设计模糊推理规则:

$ {\rm{if(}}{\mathit{\boldsymbol{s}}_{\rm{1}}}{\rm{(}}t{\rm{)is}}{\kern 1pt} N{\rm{)then(}}{k_{{N_Z}}}{\rm{(}}{\mathit{\boldsymbol{s}}_{\rm{1}}}{\rm{)is}}P{\rm{)}} $
$ {\rm{if(}}{\mathit{\boldsymbol{s}}_{\rm{1}}}{\rm{(}}t{\rm{)is}}{\kern 1pt} Z{\rm{)then(}}{k_{{N_Z}}}{\rm{(}}{\mathit{\boldsymbol{s}}_{\rm{1}}}{\rm{)is}}Z{\rm{)}} $
$ {\rm{if(}}{\mathit{\boldsymbol{s}}_{\rm{1}}}{\rm{(}}t{\rm{)is}}{\kern 1pt} P{\rm{)then(}}{k_{{N_Z}}}{\rm{(}}{\mathit{\boldsymbol{s}}_{\rm{1}}}{\rm{)is}}P{\rm{)}} $
Download:
图 2 隶属度函数 Fig. 2 Membership function

由于在模糊推理中采用了Mamdani的最大最小合成法,因此采用常用的面积重心法进行去模糊化处理,进而得到kNZ(s1)的取值。

基于上述分析可知,所设计算法中‖s1‖,‖s2‖是有界的,所以误差e$\dot{\tilde{\boldsymbol{\theta}}}_{\mathrm{m}}$有界,即系统渐近稳定。在保证混合体系统位置控制的稳定性的同时,也需考虑对目标夹持内力协调控制的有界性。将式(24)、(34)代入式(14),可有:

$ \begin{array}{l} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\mathit{\boldsymbol{J}}_{\rm{C}}}({\mathit{\boldsymbol{F}}_{{\rm{Id}}}} - {\mathit{\boldsymbol{F}}_{\rm{I}}}) = {{\mathit{\boldsymbol{\hat D}}}_{\rm{C}}}({{\mathit{\boldsymbol{\dot s}}}_1} - \lambda {{\mathit{\boldsymbol{\dot {\tilde \theta }}}}_{\rm{m}}}) + \\ {{\mathit{\boldsymbol{\hat H}}}_{\rm{C}}}({\mathit{\boldsymbol{\theta }}_{\rm{m}}},{{\mathit{\boldsymbol{\dot \theta }}}_{\rm{m}}})({\mathit{\boldsymbol{s}}_1} - {{\mathit{\boldsymbol{\dot {\tilde \theta }}}}_{\rm{m}}}) + {{\mathit{\boldsymbol{\hat H}}}_{\rm{C}}}({\mathit{\boldsymbol{\theta }}_{\rm{m}}},{{\mathit{\boldsymbol{\dot {\tilde \theta }}}}_{\rm{m}}}){\mathit{\boldsymbol{s}}_1} + \\ {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\mathit{\boldsymbol{K}}_{\rm{p}}}\mathit{\boldsymbol{e}} + ({k_{NZ}} \bar d{\rm{sgn}} ({\mathit{\boldsymbol{s}}_1}) - \mathit{\boldsymbol{d}}) \end{array} $ (36)

式(36)两端左乘JC+,可得:

$ \begin{array}{l} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} ({\mathit{\boldsymbol{F}}_{{\rm{Id}}}} - {\mathit{\boldsymbol{F}}_{\rm{I}}}) = \mathit{\boldsymbol{J}}_{\rm{C}}^ + \mathit{\boldsymbol{(}}{{\mathit{\boldsymbol{\hat D}}}_{\rm{C}}}({{\mathit{\boldsymbol{\dot s}}}_1} - \lambda {{\mathit{\boldsymbol{\dot {\tilde \theta }}}}_{\rm{m}}}) + \\ {{\mathit{\boldsymbol{\hat H}}}_{\rm{C}}}({\mathit{\boldsymbol{\theta }}_{\rm{m}}},{{\mathit{\boldsymbol{\dot \theta }}}_{\rm{m}}})({\mathit{\boldsymbol{s}}_1} - {{\mathit{\boldsymbol{\dot {\tilde \theta} }}}_{\rm{m}}}) + {{\mathit{\boldsymbol{\hat H}}}_{\rm{C}}}({\mathit{\boldsymbol{\theta }}_{\rm{m}}},{{\mathit{\boldsymbol{\dot {\tilde \theta }}}}_{\rm{m}}}){\mathit{\boldsymbol{s}}_1} + \\ {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\mathit{\boldsymbol{K}}_{\rm{p}}}\mathit{\boldsymbol{e}} + ({k_{NZ}} \bar d{\rm{sgn}} ({\mathit{\boldsymbol{s}}_1}) - \mathit{\boldsymbol{d}})) \end{array} $ (37)

对式(37)两端求范数,并化简得:

$ \begin{array}{l} {\kern 1pt} {\kern 1pt} {\kern 1pt} \left\| {{\mathit{\boldsymbol{F}}_{{\rm{Id}}}} - {\mathit{\boldsymbol{F}}_{\rm{I}}}} \right\| \le \left\| {\mathit{\boldsymbol{J}}_{\rm{C}}^ + } \right\|({D_{{\rm{cM}}}}(\left\| {{\mathit{\boldsymbol{s}}_1}} \right\| - \lambda \left\| {{{\mathit{\boldsymbol{\dot {\tilde \theta} }}}_{\rm{m}}}} \right\| + \\ {H_{{\rm{cM}}}}(\left\| {{\mathit{\boldsymbol{s}}_1}} \right\| - \left\| {{{\mathit{\boldsymbol{\dot {\tilde \theta }}}}_{\rm{m}}}} \right\|) + {H_{{\rm{cM}}}}\left\| {{\mathit{\boldsymbol{s}}_1}} \right\| + \left\| {{\mathit{\boldsymbol{K}}_{\rm{p}}}} \right\|\left\| \mathit{\boldsymbol{e}} \right\| + \\ {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {k_{{N_Z}}}\bar d\left\| {{\rm{sgn}} ({\mathit{\boldsymbol{s}}_1})} \right\| - \left\| \mathit{\boldsymbol{d}} \right\|) \end{array} $ (38)

因为JC满足特性1,故其有界,且$\dot{\boldsymbol{s}}_1$s1$\dot{\tilde{\boldsymbol{\theta}}}_{\mathrm{m}}$e满足有界性,所以夹持内力的误差值也是有界的,因而对目标的夹持内力控制是稳定的。

3 空间机器人双臂夹持捕获操作仿真算例

图 1所示的双臂空间机器人系统及目标航天器系统为例,采用本文所提控制算法进行数值仿真试验。系统结构参数选取如下:载体质量、转动惯量及O0O1长度分别为m0=200 kg,I0=50 kg·m2d0=1.064 m;刚性连杆Bi(i=1, 2, 4, 5)的质量、转动惯量及杆长分别为mi=20 kg(i=1, 2, 4, 5),Ii=10 kg·m2(i=1, 2, 4, 5),li=2 m(i=1, 2, 4, 5);刚性连杆Bi(i=3, 6)的质量、转动惯量及杆长分别为mi=5 kg(i=3, 6),Ii=2 kg·m2(i=3, 6),li=0.5 m(i=3, 6);目标航天器的质量、转动惯量及其质心至两端捕获点长度分别为mt=50 kg,It=10 kg·m2lL=0.5 m,lR=0.5 m,lB=2 m。

假设捕获前双臂空间机器人系统初始构型为:

$ \begin{array}{l} \mathit{\boldsymbol{q}} = [\begin{array}{*{20}{l}} 0&0&{{0^{\rm{^\circ }}}}&{ - {{120}^{\rm{^\circ }}}}&{ - {{60}^{\rm{^\circ }}}}&{ - {{60}^{\rm{^\circ }}}}&{{{60}^{\rm{^\circ }}}} \end{array}\\ {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} \begin{array}{*{20}{c}} {{{60}^{\rm{^\circ }}}}&{{{60}^{\rm{^\circ }}}} \end{array}{]^{\rm{T}}} \end{array} $

t0=0 s时刻,被捕获目标航天器以速度vt=[0.1 m/s0.1 m/s0.35 rad/s]T与双臂空间机器人系统发生接触、碰撞,经过极小的时间Δt后捕获完成。为保护关节电机,在发生碰撞时关节电机关闭,即τ=07,经过较短的反应时间后,开启主动镇定控制,选取镇定控制期望位形为θmd=[0°-120°-60°-60°60°60°60°]T。仿真时间选取为t=15 s。

针对控制设计相关假设,选取如下控制参数:

$ \mathit{\boldsymbol{\lambda }} = {\rm{diag}}\left( {5,5,5,5,5,5,5} \right),{k_{\rm{d}}} = 260, $
$ \begin{array}{*{20}{l}} {k{\rm{ }} = 150,{\mathit{\boldsymbol{K}}_{\rm{p}}} = {\rm{diag}}\left( {20,20,20,20,20,20,20} \right),} \end{array} $
$ {\mathit{\boldsymbol{\lambda }}_0} = {\rm{diag}}\left( {0.{\rm{ }}2,0.{\rm{ }}2,0.{\rm{ }}2,0.{\rm{ }}2,0.{\rm{ }}2,0.{\rm{ }}2,0.{\rm{ }}2} \right)。$

由捕获产生的系统扰动及标称模型分别为:τd=[3.5sin(πt/4)3.5cos(πt/4)3.5sin(πt/4)3.5cos(πt/4)3.5sin(πt/4)3.5cos(πt/4)3.5sin(πt/4)]T, $\hat{\boldsymbol{D}}_{\mathrm{C}}$=0.8DC$\hat{\boldsymbol{H}}_{\mathrm{C}}$=0.8HC。选取期望内力项为:FId=[-10 N  0 N  0 N·m  10 N  0 N  0 N·m]T。所得仿真结果如图 3~7所示。

Download:
图 3 载体姿态轨迹跟踪情况 Fig. 3 Tracking trajectory of Base attitude
Download:
图 4 左臂各关节轨迹跟踪情况 Fig. 4 Tracking trajectory of joints in left arm
Download:
图 5 右臂各关节轨迹跟踪情况 Fig. 5 Tracking trajectory of joints in right arm
Download:
图 6 左端夹持内力跟踪情况 Fig. 6 Tracking of clamping internal force in left gripper
Download:
图 7 右端夹持内力跟踪情况 Fig. 7 Tracking of clamping internal force in right gripper

图 3~5为载体姿态及左右臂各关节姿态运动情况,可发现夹持捕获后未开启主动控制的时间内,闭链混合体系统受捕获碰撞影响为失稳状态,若不开启主动控制,系统将产生翻滚现象,其将影响捕获任务的进行,严重时甚至会损坏相关组件。为使失稳系统镇定,开启本文所提基于无源性理论的滑模控制算法。通过轨迹跟踪情况可发现,由于开启模糊控制后,系统可根据实时输出来有效调节滑模补偿项的控制增益,进而保证了镇定运动的快速响应而又减少了滑模控制固有的抖振现象。通过图 3~5可知,开启模糊控制后,闭链混合体系统的镇定运动达到稳定状态。图 67为左右臂夹持内力的跟踪情况,从仿真结果可知夹持内力最后达到期望、稳定状态,进而实现了力/位置的协调控制。

4 结论

1) 本文设计了失稳闭链混合体系统镇定运动的力/位置模糊滑模控制方案。所提基于无源性理论的控制方案,不仅克服了建模误差及外部扰动项的影响,还实现了对传统滑模控制抖振现象的抑制,达到了镇定运动的稳定控制与精确跟踪。

2) 本文为理论性探索研究,当仿真试验及计算机硬件条件达到一定条件后,可为航天实际操作、应用提供技术参考。同时,上述系统经过适当扩充,可推广应用于三维运动的空间机器人系统。

参考文献
[1]
FLORES-ABAD A, MA Ou, PHAM K, et al. A review of space robotics technologies for on-orbit servicing[J]. Progress in aerospace sciences, 2014, 68: 1-26. DOI:10.1016/j.paerosci.2014.03.002 (0)
[2]
刘阳, 谢宗武, 王滨, 等. 自由漂浮空间机器人视觉伺服系统研究[J]. 哈尔滨工程大学学报, 2017, 38(2): 153-159.
LIU Yang, XIE Zongwu, WANG Bin, et al. Research on the visual servo system of a free-floating space robot[J]. Journal of Harbin Engineering University, 2017, 38(2): 153-159. (0)
[3]
DONG Hongyang, HU Qinglei, AKELLA M R. Dual-quaternion-based spacecraft autonomous rendezvous and docking under six-degree-of-freedom motion constraints[J]. Journal of guidance, control, and dynamics, 2018, 41(5): 1150-1162. DOI:10.2514/1.G003094 (0)
[4]
VIRGILI-LLOP J, ZAGARIS C, ZAPPULLA II R, et al. A convex-programming-based guidance algorithm to capture a tumbling object on orbit using a spacecraft equipped with a robotic manipulator[J]. The international journal of robotics research, 2019, 38(1): 40-72. DOI:10.1177/0278364918804660 (0)
[5]
WANG Congqing, WU Pengfei, ZHOU Xin, et al. Composite sliding mode control for a free-floating space rigid-flexible coupling manipulator system[J]. International journal of advanced robotic systems, 2013, 10(2): 124. DOI:10.5772/54640 (0)
[6]
LIU Xiaofeng, LI Haiquan, WANG Jingsen, et al. Dynamics analysis of flexible space robot with joint friction[J]. Aerospace science and technology, 2015, 47: 164-176. DOI:10.1016/j.ast.2015.09.030 (0)
[7]
YU Xiaoyan, CHEN Li. Modeling and observer-based augmented adaptive control of flexible-joint free-floating space manipulators[J]. Acta astronautica, 2015, 108: 146-155. DOI:10.1016/j.actaastro.2014.12.002 (0)
[8]
梁捷, 陈力. 执行器受限空间机器人的模糊神经网络控制[J]. 工程力学, 2014, 31(11): 190-197.
LIANG Jie, CHEN Li. Fuzzy neural network control for a space-based robot with constrained actuators[J]. Engineering mechanics, 2014, 31(11): 190-197. DOI:10.6052/j.issn.1000-4750.2012.11.0840 (0)
[9]
REKLEITIS G, PAPADOPOULOS E. On-orbit cooperating space robotic servicers handling a passive object[J]. IEEE transactions on aerospace and electronic systems, 2015, 51(2): 802-814. DOI:10.1109/TAES.2014.130584 (0)
[10]
魏承, 赵阳, 田浩. 空间机器人捕获漂浮目标的抓取控制[J]. 航空学报, 2010, 31(3): 632-637.
WEI Cheng, ZHAO Yang, TIAN Hao. Grasping control of space robot for capturing floating target[J]. Acta aeronautica et astronautica sinica, 2010, 31(3): 632-637. (0)
[11]
HUANG Panfeng, WANG Ming, MENG Zhongjie, et al. Reconfigurable spacecraft attitude takeover control in post-capture of target by space manipulators[J]. Journal of the franklin institute, 2016, 353(9): 1985-2008. DOI:10.1016/j.jfranklin.2016.03.011 (0)
[12]
GASBARRI P, PISCULLI A. Dynamic/control interactions between flexible orbiting space-robot during grasping, docking and post-docking manoeuvres[J]. Acta astronautica, 2015, 110: 225-238. DOI:10.1016/j.actaastro.2015.01.024 (0)
[13]
ZHANG Bo, LIANG Bin, WANG Xueqian, et al. Manipulability measure of dual-arm space robot and its application to design an optimal configuration[J]. Acta astronautica, 2016, 128: 322-329. DOI:10.1016/j.actaastro.2016.07.040 (0)
[14]
CHEN Gang, WANG Yuqi, WANG Yifan, et al. Detumbling strategy based on friction control of dual-arm space robot for capturing tumbling target[J]. Chinese journal of aeronautics, 2020, 33(3): 1093-1106. DOI:10.1016/j.cja.2019.04.019 (0)
[15]
王从庆, 柯昌俊. 自由浮动柔性空间双臂机器人的混合抑振控制[J]. 振动工程学报, 2009, 22(1): 31-35.
WANG Congqing, KE Changjun. Hybrid vibration suppression for a free-floating flexible dual-arm space robot[J]. Journal of vibration engineering, 2009, 22(1): 31-35. DOI:10.3969/j.issn.1004-4523.2009.01.006 (0)
[16]
程靖, 陈力. 双臂空间机器人捕获航天器后的镇定运动分块滑模自适应神经网络控制[J]. 中国机械工程, 2017, 28(12): 1427-1433, 1441.
CHENG Jing, CHEN Li. Partitioned sliding mode adaptive neural network control of calm movements of dual-arm space robot after capturing a spacecraft[J]. China mechanical engineering, 2017, 28(12): 1427-1433, 1441. DOI:10.3969/j.issn.1004-132X.2017.12.007 (0)
[17]
NGUYEN-HUYNH T C, SHARF I. Adaptive reactionless motion and parameter identification in postcapture of space debris[J]. Journal of guidance, control, and dynamics, 2013, 36(2): 404-414. DOI:10.2514/1.57856 (0)
[18]
BANDYOPADHYAY S, CHUNG S J, HADAEGH F Y. Nonlinear attitude control of spacecraft with a large captured object[J]. Journal of guidance, control, and dynamics, 2016, 39(4): 754-769. DOI:10.2514/1.G001341 (0)
[19]
ABIKO S, LAMPARIELLO R, HIRZINGER G. Impedance control for a free-floating robot in the grasping of a tumbling target with parameter uncertainty[C]//Proceedings of 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems. Beijing, China, 2006: 1120-1125. (0)
[20]
林雷, 王洪瑞, 任华彬. 基于模糊变结构的机械臂控制[J]. 控制理论与应用, 2007, 24(4): 643-645.
LIN Lei, WANG Hongrui, REN Huabin. Fuzzy-based variable structure control for robotic manipulators[J]. Control theory & applications, 2007, 24(4): 643-645. DOI:10.3969/j.issn.1000-8152.2007.04.025 (0)
[21]
MULERO-MARTINEZ J I. Robust GRBF static neurocontroller with switch logic for control of robot manipulators[J]. IEEE transactions on neural networks and learning systems, 2012, 23(7): 1053-1064. DOI:10.1109/TNNLS.2012.2196053 (0)
[22]
WAI R J, MUTHUSAMY R. Fuzzy-Neural-Network Inherited Sliding-Mode control for robot manipulator including actuator dynamics[J]. IEEE transactions on neural networks and learning systems, 2013, 24(2): 274-287. DOI:10.1109/TNNLS.2012.2228230 (0)
[23]
KRISTIANSEN R, NICKLASSON P J, GRAVDAHL J T. Spacecraft coordination control in 6DOF:integrator backstepping vs passivity-based control[J]. Automatica, 2008, 44(11): 2896-2901. DOI:10.1016/j.automatica.2008.04.019 (0)
[24]
REN Chao, MA Shugen. Passivity-based model free control of an omnidirectional mobile robot[C]//Proceedings of 2015 IEEE International Conference on Mechatronics. Nagoya, Japan, 2015: 262-267. (0)