Research On Control Method Of Space Manipulator Capturing A Malfunctioned Tumbling Object

In recent decades,with increasing number of malfunctioned space-crafts,on-orbit crafts' safety is confronted with more critical threat.It has great potential to clear orbit by space robotic manipulators.Most malfunctioned crafts are in the nutational motion,and with lack of specific capture interface and targets for visual measurement.Capturing such tumbling satellite of large inertial is a challenging task for the manipulator's controller.In this thesis,my work focuses on impedance control,dynamic parameter identification,state observation,disturbance suppressing and motion planning for a space manipulator capturing the malfunctioned tumbling space-crafts.A key factor to capture a malfunctioned target with large inertia is adjusting the stiffness and damping of the space manipulator.The toal capture process is consisted of three stages: pre-capture stage,capture stage and stabilization of the coupling system,in which different stiffness and damping of the space manipulator are expectd.In the pre-capture stage,the space manipulator should exhibit high stiffness to ensure high-precision target-tracking,while in rest stages,the stiffness and inject joint damping should be reduced to dissipate the energy caused by rigid capture,and smoothing the redistribution process of momentum among the coupling system.Impedance control is an idel solution treating such contradiction between high-stiffness servo and low-stiffness capture.The space manipulator is driven through lightweight harmonic drivers and equipped with joint torque sensors.The impact of joint elasticity on the performance of the impedance controller is non-negligible.Meanwhile,joint angular error is easy to be amplified by long arm,so requirements for the controller become more stringent given the same requirement w.r.t.end-effector accuracy.With flexible joint manipulator model,a space manipulator's Cartesian impedance control method is proposed,which is composed of a motion inner loop stabilizing flexible joint torque and an impedance outer loop adjusting stiffness and damping characteristics.The motion inner loop eliminates the elastic tremor caused by the flexible transmission in the motion,also reshapes the flexible joint manipulator into a rigid manipulator.The Cartesian impedance outer loop establishes the task space position impedance equation and the attitude impedance equation.The desired trajectory is shaped given set desired impedance and the end force feedback,then the compliant trajectory is tracked by the motion inner loop to realize end-tip stiffness and damping adjustment.The performance of the impedance controller depends on the accuracy of external-force measurement.To achieve high-precision motion control of the impedance controller in free space,the inertial moment and Coriolis moment should be removed from joint torque sensor signals,so accurate identification of the dynamic parameters of the manipulator is expected.A parameter identification method based on the principle of immersion and invariance is proposed.The parameter estimation correction term is introduced into the parameter update law,the parameter estimation error is used as the manifold,and the manifold is invariant and attractive as the condition to construct the correction term.Compared with classic adaptive control methds,based on the certainty-equivalence principle,the proposed parameter identification method can obtain more accurate dynamic parameters.It is quite important to realize the unification of high-precision motion control and low-stiffness compliance control of the impedance controller.In addition to the uncertainty of dynamic parameters,a complete impedance controller must achieving motion state observation and disturbance repressing.Currently,just motor position and joint position sensors are intergrated in the joints,and the torque sensor has a measurement noise of about ±2Nm.In this paper,a Kalman observer based on the Brownian motion model is designed,in which joint torque with high-order information and motor speed are estimated simultaneously,then the joint states are obtained.This solves the contradiction between the limited sensor configuration of the manipulator and the high-precision control.In addition,to reject the nonlinear friction disturbance inside motors,a sliding-mode observer is introduced in the motion inner loop to compensate for the viscous friction and Coulomb friction,realizing the robustness to friction disturbance.The drawback of the control-signal tremors of classic slide-mode method is overcomed.The impedance controller realizes the space manipulator good control performance,and a motion planning method based on visual feedback is proposed providing motion instructions to approach and track the tumbling target.The method is consisted of two parts: coarse servo and fine servo.During coarse servo stage,the visual measurement feedback acts as a flag.Through the acceleration-deceleration adjustment of the end speed of the manipulator,achieving the manipulator smoothly approaching the target in the case of long-distance measurement with large noise.Fine servo is resposible to eliminate the initial deviation between the end-effector and the target within the set time,and achieves accurate tracking of the moving target.Finally,an experimental system are built to simulate on-orbit tasks.On the three-dimensional hardware-in-loop semi-physical system,a space robot visual tracking non-cooperative tumbling target experiment was carried out to verify the effectiveness of the proposed planning algorithm and the motion control performance of the impedance controller.On the two-dimensional microgravity system,experiments of space manipulators capturing large-inertia moving targets have been carried out to verify the compliance and force-control performance of the impedance control method.The work introduced in this thesis is oriented to capture the non-cooperative tumbling target with space manipulator,and also can be implemented in other application,e.g.on-orbit tasks of the space manipulator,industrial robotic arm planning,motion and force control etc.