Path Planning And Experiment Study Of Space Robot For Target Capturing

With the development of space exploration, space robotic systems are expected to play an increasingly important role in the future. The technology of robotic on-orbital service, whose key is the capturing technology, becomes research hot in recent years. Therefore, this thesis will focus on resolving some challenging problems about target capturing, which include Cartesian Path Planning, Autonomous Path Planning for Target Capturing and Path Planning for Target Berthing and Reorientation after Capturing. And it is required to set up the ground experiment system to verify and evaluate the corresponding approach. The main contributions of the thesis are as follows:Firstly, point-to-point and continuous Cartesian path planning are studied. The approach for the point-to-point path planning only uses the direct kinematic equations, so it is not affected by the dynamic singularity. The joint trajectories are parameterized by polynomial fuctions. Then, the joint functions are normalized and the system of equations about the parameters is established by integrating the differential kinematics equations. Lastly, the parametes are solved by the iterative Newtonian algorithm. The method settles the shortcomings of similar approaches, i.e., it satisfies the limits on the manipulator and it supplies a general criterion to assign the initial guess of the unknown parameters. For the continuous Cartesian path tracking, a method is proposed to plan the joint path to control six arbitrary variables of the nine: satellite attitude, end-effector position and end-effector orientation. Based on the differential kinematical equations and the momentum conservation equations, a numeric algorithm is presented to plan the motion of the space manipulator. Five cases are studied: a)Continuous pose tracking; b)Continuous position tracking without disturbance on the base; c)Continuous orientation tracking without disturbance on the base; d) Continuous position tracking and adjusting the base attitude; e) Continuous orientation tracking and adjusting the base attitude. The method is of theoretic and practical significance.Continuous Cartesian path tracking will be affected by the dynamic singularity, and existing methods to avoide it are complex. A practical approach is proposed to avoid the dynamic singularity, which transforms it into real-time kinematic singularity avoiding problem. Firstly, according to the characteristic of the PUMA manipulator, the method"Singularity Condition Separation Plus Damped Reciprocal"is presented, which separates the singularity parameters from the inverse of the Jacobian matrix, and replaces the reciprocal of the singularity parameters by the Damped Reciprocal. Since the SVD decomposition and the estimate of the minimum singularity value are not required, the algorithm is more efficient. And not all of the velocity components of the end-effector are sacrificed in accuracy, i. e. the effect of the singularity is minimized. Combining the results with the momentum conservation laws, the practical singularity avoiding algorithm for space robot is proposed.The autonomous path planning for capturing target is also studied. Two algorithms, i.e. the position-based and image-based algorithms are proposed. The term autonomous means that"to predict the target motion autonomously","to avoid the singularity autonomously"and"to adjust the manipulator motion without excess disturbance on the base autonomously". For the position-based algorithm, the task is described in Cartesian space and it can drive the manipulator to approach the target along the closest path. But the accuracy is largely affected by the hand-eye calibration error. For the image-based algorithm, it is not sensitive to the calibration error and it avoids the complex 3D reconstruction process. But it needs to compute the image Jacobian matrix on real time. The singularity of the image Jacobian matrix will disable the algorithm. Therefore, the ideal case is to combine the two algorithms in the application.After capturing, the attitude of the coupled pair inclines. It is required to re-orientate the system and to berthing the target in the prescribed pose for the on-orbital service. A non-holonomic path planning based on Genetic Algorithm is presented, which is used to realize the"target berthing"and"attitude re-orientation"at the same time. The method has the following traits: 1)The kinematic and dynamic constraints of the manipulator are considered in the planning process; 2) The dynamic singular doesn't affect the algorithm; 3) The planned path is very smooth, and so on.In order to verify the path planning methods, the author combines the dynamic simulation with the kinematic equavallence, and proposes two realization modes of the ground experiments. Then an experiment system is set up using cheap devices. The system can emulate the capturing process observed from the space base or the inertia frame. And the geometry and mass properties of space robot are not limited. The system can be extended by small modification to verify different technology of target capturing. The real-time 3D simulation system visualizes the capturing process intuitively. The path planning methods proposed above are tested and verified using the experiment system.