Parameter Identification Of Target In Space Robot Grasping Process

The quality of space robot changes with payload and fuel consumption, and inertia parameters of non-cooperative target are unknown. Parameter identification of target is necessary for precise dynamics control in space robot grasping unknown target. With free-flying space robot as objects, this paper is concerned with target inertia parameters identification in space robot grasping unknown target.Firstly, kinematic analysis of free-flying space robot is advanced, and dynamics models of space rigid robot utilizing Lagrange equations and dynamics models of target applying Newton-Euler equations are established. Target grasping area models of different end-effectors are established based on the definition of grasping area. Impact model between end-point and target is established to describe the dynamic process of space robot capturing object.Then two methods are proposed to estimate inertia parameters of target (mass, centroid position and inertia matrix). One method is based on the conservation principle of linear and angular momentum and the other is based on Newton-Euler equations of motion. The equations are linearized with respect to the unknown inertial parameters (mass, the three components of centroid position with respect to the target coordinates and the six independent elements of inertia matrix). Only the linear/angular velocities of space robot and target are used in the identification method based on momentum conservation with no information about the force and torque utilized. The method based on momentum conservation is suitable to parameter identification of free-flying space robot and unknown target.The linear parameterization model is established above, and on-line estimation of target parameter based on least square method is utilized. The influence of null space compensation and measure noise disturbance to parameter identification precision is analysed from linear equations least square theory. The feasibility of the method is demonstrated by numerical simulation.