Point-to-Point Motion Planning of a Free-Floating Space Manipulator Using the Rapidly-Exploring Random Trees (RRT) Method

Abstract

SUMMARYIt is usually proposed to use a robotic manipulator for performing on-orbit capture of a target satellite in the planned active debris removal and on-orbit servicing missions. Control of the satellite-manipulator system is challenging because motion of the manipulator influences position and orientation of the chaser satellite. Moreover, the trajectory selected for the capture manoeuvre must be collision-free. In this article, we consider the case of a nonredundant manipulator mounted on a free-floating satellite.We propose to use the bi-directional rapidly-exploring random trees (RRT) algorithm to achieve two purposes: to plan a collision-free manipulator trajectory that, at the same time, will result in a desired change of the chaser satellite orientation. Several improvements are introduced in comparison to the previous applications of the RRT method for manipulator mounted on a free-floating satellite. Feasibility of the proposed approach is demonstrated in numerical simulations performed for the planar case in which the chaser satellite is equipped with a 2-DoF (Degree of Freedom) manipulator. The obtained results are analysed and compared with the results obtained from collision-free trajectory planning methods that do not allow to set the desired final orientation of the chaser satellite.

A Review of Spatial Robotic Arm Trajectory Planning

“…Moreover, since the RRT algorithm does not require an explicit obstacle representation, it is widely used in various motion planning tasks [58], such as space obstacle avoidance detection tasks. Rybus et al [59] used the RRT algorithm to plan a collision-free manipulator trajectory. The bi-directional approach is used in the construction of the tree in order to plan a trajectory from the given initial state to the specific final state.…”

“…In summary, the current status of space robotic arm trajectory planning research based on obstacle avoidance requirements is summarized in Table 1. √ Dynamic Obstacle Avoidance [40] √ Reaction Zero and Jacobi Transpose [41] √ Adjusting Gain [43] √ Use a Mass-spring-damping System [46] √ Convex Quadratic Programming [47] √ CCDJAP-IK [49] √ AccPF [50] √ - [52] √ Based on C-space Topology [53] √ Two-way Heuristic Search Method [55] √ FTWLCC [56] √ Combine the Dynamic Window [59] √ - [60] √ A New Time-scale Transformation [61] √ Improved RRT [62] √ Combine with APF…”

Kinodynamic planning with reachability prediction for PTL maintenance robot

“…In this article, the kinodynamic planning is the center of attention which is performed at the beginning of the maintenance operation. 28 The planning of dynamic motion is usually carried out in the state space of the robotic joints, that is, the vector of the effective state x = ½ c T _ c T T in the kinematics, where c is the configuration vector of 12dimensional generalized coordinates, and determines the position and posture of all the links at a given moment; _ c is the derivative of c with respect to time and describes its velocity. The torque of the actuator is set to six-dimensional time-varying action vector u = u(t) 2 U, where U represents the set of effective actions under the constraint of driving force limiting the system acceleration indirectly.…”

“…In this article, the kinodynamic planning is the center of attention which is performed at the beginning of the maintenance operation. 28…”

ADD-RRV for motion planning in complex environments

“…Motion planning problems, requiring to compute a path from start to goal in the configuration space without colliding with obstacles, have been widely studied mainly in robotics, for example, autonomous driving [1], unmanned aerial vehicles [2], and on-orbit operations [3], as well as other research areas, such as computer animation, part assembly [4], and computational biology [5] since the past decades.…”