Research On Motion Planning Of Free-floating Space Robot For Target Capture

Along with the development of space technology, activities of space explorationand exploitation are becoming frequent with the help of spacecrafts, and how to ensurethe normal operation of spacecrafts is an urgent problem to solve. Due to the limitationof early technology, most of the on-orbit servicing missions are completed byextravehicular astronauts, who have to be confronted with the dangerous spaceenvironment. In recent years, space robots are used widely to assist or replace astronauts,which can lower risks and costs largely.Motion planning is often conducted in advance to realize that constraints andperformance indicators are satisfied when the space robot operates. There existsdynamic coupling between the base and the mechanical arm, which is different from thefixed-base robot on the ground, so it is necessary to study the rule of motion planning ofthe free-floating space robot.Firstly, the motion planning for point-to-point motion of space robot is studied.Based on trapezoidal velocity profile, the planning algorithm adopts the method ofdamped least square to solve singular problem at the cost of decreasing accuracy, andthe velocity planned cannot adjust in real time, which leads to accumulation error, andthe end-effector cannot reach the desired position. To solve this problem, a novel motionplanning algorithm, which combines trapezoidal velocity profile and proportionalnavigation method, is then presented. By the proposed algorithm, the trapezoidalvelocity profile method is utilized in the first stage, followed by the proportionalnavigation method when the end-effector is close to the desired position in a certainsmall distance. For the reason that the velocity of the end-effector can adjust on line, thevelocity points to the desired position all the time, which guarantee that the end-effectorcan reach the desired position. The simulation result shows the feasibility of this novelalgorithm.Secondly, the motion planning for target capture is studied. Based on proportionalnavigation method, the planning algorithm is adopted, and simulation on a space robotcapturing a target with constant speed in a straight line proves that the algorithm iseffective. But capture time cannot be predicted, which makes it difficult to capture thetarget when it moves out of the workspace of the mechanical arm. To solve the problem,a novel motion planning algorithm is then proposed, which combines trapezoidalvelocity profile and proportional navigation method. With the use of sensors, the motioncurve of the target can be gained, so a point in the workspace of the arm can be selectedas a desired position, and the time to reach the position can be gained. Then the end-effector can reach the position near the desired before the target arrives, duringwhich the proportional navigation method is adopted. When the end-effector is close tothe target in a certain small distance, trapezoidal velocity profile is applied as a relay.Finally simulation on a space robot capturing a target with constant speed in a straightline shows that the algorithm can work effectively.