Research On Motion Planning And Control Of A Dual-manipulator Strongly Coupled System

For the single non-redundant manipulator(NRSM), there exist big limitationsin flexibility, adaptability and operating capacity for the current level of robottechnology. In order to meet the needs of complex tasks, dual-redundantmanipulators (DRM) come into being, compared with the NRSM, the controllerdesign, solution for inverse kinematics, motion planning and coordination controlfor the DRM are more complex and difficult.According to the needs of the DRM’s control, the hierarchical controlarchitecture of central controller–multi joint controller is used. According to theneeds for joint control, joint controller based on FPGA(Field-Programmable GateArray) which embedded NIOS II microprocessor is designed, the testing resultsshowed that the joint controller can quickly acquire the sensor information, andcommunicate with the central controller at a high speed.Firstly, kinematic constraints for the strongly coupled dual-manipulator system(SCDMS) which the DRM grip a rigid body is given. Combined with therequirements of the mission, a seven-segment Cartesian trajectory planning methodwas used, the trajectory planned by this method owns continuous velocity andacceleration, and the velocity, acceleration and jerk are limited.Then, an improved gradient projection method is deduced to solve the velocity;this method can effectively handle manipulator singular position and avoiding jointspeed limits. For the7DOF manipulator with offset-wrist, the deficiencies of thecommonly used analytical method for inverse solution of position is given, and thenan analytical solution called virtual joint method is proposed, the simulation resultshows that this method has a high position accuracy.Finally, the dynamic model of the system is assigned, and an arm positioncontroller using a PD+Gravity compensation technique is designed. Model ofSCDMS is build based on SimMechanics, and the simulation result showed that thecontact force between manipulators and object changes frequently and leads to bigrelative position error. In order to control the contact force, impedance control inCartesian space is applied and a position/impedance coordination controller isdesigned. The simulation result shows that this controller can track the givencontact force well and reduce the relative position error.