Research On Multi-task Motion Planning Of Supernumerary Robotic Limbs

In the industrial fields such as the operation and maintenance of nuclear power plant and the production and manufacturing of large aircraft,there are a large number of tasks which can only rely on manual operation,with complex operation processes,high intensity of work,high flexibility requirements,and difficult to achieve mechanization and automation.In the process of operation,the efficiency of work and personal safety of operators are difficult to guarantee.Therefore,the supernumerary robotic limbs,which could coordinate with human and realize the automation of part of the tasks,have become a new technical means to solve the above problems.The supernumerary robotic limbs are a high dimensional system with multiple working conditions and multiple task space.The multi-task motion planning and control method for the system is the key to improving cooperative operation ability of robots.The main research work are as follows:1.The kinematics model was established for the virtual prototype of supernumerary robotic limbs.The forward and inverse kinematics and Jacobian matrix of single limb and backplane were derived by D-H parameter method,and the dynamic model of single limb was established based on Lagrange.2.Based on quintic polynomial interpolation algorithm,the trajectory of single limb and backplane was planned.Combined with the working scene of supernumerary robotic limbs,the quintic polynomial interpolation algorithm was used to plan the trajectory of the single limb and the backplane motion,which could ensure the continuity and smoothness of the motion.According to the planned trajectory,the velocity controller based on Jacobian matrix was designed to verify the kinematic model of single limb.The kinematic model of the backplane is verified by the planned backplane back-and-force motion trajectory.3.The multi-task motion planning and control of single limb end was performed based on null space projection.The redundancy of the single limb system was analyzed,and the task decomposition of the pick-and-play and obstacle avoidance operations was carried out.Then the Jacobian matrix was derived.The multi-task priority structure of pick-and-play and obstacle avoidance operations was established using null space projection.The multi-task control of two operation modes performed well in the single limb virtual prototype.4.The multi-task motion planning of multiple limbs end with whole body control was carried out.In the suspension working condition of supernumerary robotic limbs,the whole body motion planning method based on null space projection was studied,and the whole body control framework was designed,which was based on limb task and backplane task priority structure.The multi-task motion planning and control of backplane and single limb end performed well in the virtual prototype.