Trajectory Planning Of 7-DOF Humanoid Manipulator Based On Multi-objective Optimization

With the continuous improvement of the automation level and the increasing complexity of task requirements,the demand for high-degree-of-freedom humanoid manipulator which can perform more flexible movement is increasing as well.As an important part of the humanoid manipulator task,Trajectory planning,the basis of smooth and efficient operation,has significant practical meaning and value.In this paper,based on the7-DOF humanoid manipulator,and the multi-objective trajectory is studied by analyzing its kinematics and dynamics.The main content is as followings:(1)Aiming at the multi-objective optimization problem of trajectory of humanoid manipulator,in order to make the optimal solution set of Pareto can approximate the real solution,the characteristics of multi-objective optimization problem are analyzed.Aiming at the problems existing in traditional teaching and learning optimization algorithms,the second teaching strategy and diversity stage are introduced,and an improved multi-objective teaching and learning optimization algorithm(IMTLBO)is proposed,and the basic implementation steps and flow chart of the algorithm are given.Finally,the performance of the algorithm is analyzed and verified by simulation experiments(2)First,the joint structure model of the 7-DOF humanoid manipulator is established,and the forward kinematics analysis is carried out.The inverse kinematics of 7-DOF humanoid manipulator is solved by geometric analytical method,and the joint velocity is obtained by Jacobi Matrix.Aiming at the complexity of the dynamic modeling of 7-DOF humanoid manipulator,this paper presents a solution based on Newton-Euler dynamic equation,which uses spatial vector theory to deduce the inverse dynamics of humanoid manipulator,which can effectively reduce the computation of dynamic equations,is easy to be programmed,and is helpful to the trajectory planning process of humanoid manipulator(3)Taking the task of the humanoid manipulator for hitting ping-pong as an example,in order to reduce the energy consumption of humanoid manipulator in the course of operation,this paper proposes a mathematical model of energy consumption based on dynamics,which can more realistically reflect the energy consumption in the process of imitation of humanoid manipulator than the energy consumption model analyzed from kinematics angle.Taking the energy consumption and arm posture configuration comfort in the process of humanoid manipulator operation as a double optimization index,considering the mechanism physical constraints and space obstacle constraints in the motion process,the redundant variables are obtained by imitating the trajectory of the humanoid manipulator,and based on the MOTLBO,the multi-objective trajectory planning scheme of the imitation human arm to table tennis is proposed.(4)Taking the 7-DOF humanoid manipulator for hitting ping-pong as an example,numerical simulation and experimental test are carried out on the dynamic energy consumption model and the multi-target trajectory planning scheme of humanoid manipulator.The results show that the dynamics-based energy consumption model and kinematics-based compared with the energy consumption model,it can better reflect the energy consumption of the humanoid manipulator during the operation.The multi-objective trajectory planning experiment of the humanoid manipulator shows that the optimal results of the multi-objective trajectory planning scheme can be better approximated to the real Pareto frontier.In the optimal solution set,a non-dominated solution is selected for the humanoid manipulator online planning,and the results verify the validity of the above research contents.