Research On Trajectory Planning Algorithm Of Manipulator Based On Energy Optimization

In the modernization of production and information technology, the robots have been widely used in industry, education, medical, military and other fields, as they could liberate and develop the productive forces and complete tasks with high efficiency. The quality of trajectory planning is one of the key factors that influences the performance of robot. However, trajectory planning in manipulator was stuck in polynomial interpolation and spline function methods, which has some disadvantages like a available path should be replanned, the whole trajectory is affected by a single path fluctuation and low efficiency. A new automatic trajectory planning algorithm derived from both Ripidly-Exploring random tree(RRT) and A * algorithm called heuristic sampling algorithm has been put forward, which aims at improving trajectory planning efficiency and energy optimization. Such merits act as follows: perfect the efficiency by planning smooth curve of joints directly instead of quadratic programming; meet the dynamic requirements which avoids the overflow of torque; achieve automatic trajectory planning optimization which changes from preestimating available route to plan automatically. The 2 degree of freedom(2 DOF) manipulator's throwing model is regarded as the main research subject in this thesis. The contents are as follows:Firstly, a 2 DOF manipulator is established and throwing kinematic and dynamic equations are derived through Lagrange method. The research and analysis in the speed and direction of throwing position via motion model provides a theoretical basis for casting loads to reach the goal.Secondly, the application of the heuristic sampling algorithm in 2 DOF manipulator is described in detail. Energy optimization acted as heuristic evaluation guides the sampling planning between the initial and final position. Sampling unit, number, method and time interval are briefly summarized. Grid method is applied to select approximate nodes and the sign of angular velocity is used for the judgment of moving direction. All mentioned above lays the foundation for the dynamic simulation in the subsequent cast model.Finally, the dynamic simulation is carried out in MATLAB using C++ language, which intends to realize the application of heuristic sampling algorithm in 2 DOF manipulator. The analysis of angular acceleration, angular velocity and torque curves under multiple loads verifies the feasibility of the algorithm the thesis put forward. And heuristic sampling algorithm has been proved its superiority in planning time and stability of joints by comparing to B-spline algorithm.