Analysis And Optimization Of Parallel Robot Mechanism

Parallel robot mechanism is used widely in many fields such as parallel robot and parallel machine tool because its advantages of rigidity, load capacity, high control precision and good dynamic performance. Nowadays, the study of parallel robot mechanism is unfolding and is concerned by many researchers. In this paper, Monte Carlo methods, particle swarm optimizing algorithm, artificial neural network and differential evolution algorithm were used for the analysis and optimization of parallel robot mechanism.Firstly, aiming at the complexity and multi-result of the forward position of parallel mechanism, sharing fitness of particle swarm optimization was proposed for exploring all the results, PSO-Newton algorithm which combined sharing fitness of particle swarm optimization and Newton iteration algorithm was designed for improving the convergence speed. The algorithm was applied to the solutions of the position of the 2PR&2SPS parallel robot, and the result shows that the algorithm works well. It is also a new and fast algorithm for solving the parallel mechanism's forward positional.Secondly, aiming at the complexity of analysis the workspace of parallel robot mechanism, Monte Carlo method was used for analyzing the shape and size of workspace. But the Monte Carlo method is not very effectual, so a new kind of BP network, which was trained by quantum particle swarm algorithm, was proposed. The efficiency was proved by example.Thirdly, a multi-objective differential evolution algorithm (MODE) based on modified differential evolution algorithm was proposed and the algorithm was designed for multi-objective optimization problem of parallel robot mechanism.Finally, how to build parallel robot mechanism's multi-objective model was proposed. Taking a 3-TPT parallel robot mechanism as an example, a multi-objective model was presented, the objectives of which were the mechanism's work space and global kinematics performance index, and the model was solved by MODE.The results show that it is beneficial for the theoretical analysis, design and application of parallel robot mechanism to use some advanced algorithms in analysis and optimization of mechanism. The algorithms in this paper are robust and suitable for the other similar problems.