Research On Inverse Kinematics Of Redundant Manipulator Based On Swarm Intelligence Optimization Algorithm

For a redundant manipulator whose geometrical structure does not meet the Pieper criterion,its inverse kinematics solution cannot be obtained by closed-form methods,but can only be solved by numerical methods.The conventional numerical iterative methods may cause considerable computational load,and there are cumulative errors and singularities.In response to these problems,the application of swarm intelligence optimization algorithm to solve the inverse kinematics of redundant manipulator has made good progress.However,there are still some problems to be studied,such as slow convergence speed,low convergence accuracy and easy to fall into local optimum.In order to effectively improve the quality of solving the inverse kinematics problem of the redundant manipulator,the typical swarm intelligence optimization algorithms,including particle swarm optimization(PSO),clonal selection algorithm(CSA),and fruit fly optimization algorithm(FOA),are improved.The main works of this paper are as follows.(1)Improvements of PSO and inverse kinematics solution of the redundant manipulator.In this paper,two improved particle swarm optimization algorithms,namely comprehensively improved particle swarm optimization(CIPSO)and hybrid mutation particle swarm optimization algorithm(HMPSO)are proposed.The CIPSO is comprehensively improved from the aspects of swarm initialization,inertia weight adjustment strategy,differential mutation operation,the boundary violation treatment of particle position and flight velocity as well as local variable-depth search.In HMPSO,an inertia weight value updating strategy with a random factor is adopted,and then the hybrid mutation evolution is introduced.CIPSO and HMPSO are applied to solve the inverse kinematics problem of a 7-DOF redundant manipulator.The experimental results show that both CIPSO and HMPSO have achieved good success rates in finding the inverse position solution,while their success rates in finding the inverse pose solution are 21%and 65%,respectively.(2)Improvements of CSA and inverse kinematics solution of the redundant manipulator.Two improved clonal selection algorithms with a bilevel coevolution mechanism,namely bilevel coevolutionary clonal selection algorithm(BCECSA)and improved BCECSA(IBCECSA)are presented,in which each level of evolution evolves with a different evolutionary scheme.Through information sharing,the co-evolution between the levels is realized,and an evolutionary model of intra-level competition and inter-level cooperation is formed.Although BCECSA and IBCECSA adopt the same algorithm framework,their evolutionary operators and the objects for immune operation are different.The improved algorithms are used to solve the inverse kinematics problem of a 7-DOF redundant manipulator.The experimental results reveal that both BCECSA and IBCECSA have obtained ideal success rates in finding the inverse position solution,while their success rates in finding the inverse pose solution are 25%and 83%,respectively.(3)Improvements of FOA and inverse kinematics solution of the redundant manipulator.Two improved fruit fly optimization algorithms are developed,which are double strategies co-evolutionary fruit fly optimization algorithm(DSCFOA)and hybrid mutation fruit fly optimization algorithm(HMFOA).In DSCFOA,the good point set method is used to initiate the fruit flies'position,and the swarm leadership is introduced to collectively guide the optimization search of individuals.In addition,a co-evolutionary mechanism based on double strategies is constructed in DSCFOA.An olfactory search based on multiple mutation strategies and a visual search based on the dynamic real-time updates are adopted in HMFOA.The improved algorithms are used to solve the inverse kinematics problem of a 7-DOF redundant manipulator.The experimental results indicate that both DSCFOA and HMFOA have also achieved good success rates in finding the inverse position solution,and their success rates in finding the inverse pose solution are as high as 86%and 96%respectively.In this paper,a systematic study of the inverse kinematics solution optimization algorithm of the redundant manipulator based on swarm intelligence optimization algorithms is carried out.Aiming at the shortcomings of PSO,CSA and FOA,relevant improved algorithms are proposed.Among them,HMPSO obtained the fastest speed of inverse position solution,and HMFOA has the highest success rate in finding the inverse pose solution.