A neural network based inverse kinematic problem

Solution of inverse kinematic problem of robotic manipulator is a complex task yielding no unique results. This paper presents a neural network based solution to inverse kinematic problem. There exist quite a few methods to solve inverse kinematic problem, but as the number of joints increase these methods find it difficult to solve the Inverse kinematic problem. Normally it is difficult to model the inverse kinematic problem of a manipulator having large degrees of freedom and hence the solution becomes complex and sometimes inconclusive. Even methods like neural network having their own limitations to solve problems accurately. If the accuracy is compromised there may be real problems in handling the jobs or carrying out the desired tasks. The neural network needs to be systematically and properly trained to use it for finding solution. The developed algorithm has been used to solve the inverse kinematic problem and has been validated through comparison.