Inverse kinematics of a 7 DOF manipulator using Adaptive Neuro-Fuzzy Inference Systems

This paper was carried out objectively to explore and describe the inverse kinematics solutions of an anthropomorphic redundant robotic structure with seven degrees of freedom and human like workspace. Traditional inverse kinematics methods can have an unacceptably slow pace for the today´s extremely redundant systems. The presented method uses the Adaptive Neuro-Fuzzy Inference Systems (ANFIS) editor and the Fuzzy Logic toolbox from MATLAB® which allow the investigation of various kinematical suitable solutions. ANFIS supports the determination of one degree of freedom, remaining therefore only six undetermined degrees. For better understanding of the simulations a CAD model that mimics the real robotic arm was created into SolidWorks® environment, followed by the conversion of the CAD parts into Simulink®-SimMechanics model. The results of this method are more interactive in response and more reliable.