Robust control of a 3-DOF parallel cable robot using an adaptive neuro-fuzzy inference system

Cable-driven parallel robots inherit the advantages of serial and parallel manipulators, besides these robots do not have parallel robot's structural restriction. Utilizing the cables instead of conventional rigid arms in parallel and serial mechanisms causes challenges. Due to intrinsic properties of the cables, one of the challenges in designing and controlling these mechanisms which has to be taken into account is the inherent flexibility of cables. The cable flexibility makes the cable-driven robots highly sensitive to noise and disturbance. Due to this sensitivity, this research presents an adaptive-neuro fuzzy controller which is robust to disturbance and can reduce sensitivity of the system to noise. Kinematic and dynamic equations of cable robot motion have been obtained to investigate the system behavior by computer simulation. Designing an ANFIS controller requires a large training dataset. Therefore, a designed PID controller has been used as a supervisor for the ANFIS controller. The resultant data from this PID controller used to design a suitable neural network with an acceptable error. The ANFIS controller used to control the cable robot system and the results have shown better performance of the controller. Eventually, performance of the ANFIS controller has been compared with performance of the PID controller in presence of noise which the results show the effectiveness of the designed ANN-controller.