This paper implements multi-layered ANFIS controller in PIC16F886 micro-controller as a supervisory control for a 6 DOF robotic arm. The complexity in mathematical modelling demands for machine-learning techniques, which rely less on precise mathematical analysis. ANFIS is one such machine learning technique which helps in decision making for the control of robotic arms. Standard PD controllers could be used as servos to guarantee precise tracking. Based on real physical parameters of Dexter ER-1, a model is developed in Sim –Mechanics to capture the actual dynamics of robot arm. The time dependent reachable set is generated out of which it gave nearly 40,000 data samples. This data is used as inverse training data for ANFIS network and is implemented as a supervisory controller in microcontroller. The controller is tested with predefined paths and random position targets and results are shown to act satisfactorily.
M. Madhu Dharmana, Shashidhar, S., Kumar, S., and , “Embedded ANFIS as a Supervisory Controller for a 6-DOF Robotic Arm”, International Journal of Engineering Research, vol. 3, no. 5, pp. 318-320, 2015.