Positional Control of a 3D-Printed Indigenous 6-DoF Feeding Robotic Arm Using ANFIS and Fuzzy-PID Controller

The current chapter demonstrates the use of an Adaptive Neuro-Fuzzy Inference System (ANFIS) and fuzzy proportional integral derivative (PID) controller to control the position of a 6-DoF custom 3D-printed feeding robotic arm. The chapter uses an open-source platform to show how machine vision, robot kinematics, and positional control may be coordinated. Throughout its function, the robotic arm travels through several intermediary locations in the 3D Cartesian space. The robot is designed to collect the food from the bowl and bring it up close to the user’s face. The destination point, which is dependent on where the patient’s face is, is in this instance the most important component. The vision camera provides the location of the face, and the inverse kinematics algorithm gives the joint angles for that location. To control the position of six servo actuators for the destination point, we employed an ANFIS and FPID-based positional controllers. The last three medium-torque servos are controlled by FPID, and the initial three high-torque servos are controlled by ANFIS. The FPID offered an average 15% overshoot, whereas the ANFIS offered a mere 5% overshoot. It was decided to control just high-torque servos using ANFIS and the rest actuators using FPID because the complete system was running on an affordable open-source platform like an Arduino or Raspberry Pi, and we wanted to reduce the computational overload on the Raspberry Pi.