The intelligent controller for spherical robot

Abstract

This research sought to design the intelligent controller for a spherical robot. Spherical robot is a ball-shaped mobile robot that capable to move from one place to another to perform the desired tasks. A error driven Fuzzy Logic Controller (FLC) which consist of two Proportional-Derivative type FLC and Proportional-Integral type FLC are designed to control the position of the spherical robot, angle of the pendulum and the angular velocity of the spherical robot. The intelligent controller like FLC is proposed to reduce the dependency on its mathematical model, FLC is a knowledge-based controller which can be design without using a complex mathematical model. Then, the designed controller is optimized with tuning the input and output gain of the controller with Particle Swarm Optimization (PSO) method. Besides, the effect of different type of membership function (MF) and number of variables used for MF are also been studied by heuristically tuned these two parameters. As a result, five variable of triangular and trapezoidal MF are selected as the best MF to be used in both PD-type FLC for position controller while gaussian and sigmoidal are selected to be used in PI-type FLC to control the angular velocity of the spherical robot. With PD-type FLC, rise time and settling time of the position control of the robot managed to be reduced by 84% and 87% respectively. In addition, PD-type FLC also capable to eliminate the overshoot and reduce the settling time of the pendulum control by 60 % compared to when the robot run without the controller. The optimized PI-type FLC used to control the angular velocity of the spherical robot manage to eliminate the steady state error and achieved its desired setpoint in 1.47s. Finally, the spherical robot with monitoring devices like camera, temperature sensor and the gas sensor are successfully developed, and the designed controllers are integrated to control the position, angle of pendulum, and the velocity of the spherical robot. The experimental performance analysis is studied and show that the integration of the controller managed to control the position with settling time 60% faster than simulation. While the PD-type FLC used to control the angle of pendulum managed to control its angle to it desired setpoint but having some problem on maintaining the angle due to the rotation of the spherical robot. For PI-type FLC, the experimental value achieved 83% of its desired angular velocity and need to be optimized to really achieved it desired setpoint velocity. Therefore, it concluded that the objective of the research is achieved to design and studies the integration of the intelligent controller for the developed monitoring and surveillance spherical robot.