The major issue for the wheeled mobile robot is the low level controller
gains tuning up especially in the robot competition. The floor surface can be
damaged by the robot wheels during the competition, therefore the surface
coefficient can be changed over time. PI gains have to be tuned before every
match along the competition. In this research, the torque controller is defined
and implemented in order to solve this problem. Torque controller consists of a
PI controller for the robot wheel's angular velocity and a dynamic equation of
brushless motor. The motor dynamics can be derived from the energy conservation
law. Three different carpets, which have the different friction coefficients,
are used in the experiments. The robot wheel's angular velocity profiles are
generated from the robot kinematics with different initial conditions. The
output paths of the robot with the torque controller are compared with the
output paths of the robot with regular PI controller when the same wheel
angular velocity profiles are applied. The results show that the torque
controller can provide a better robot path than the normal PI controller.