We developed an autonomous omni-directional wheeled mobile robot which has three driving omniwheels arranged in a triangular manner. We adopted Xscale processor as a main controller which is based on the ARM architecture enabling high-performance computing with low power consumption. Also by using Real-time Linux as OS, we constructed a Real-time control system. And we proposed the Encoder Trailer that is is composed of four encoders. we designed and tested it for more accurate self-localization. To control the motors and sensors, we designed and implemented an interface board for Xscale processor. And we used wireless LAN to monitor and control the mobile robot in remote without any wire connection.
Main task of this robot is to arrive at the goal position autonomously by using a given global map. So the sensors are selected and implemented on the purpose to apply obstacle-avoidance, path planning and self-localization of mobile robot. IR Scanner was used for the environment detector with 180 angle scan per 0. 1 sec and encoder trailer for localization.
To arrive at the goal in the global map, the robot performs the global path planning using the map with visibility graph technique and local obstacle avoidance when the robot faces an unexpected obstacle. Also the robot creates the smooth path from the segment-based path for efficient movement and follows the path.
The results of experiments are described to show the motor control of omni-directional wheels, the performance of encoder trailer, the accuracy of IR Scanner and the validity of autonomous navigation with path planning.