Basic vision experiments with the mobile robot HERMIES-II

The research team of the Center for Engineering Systems Advanced Research (CESAR) uses the mobile robot HERMIES-II (Hostile Environment Robotic Machine Intelligence Experiment, Series II) for experiments in autonomous obstacle avoidance, path planning and navigation. Sensor feedback from the robot comes from several sources, including sonar and a binary vision system. It is this vision system with which this work will primarily be concerned.

The vision system implemented on the robot is based on the MicronEye OpticRam camera from Micron Technology Inc. The camera was interfaced to the robot and several image processing routines were written in the language FORTH. The general purpose assembly language routines provided with the MicronEye were rewritten and condensed for this specific application. "Primitives" were written to allow external computers such as CESAR'S LISP computer to access information provided by the camera. Routines were writ ten to do basic tasks such as calculating the angle from the center of the robot to the edge of an obstacle. These routines were used to develop navigation routines. The LISP computer previously had used only sonar feedback to determine the locations of obstacles and controlled the robot by passing commands over an RS-232 serial link. The new routines allowed more processing to be done in the robot and decreased the time needed to accomplish an overall task.