Mobile robot navigation using sensory feedback

Sensor-based algorithms to navigate a mobile robot from one geographical location to another as fast as possible without using vision are discussed. The environment is completely unknown a priori, but certain assumptions are made about the environment. An algorithm to navigate a robot in a manufacturing environment without chemical paths or tracks using a minimum amount of a priori information is described. Solid-state color sensors, ultrasonic and infrared sensors are used for path planning and recalibration to improve the bandwidth of the operation. The algorithm has fast and slow modes. Fast motions can be achieved through appropriate switching between these modes. The algorithm was implemented on research robot.<>