Environment recognition for a mobile robot using double ultrasonic sensors and a CCD camera

To move efficiently in an unknown or uncertain environment, a mobile robot must take observation from various sensors to provide information for path planning and execution. A sufficient representation of the external world would also be very useful for self-localization. One of the merits of applying multiple sensors to a mobile robot is the enhancement of environment recognition. In this paper, sensory information combined from double ultrasonic sensors and a CCD camera is provided for this purpose. The authors used ultrasonic sensors for distance measurement and a vision system for object boundaries detection. The authors developed an algorithm to eliminate errors due to the beam opening angle of ultrasonic sensors based on a dual transducer design. An extended discrete Kalman filter was used to fuse raw sensory data and reduce the influence of specular reflection of ultrasonic type transducers. Therefore a more reliable representation was obtained for environment recognition. Computer simulation as well as practical experimental results show that this sensory system can provide useful and robust environment recognition for intelligent robotics