Vision-based autonomous robot control for pick and place operations

Typical robotic pick and place operations involve sensors such as touch and vision attached to the robot. In the presence of ionizing radiation this approach becomes difficult because most customary sensors degrade over time due to radiation exposure. The alternative is the placement of multiple cameras outside the radiation zone, and to identify and infer the location and dimensions of parts within the robot's workspace through image analysis. The paper describes strategies for the reliable and flexible control of the material handling robots inside a hot cell by means of automated stereo vision systems. Since the cameras can be positioned outside the hot cell, such systems would have significant advantages over sensors inside the hot cell, resulting in potentially reduced system maintenance and increased system reliability.