A UMI RTX robot, modified with limited end-effector sensors and a restricted but effective vision system, is currently used in a developmental education setting for severely physically disabled children. The low physical and cognitive abilities of the children involved in the project require a semi-autonomous robot with environmental sensing capability to operate in a task oriented mode. A variety of low-cost sensors including proximity, distance, force and slip sensors, have been investigated for integration in end-effectors for the RTX robot. The sensors employed on a modified end-effector are detailed and experimental results are presented. A design for an end-effector with integrated sensors is discussed. The integration of the sensor information into a high-level, task-oriented programming language is detailed and examples of high-level control sequences using sensor inputs are presented. Finally, the development of intelligent gripping strategies based on sensor information is discussed.
Read full abstract