This paper describes a telemanipulation system to assist persons with disabilities performed dexterous manipulation tasks. This research is expected to enhance the teleoperation performanc through the use of scaled mapping from master to slave manipulation based upon sensory data. This is particularly useful when the user has restricted range of movements due to certain disabilities such as muscular dystrophy, a stroke, or any form of pathological tremor.

Assistance functions are used for mapping such that human input is enhanced rather than superseded by the computer. A common vocational rehabilitation test referred to as Box and Blocks was chosen to test the effectiveness of this sensor-assisted function. A variable scaling scheme was developed using available sensory data.

In the simulation mode, a visual environment was created for the Box and Blocks test. This was used to predict if a person with disabilities would be able to perform a task comfortably. The real test was performed using a master and slave manipulator system with a camera and laser rang finder. A motion constraint was added to the master to simulate a user with disabilities. The results demonstrated that the sensor assistance not only reduced required input motion, idle time, and execution time, but also increased manipulation accuracy during the Box and Blocks test.