Automation of short-run assembly cannot be justified economically. The problem is how to reduce the labor cost for runs of medium length. It was hoped that this could be accomplished by use of robots which are reprogrammable. Two approaches have been taken. One is to develop visual, tactile, and force sensors for the next generation of robots so that they can find the parts to be assembled and then guide them to their appointed positions in the assembly. The other approach is to use the present generation of robots, but to reduce the cost of equipment which brings the piece parts to locations where the robot can find them and pick them up. Much time and money have been spent on the first approach but the results have been far from satisfactory. This effort deals with the second approach. Four tasks are undertaken: design, fabrication, test, and evaluation of low cost, easily programmable part feeders that convey parts, one by one, from a jumbled collection to a point where the robot can find them, properly oriented; design, fabrication, test, and evaluation of three- and four-finger grippers versatile enough to handle a large variety of parts; establishment of an assembly work station based on the foregoing tasks; and development of rules for designing parts and products so that they are suitable for robot-based assembly systems.
