will deploy properly. Doing this by building a physical prototype of the antenna is impossible. The antenna is made primarily of cables and will not support its own weight under gravity. Even if one could build a physical prototype, one would be reluctant to test a dejamming procedure because if the antenna should jam in deployment, the procedure might destroy the prototype. In this situation, computer modeling and simulation are an essential tool, allowing one to electronically simulate deployment and dejamming without risk.
Figure 2 is a picture of an electronic prototype of the 3-finger JPL gripper. We wanted to experiment with gripping and rotating objects and selected this particular gripper to prototype electronically because we had access to the engineering drawings. The prototype is more than just a solid model of the gripper. The model has properties such as mass and inertia and the ability to represent forces on the cables in the fingers. The modeling system, NEWTON (Cremer, 1989), takes a description of an object such as this gripper and allows one to program the gripper to grasp or manipulate an object and to view the simulation on a workstation. What is novel about NEWTON is the automatic creation of a simulation from a physical description of an object, and the capability of an event handler to detect if a collision takes place during a simulation. If a collision does take place, NEWTON also appropriately modifies the equations governing the simulation to account for the new constraints.
Since the model of the gripper is only a computer file, one can edit it. Figure 3 shows a 4-finger version. One quickly realizes that one has the ability to explore finger placement, length and number of fingers, size and shape of gripped objects, and so on. Thus, a designer could explore a large design space, which would be impossible if he or she were dealing with a physical gripper. One can try out designs by simulating actual situations in which a dextrous gripper would be used. This ability to design and test objects by simulation will facilitate the creation of new, better, and less expensive objects for all.
Daniela Rus, a graduate student at Cornell University, used the electronic prototype of the gripper in her Ph.D. thesis on manipulation of objects (Rus, 1991). Her work extends earlier research on grasp