Interactive synthetic environments with force feedback

Abstract

The evolution of the visual display technology used in synthetic environments is fueling the development of numerous applications. The results of these initial expeditions into virtual worlds have been promising. However, these initial investigations have also highlighted the need for force feedback in synthetic environments to make the virtual experience more immersive and easier for the traveler to interact with the objects that populate the synthetic environment. In addition the inclusion of force feedback in a synthetic environment will provide another input channel that can provide information to the traveler beyond the typical visual and audio input modes. Research in the area of force feedback for synthetic environments thus far has focused on the design and construction of specialized interface devices. These new haptic devices can be used to provide force interaction, however because these devices are unique prototypes it is difficult if not impossible to reproduce and extend results obtained at different facilities. This work proposes a new approach to force interaction in synthetic environments, virtual manipulators. The virtual manipulator control concept can be applied to any available six degree of freedom robot manipulator. Therefore experimental results obtained using the virtual manipulator control law can be reproduced at any research facility with a six degree of freedom robot. This work will develop the virtual manipulator control approach as well as investigate the stability characteristics of the control law operating on a general six degree of freedom robot. Experimental results will be presented for various virtual manipulators including the time varying extension of the virtual manipulator concept. In addition to the virtual manipulator concept this work will also develop a physically-based modeling technique that can be used to assimilate a force feedback device into a synthetic environment. This modeling approach uses finite element analysis techniques but uses the NURBS basis functions instead of the typical interpolation basis functions. As a result the dynamics of the model can be represented using the same characteristic parameters as the geometric model. Results of this modeling approach will be presented for one and two dimensional dynamic models.