Haptic augmented reality interface using the real force response of an object

Reviewer: John S. Edwards

Kurita et al. describe a technique to assist with the perceived sense of touch in a virtual environment. Commercially available haptic devices are currently employed in virtual reality environments, to provide sensory feedback to the user. While they can provide three or six degrees of freedom, the authors describe limitations of these devices-they do not model the complex force responses of visco-elastic objects well. As Kurita et al. state: "to enhance the feedback force, the actuators and sensors have to be improved and it [would come] with a high cost." In an attempt to improve the response, the authors use an inexpensive haptic device and propose an augmented reality (AR) system where the haptic device can interact with a base object. The resultant AR force is then the sum of the virtual force of the haptic device and the real force exerted by the base object. In the experiment, three objects ( A , B , and C ) with varying degrees of stiffness and the base object were mechanically measured. (The stiffness of the objects was: C > B > A > base.) When the real static force responses were measured using a formulated value for the haptic response, the VR response (generated by a haptic device only) was adequate for object C , but still shy of the static response of object A . However, when the base object was placed in contact with the haptic device, the additional force of the base object complemented the VR to create the AR response. This AR static response closely resembled the real forces of A , B , and C . Then, the frequency responses were measured mechanically. The results show that the AR system matched the real response more closely than the VR system alone. Finally, a set of nine volunteers replaced the mechanical attachments. They measured the response of the three objects, using the real, VR, and AR setups. (The tests were performed under a hood, so the volunteers could not see the setup.) The subjects had to identify A , B , or C and indicate their degree of confidence. The results indicate that the stiffer objects were identified more reliably by the AR method than the VR method. As Kurita et al. conclude, this method could have applications in industries such as meat cleaning and even surgery, where feedback systems could train individuals to sense resistance and perform reliable procedures. The paper is interesting and could become even better with further refinements. Online Computing Reviews Service