Usability studies are an essential and iterative component of technology development and ease its transfer from the laboratory to the clinic. Although such studies are standard methodology in today's graphical user-interface applications, it is not clear that current methods apply to new technologies such as virtual reality. Thus experimentation is needed to examine what existing methods can be viably transferred to the new user-interaction situations. In this paper, 5 integrated interfaces with 3 simultaneous users are evaluated via a set of usability studies, which adapt traditional methods for assessing the ease of use of the interface design. A single expert domain user was run in an intensive study that examined the therapist manual and interfaces of the Rutgers Ankle Rehabilitation System (RARS). The interface and manual were extensively modified based on this evaluation. A second study involving 5 therapists was then conducted to evaluate the telerehabilitation component of the RARS system. In both studies, the tester and developer's observations, along with the session videotapes and therapist-user questionnaires, were triangulated to identify user problems and suggest design changes expected to increase the usability of the system. Changes that resulted from the analysis with the domain expert are described and recommendations for how to conduct usability studies in such multiuser remote virtual reality situations are proposed. Results from the pilot usability telemonitoring studies are also presented. The validity of usability studies in the development and refinement of rehabilitation technology is highlighted.

We developed a novel interface that gives upper-limb amputees a virtual hand that can manipulate objects in a challenging environment. The interface registers specific myokinetic activity of the residual limbs, and encodes the intended voluntary movements that are then actualized as virtual hand motions. The composite myokinetic interface-virtual reality (MKI-VR) system consists of an array of pressure sensors mounted in an arm sleeve, sensors of elbow- and shoulder-joint angles, a trained filter derived from the pseudoinverse of a response matrix, and a virtual hand model, programmed in Java 3D. Users can manipulate virtual objects such as balls and pegs in a 3D training environment, while their performance at various difficulty levels is scored. In preliminary tests, upper-limb amputees readily gained the ability to grasp and release virtual objects. We propose the utility of the MKI-VR system both as an assessment tool for rehabilitation engineers, and as a motivator for amputees to exercise and thereby maintain their residual motor ability.

This article describes a new and unified approach to computerized hand diagnosis and rehabilitation. The system uses a workstation to automate diagnosis data collection and analysis and to assess the rehabilitation progress. A new diagnosis glove was developed and tested. This device measures grasping forces applied to 16 regions of the hand. A physician using this system can also utilize modern diagnosis devices such as electronic dynamometer, pinchmeter, and goniometer. Additionally, three VR rehabilitation exercises were created using WorldToolKit graphics library and run on the same workstation. These exercises were modeled after standard hand rehabilitation procedures and involve manipulation of virtual objects and transparent real-time data gathering. Grasping forces were modeled and fed back using the Rutgers Master worn on the patient's hand. An Oracle database was used to store, analyze, and integrate the patient's diagnosis and rehabilitation data. The system is presently undergoing clinical trials.

Research on virtual environments (VE) produced significant advances in computer hardware (graphics boards and i/o tools) and software (real-time distributed simulations). However, fundamental questions remain about how user performance is affected by such factors as graphics refresh rate, resolution, control latencies, and multimodal feedback. This article reports on two experiments performed to examine dextrous manipulation of virtual objects. The first experiment studies the effect of graphics frame rate and viewing mode (monoscopic vs. stereoscopic) on the time required to grasp a moving target. The second experiment studies the effect of direct force feedback, pseudoforce feedback, and redundant force feedback on grasping force regulation. The trials were performed using a partially-immersive environment (graphics workstation and LCD glasses), a DataGlove, and the Rutgers Master with force feedback. Results of the first experiment indicate that stereoscopic viewing is beneficial for low refresh rates (it reduced task completion time by about 50% vs. monoscopic graphics). Results of the second experiment indicate that haptic feedback increases performance and reduces error rates, as compared to the open loop case (with no force feedback). The best performance was obtained when both direct haptic and redundant auditory feedback were provided to the user. The large number of subjects participating in these experiments (over 160 male and female) indicates good statistical significance for the above results.

Dextrous masters control robots and artificial environments through hand gestures. Commercial products have open-loop control, without force feedback to the operator. There is a need for portable systems that have force feedback, but are still sufficiently compact to be desktop. In this paper we discuss a prototype master providing force feedback for the Dataglove. The master structure and its actuator characteristics are presented first. Then a control model is given based on finger parameters and joint coupling. The glove calibration is subsequently discussed, taking into account the influence of the feedback structure. The experimental setup and initial results are presented last.