Manual Resolution of Compliance When Work and Force Cues Are Minimized

This paper summarizes new experiments on compliance discrimination in which work cues were eliminated and force cues were minimized. The average JNDs for compliance ranged from 15% to 99% and were much larger than the average JND (8%, see Tan, Pang & Durlach, 1992) obtained from previous compliance discrimination experiments in which both work and terminal force cues, as well as compliance cues, were available to the subject. By converting results to corresponding JNDs in terminal force, we obtained a force JND of 5.2% ±0.8%; this value was found to be consistent with force JNDs (6-8%, see Pang, Tan & Durlach, 1991; Tan et al., 1992) we had obtained earlier from force discrimination experiments. We conclude from these compliance experiments and our previous experiments on force and compliance perception that manual resolution of compliance deteriorates when force and/or work cues are reduced or eliminated. INTRODUCTION Our interest in the study of manual resolution of physical parameters such as length, force, and compliance are twofold: (1) we want to measure how well humans can discriminate these parameters and thus understand performance in manual object recognition; (2) we want to apply the knowledge gained from these psychophysical studies to the design of better haptic interfaces. When one designs and constructs haptic interfaces for teleoperation or virtual environment systems, it is important to match the resolution of the haptic display to that of the human sensory systems. Compliance is one of the many important object properties (others are shape, mass, viscosity) that can be effectively perceived through the sense of touch. For example, a telediagnostic interface requires realistic display of compliance during palpation. Compliances associated with normal and abnormal health conditions as revealed from palpation should be displayed in a manner that facilitates easy discrimination by the physician at the remote site. In general, our continuing work on manual resolution of object properties provides quantitative guidelines for engineering design. Our earlier work can be found in Durlach, Delhorne, Wong, Ko, Rabinowitz & Hollerbach (1989; length resolution), Pang et al. (1991; force resolution), and Tan et al. (1992; a review). In an earlier paper (Tan et al., 1992), we reported the degrading effect of roving displacement on both force and compliance discrimination. Average compliance JNDs (justnoticeable-differences) were 22% and 8% for roving and fixeddisplacement experiments, respectively. Average force JNDs were 14% and 6% for roving and fixed-displacement experiments, respectively. Further analysis of the data from both force and compliance experiments with roving-displacement seemed to support a “Work Hypothesis”; i.e., subjects tended to respond “bigger force” or “smaller compliance” when the work involved in the active finger motion was greater than some fixed criterion work value. The data from compliance discrimination experiments with roving-displacement also seemed to be consistent with a “Terminal Force Hypothesis”; i.e., subjects tended to respond “smaller compliance” when the terminal force was greater than some fixed criterion value of force. In the design of our earlier compliance experiments, resistance force f was related to pushing distance d by f = d C , where C is compliance (see Figure 1A). The terminal force increment ∆F F0 was equal to ∆C (C0 + ∆C). Thus for ∆C C0 of 22% and 8% (compliance JNDs for roving and fixed-displacement paradigms), ∆F F0 was 18% and 7%. These results were very close to the force JNDs for roving and fixed-displacement paradigms, respectively. Therefore, subjects could have relied on terminal force cues for compliance discrimination with this experimental design. For this force-distance profile, work