Fuzzy Evaluation Method for Haptic Perceptual Similarity

Abstract

Tactile rendering in virtual interactive scenes plays an important role in improving the quality of user experience. The subjective rating is currently the mainstream measurement to assess haptic rendering realism, which ignores various subjective and objective uncertainties in the evaluation process and also neglects the mutual influence among tactile renderings. In this paper, we extend the existing subjective evaluation and systematically propose a fuzzy evaluation method of haptic rendering realism. Hierarchical fuzzy scoring based on confidence interval is introduced to reduce the difficulty of expressing tactile feeling with deterministic rating. After the fuzzy statistics based on the membership function, we further use close-degree and transitive closure to calculate the fuzzy equivalence matrix between different tactile renderings. Fuzzy clustering is carried out to complete the comprehensive evaluation in the case of multiple indicators. Five tactile objects are used to simulate various situations of tactile rendering. The experimental results of haptic perceptual similarity evaluation show the existence of fuzziness in the subjective evaluation and verify the feasibility of the proposed method applied to multi-indicator evaluation. We also conclude that the proposed method outperforms the existing methods in terms of time cost and labor cost.