Hybrid signal-based and geometry-based prediction for haptic data reduction

Haptic data reduction schemes address the high packet-rate requirements of networked haptics. Perception-driven predictive coding approaches enable strong packet rate reduction while keeping the introduced distortion below human haptic perception thresholds. The performance of predictive coding is strongly influenced by factors such as human behavior, system characteristics, geometric and impedance properties of the environment, etc. In this paper, we first describe a novel surface geometry-based prediction approach for haptic data reduction where local object surface features are approximated with the help of simple geometric models. Secondly, we present a hybrid framework that combines signal-based and geometry-based prediction. Psychophysical experiments are performed to validate this framework. The results of the proposed geometry-based prediction show an improvement in haptic data reduction of about 54% as compared to the signal-based prediction (linear predictor). Furthermore, the presented hybrid prediction technique allows for an additional gain of 15%.