Perception-aware packet-loss resilient compression for networked haptic systems.

Abstract

Haptic systems are used increasingly in various applications such as virtual training, remote presence and telesurgery. The effectiveness of such systems depends greatly on the human perceptual characteristics, the nature of the application being implemented (precision needed, speed and movement type involved, etc.) and the characteristics of the communication channels used. In addition, human perception characteristics play a major role in the efficient design of perceptual compression methods for haptic systems. However, the performance of these methods is jeopardized when packet loss is introduced in the network. This thesis first assesses the variations in perceptual limitations of the dominant and non-dominant hand, when engaged in a synchronous movement, as no previous work has assessed the influence of such movement on the force perception. Results showed that both hands are less sensitive to force variation when operating together. Moreover, this thesis presents two packet loss resilient perceptual compression methods, the “magnitude –based adaptive deadband “ which modifies the value of the deadband based on the magnitude of the force, and the “modified prediction deadband” method, a variation of the published linear prediction deadband method. The proposed methods showed a significant improvement in the user experience compared to other previously published methods, without a considerable loss in the compression ratio. Moreover, these two methods present many advantages over other published error resilient compression methods, since they have a low footprint on the computational cost and on the memory consumption of the system. In addition, they do not require any previous knowledge or statistics regarding the state of the network.