Semantic Communication for Multiple Vibrotactile Sensors

Sending tactile information over wireless channels will pave the way for the realization of immersive Extended Reality (XR) applications. Existing vibrotactile communication solutions have integrated traditional source and channel coding to provide high-quality vibrotactile information over band-limited and error-prone wireless channels. However, due to the bit-error vulnerability of source and channel coding and the unrecoverable distortion of source coding, the existing solutions suffer from catastrophic quality degradation. This paper proposes a novel tactile communication scheme inspired by semantic communication. The main contribution to the existing solutions is to replace the source and channel coding with the semantic vibrotactile coder and transceiver. Specifically, the semantic encoder compresses the vibrotactile information using a two-dimensional discrete cosine transform (2D-DCT) without binarization, and the transmitter directly maps the DCT coefficients to the transmission symbols with the optimized power allocation to avoid the catastrophic quality degradation due to the traditional source and channel coding and to maximize the vibrotactile quality under the wireless channel condition. In addition, we design an overhead-less semantic receiver to reduce the communication overhead required for the optimized power allocation. We implement a vibrotactile acquisition system for empirical evaluations, and the results showed that the proposed scheme provides better vibrotactile quality even in low-quality and high-quality wireless channels compared to the tactile communication schemes with the traditional source and channel coding.