NeuroBCI: Multi-Brain to Multi-Robot Interaction through EEG-Adaptive Neural Networks and Semantic Communications

Recent advancements in EEG-based BCI technologies have been explored to assist individuals in executing brain-to-robot tasks. Anticipated is the widespread adoption of EEG devices in the future, as an increasing number of individuals are likely to gravitate towards using BCI systems in both personal and professional domains, such as in daily life or manufacturing environments. One promising application is facing home life, to enable people using commercial EEG equipment to implement daily tasks through the portable BCI system. However, current BCI studies mainly focus on single-brain-to-single-robot interaction, which has limitations in representing diverse human intentions. One conventional EEG device can recognize limited commands; nevertheless, a BCI system with multiple EEG devices allows the system to achieve more commands to precisely or collaboratively control robots. Therefore, it is imperative to extend the BCI techniques into a future collaboration scenario. In this paper, we present a new system, NeuroBCI, for multi-brain-to-multi-robot interaction through the integration of sensing, computing, communication, and control. To improve sensing efficiency, NeuroBCI employs a sparse attention mechanism to extract joint features from heterogeneous EEG data. Parallel computation and transmission for multi-user multi-task scenarios are handled by semantic autoencoder and autodecoder communications. Correspondingly, a versatile code map is designed to achieve concurrent control from multi-brain to multi-robot, and a series of model compression methods are used to address edge computing on both transmitter and receiver sides. Our experiments in comparison with state-of-the-art works show the superior performance of NeuroBCI on sensing, computing, communication, and control as a holistic system.