Optimizing Energy-Distortion Trade-Off for BCI-enabled Metaverse

The integration of Brain-Computer Interfaces (BCI) within the burgeoning Metaverse promises a new era of immersive and interactive virtual experiences. Meanwhile, the wireless Electroencephalogram (EEG) system has facilitated portable and continuous monitoring in healthcare and BCI applications, supporting simultaneous use by multiple users. However, two pivotal challenges remain apparent: devices are constrained by the small energy capacities of micro batteries, and wireless data transmission is restricted by limited bandwidth. Therefore, efficient data compression and effective wireless resource management are essential to cope with these limitations. In this paper, we introduce a Convolutional Neural Network (CNN)-based network to perform EEG data compression and reconstruction, and we empirically study the trade-off between compression ratio and reconstruction distortion. Building on our analysis, we formulate an optimization problem that simultaneously minimizes reconstruction distortion and energy consumption across multiple devices while satisfying delay constraints. This involves the joint optimization of compression ratio, bandwidth allocation and transmission power selection. To address the long-term constraint, we apply the Lyapunov optimization method, which decouples the problem into real-time optimization tasks solved by our proposed iterative procedure. Simulation results show that our proposed scheme outperforms the other three baselines, benefiting from the joint optimization of compression ratio scheduling, bandwidth allocation, and transmission power control. This approach enhances the efficacy and applicability of wireless EEG systems in real-world scenarios.