Wireless human activity sensing has drawn significant interest from researchers due to its potential for transformative applications across diverse domains. However, conventional Radio Frequency (RF) signal-based and acoustic-based approaches face several challenges, including (i) interference with existing communication signals, (ii) dependence on dedicated devices and signals, and (iii) potential health concerns from prolonged exposure. Recently, researchers have explored using optical signals (e.g., LED light signals) for human gesture recognition, achieving promising performance. However, these approaches require modifications to existing lighting infrastructure to generate dedicated light signals and multiple optical receivers. To address these challenges, in this paper, we propose LightTells, an innovative single activity sensing system for one active human object that uses a single optical receiver to capture and analyze ambient light signals without modifying the existing LED lighting infrastructure. To achieve this, LightTells first analyzes the received ambient light signals and extracts activity-correlated information. Based on these features, we design a multi-modal human activity recognition model to recognize the corresponding human activity. We implement LightTells using commercial off-the-shelf components and evaluate its performance by conducting extensive evaluations under various real-world scenarios, including common indoor spaces (e.g., Lab, classroom, and lobby lounge), and RF-restricted hospital environments. The results demonstrate that LightTells is agnostic to LED sources, light receivers and its placement positions, human objects (e.g., type and clothing), and sensing distance, achieving average recognition accuracy of up to \(94.25\%\) for 12 common human activities. Furthermore, two practical use-cases: fall detection and fatigue detection, showcase LightTells’s effectiveness, achieving up to \(100\%\) accuracy in both scenarios.

Hybrid visible light communication (VLC)/radio frequency (RF) networks have emerged as a promising next-generation networking technology that can provide high data rate and wide coverage while overcoming blockage problems of VLC and the limited available radio spectrum of RF. However, there has been limited research on the joint user association and power allocation problem in indoor hybrid VLC/RF networks with consideration of networking dynamics. To bridge the gap, we first formulate a joint user association and power allocation problem with the objective of maximizing the sum throughput of the hybrid networks with consideration of user mobility, time-varying blockages and interference between VLC access points (APs) and users, and practical capacity limitations of APs. Then we propose a distributed cooperative proximal policy optimization (DC-PPO) algorithm to solve the formulated mixed integer nonlinear programming problem, where each user is modeled as a hybrid agent that distributedly determines the user-AP association (discrete action) and the desired transmission power (continuous action) variables simultaneously based on local observations and limited information exchanged between neighboring nodes. The performance of our proposed method is compared with three baselines, including deep Q-network (DQN), max-power-max-signal-to-noise ratio (SNR), and random-power-max-SNR. Simulation results show that the proposed method is mobility and blockage-ware and outperforms the baseline solutions in terms of optimality, convergence, scalability, and robustness. Specifically, our proposed method can achieve $127 \%$ throughput improvement compared with that of DQN in 30mobile-user scenario.

Hybrid Light-Fidelity (LiFi) and Wireless-Fidelity (WiFi) indoor networking has been envisioned as a promising technology to alleviate radio frequency (RF) spectrum crunch to accommodate the ever-increasing data rate demand in indoor scenarios. The hybrid LiFi/WiFi indoor networks can utilize the advantages of fast data transfer from LiFi and wider coverage of WiFi, thus complementing well with each of these technologies and increasing the network performance compared to the standalone networks. This work investigates the user-access point (AP) association problem to achieve better network-wide spectral efficiency. We first derive a mathematical formulation of the sum data rate maximization problem by determining the AP selection for each user in indoor networks with either static or mobile users. It is shown that the formulated network control problem is a mixed integer nonlinear nonconvex problem. A proximal policy optimization (PPO) based reinforcement learning method is then proposed to determine the AP assignment. The network performance is evaluated by comparing it with exhaustive search, signal strength strategy, and trust region policy optimization methods. The results indicate that the designed method achieves a network-wide data rate with up to 100% and 97.7% optimality in static and mobile user scenarios, respectively.