Reinforcement Learning (RL) is a machine learning paradigm in which an agent learns optimal decision-making policies by interacting with an environment, receiving scalar reward signals, and iteratively updating its behavior to maximize cumulative long-term reward. In the context of WiFi CSI sensing, RL is particularly relevant for adaptive resource management tasks such as dynamic duty-cycle scheduling, transmission power control, and intelligent sensor activation, enabling systems to balance sensing accuracy against energy consumption without requiring pre-labeled datasets. Key variants include model-free approaches such as Q-learning and Deep Q-Networks (DQN), policy gradient methods, and actor-critic architectures, with deep RL variants being especially valuable when the state and action spaces are high-dimensional, as is common in multi-antenna CSI environments and multi-agent sensing deployments.
Source Papers
- A Survey on Green Wireless Sensing: Energy-Efficient Sensing via WiFi CSI and Lightweight Learning ↗ — A Survey on Green Wireless Sensing: Energy-Efficient Sensing
- Data Assimilation for Agent-Based Models ↗ — Data Assimilation for Agent-Based Models
- Radio Radiance Field: The New Frontier of Spatial Wireless Channel Representation ↗ — Radio Radiance Field: The New Frontier of Spatial Wireless C