Multi-Objective, Mobility-Aware, and Lightweight Deep Reinforcement Learning for Adaptive Signal Processing in Real-Time 6G Edge Networks

As sixth-generation (6G) wireless networks advance, real-time adaptive signal processing is critical for mobility and multi-objective constraints. Traditional methods fall short in dynamic edge environments. We propose a DRL model with dual-agent actor-critic, convolutional features, and mobility forecasting for tasks like modulation and power allocation. It learns in real-time, reducing end-to-end latency by up to 27% relative to tabular RL baselines and by 20% relative to A3C, improving energy efficiency by 13–16% compared to fixed-weight and mobility-unaware configurations, and maintaining low bit error rates in vehicular and aerial simulations. This enables applications in transportation, drones, and IoT.