Uav base stations use learned models to follow shifting traffic demand
DSWM: Decomposed Spatio-Temporal World Model for Demand-Driven UAV Base Station Repositioning
Networking and Internet ArchitectureArtificial IntelligenceMachine Learning
Summary
Mobile base stations carried by drones (UAVs) need to move around to serve people as traffic demand changes by time and place. The authors created a model called DSWM that helps drone controllers predict where demand will be and plan their moves accordingly. Instead of just reacting or solving new problems each time, this model learns patterns and imagines future scenarios to make smarter moves. Tested on real-world data from three cities, their approach helped drones serve more user requests than other methods.
What this means in practice
- •For telecom network operators: Improve drone-based mobile base station placement to cover dynamically changing user demand more effectively in urban areas.
- •For wireless network infrastructure teams: Coordinate fleets of aerial base stations by using learned models to quickly replan positions based on recent observations.
Authors
Shengjie Zhong, Zhongliang Zhao, Jingxuan Chen, Xianbin Cao, Xinmei Qiang, Dapeng O. Wu, Tony Q. S. Quek
Abstract
Uncrewed aerial vehicle base stations (UAV-BSs) are expected to cover traffic demand that shifts across space and time, yet most repositioning schemes either re-solve an optimization problem per slot or learn reactive policies without an explicit demand model. We cast demand-driven fleet repositioning as latent-space decision-time planning and propose DSWM, a decomposed spatio-temporal world model: an agentic controller that perceives the demand field through a rolling observation window, retains operational context in a latent recurrent state, reasons about candidate motions by imagined rollouts under an uncertainty penalty, and coordinates the fleet through replanned first actions. DSWM learns a recurrent state-space model shaped by an exponential-moving-average (EMA) based latent predictive objective with variance regularization. It attaches a differentiable service simulator that replays the association, probabilistic line-of-sight channel, and Shannon rate chain inside latent rollouts. Planning uses a cross-entropy method whose imagined demand is anchored on the current observation window with mixing coefficient $ρ=0.95$. On a unified pipeline over three real datasets (Milan CDR (call detail record), Shanghai Telecom, YJMob100K) and 14 methods including five reproduced IEEE baselines, DSWM attains weekday served ratios of 0.889, 0.908, and 0.898, ranking first among non-ablated configurations on every dataset. On Milan it improves over the strongest non-learning baseline (Greedy, 0.780) by 0.109, a margin that comes from decision-time use of observations rather than prediction accuracy.