DSWM: Decomposed Spatio-Temporal World Model for Demand-Driven UAV Base Station Repositioning

Shengjie Zhong; Zhongliang Zhao; Jingxuan Chen; Xianbin Cao; Xinmei Qiang; Dapeng O. Wu; Tony Q. S. Quek

cs.NISep 29, 2026

DSWM: Decomposed Spatio-Temporal World Model for Demand-Driven UAV Base Station Repositioning

Authors: Shengjie Zhong, Zhongliang Zhao, Jingxuan Chen, Xianbin Cao, Xinmei Qiang, Dapeng O. Wu, Tony Q. S. Quek

Organizations: School of Electronic and Information Engineering, Beihang University, Beijing 100191, China

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.

Figures & tables

Fig. 1: Scenario and motivation. A fleet of $K$ UAV-BSs at altitude $h_{\mathrm{fly}}$ covers a gridded demand map $x_{t}$ that drifts over the day (commute bimodal pattern). Because the demand map is strongly structured in space and time, the bottleneck for coverage is how the current observation window is exploited at decision time, not raw forecasting accuracy. DSWM therefore plans in a learned latent space with the imagination anchored on the current observation.

Symbol	Meaning
$t$ , $T$	slot index; slots per day ( $T{=}144$ )
$T_{\mathrm{slot}}$	slot duration (10 min)
$k$ , $K$	UAV-BS index; fleet size ( $K{=}16$ )
$g$ , $G$	cell index; number of cells ( $G{=}400$ )
$c$	cell side length (235/470/500 m)
$x_{t}\in\mathbb{R}_{+}^{G}$	normalized demand map

TABLE I: Main Notation

Fig. 2: DSWM framework. The encoder and RSSM maintain a latent state from the observation window. The CEM planner imagines $H$ -slot rollouts, anchors the imagined demand on the current observation with mixing coefficient $\rho=0.95$ , scores each rollout with the decomposed service head minus an uncertainty penalty, and executes the first action. Training (dashed) is offline over replayed sequences.

Fig. 3: Decomposed demand-service physics head. An $E$ -head ensemble predicts the next demand map in symlog space; a differentiable service simulator replays association $\rightarrow$ probabilistic LoS ( 1 ) $\rightarrow$ Shannon rate ( 3 ) $\rightarrow$ capped per-cell fulfillment $\rightarrow$ aggregate $\widehat{\mathrm{SR}}$ ( 17 ). Ensemble disagreement defines the uncertainty $u_{t}$ .

Fig. 4: Observation-anchored CEM planning. Candidate action sequences are imagined for $H=15$ slots through the RSSM prior; at each imagined step the demand estimate is anchored on the current observation with mixing coefficient $\rho=0.95$ ( 20 ); rollouts are scored by $J$ in ( 19 ); elites refit the sampling distribution and the first action is executed.

Parameter	Value	Basis
Fleet size $K$	16	city-wide; $\{2,8\}$ scaling
Altitude $h_{\mathrm{fly}}$ / speed $V_{\max}$	100 m / 25 m/s	low-altitude [ 5 ]
Move limit $D_{\max}$ / slot	1500 m / 10 min	per-slot motion limit
Carrier $f_{c}$ / bandwidth $B$	2 GHz / 20 MHz	sub-6GHz access
Power / noise $N_{0}$	0.5 W / $10^{-20.4}$ W/Hz	small-cell
Rate demand $r_{0}$ / threshold $U_{\mathrm{peak}}$	0.5 Mbps / 0.8	per-user

TABLE II: Scenario Parameters and Justification

Method	Milan	Shanghai	YJMob
DSWM (ours)	0.889/0.903	0.908/0.862	0.898/0.891
TD-MPC	0.769/0.805	0.875/0.839	0.818/0.831
PPO	0.723/0.791	0.810/0.748	0.879/0.870
GA-MATR	0.770/0.854	0.745/0.708	0.820/0.790
Greedy	0.780/0.792	0.746/0.741	0.762/0.760
INS-WOA	0.744/0.752	0.816/0.788	0.753/0.756

TABLE IV: Cross-Dataset Results: Weekday / Holiday Served Ratio (mean; 5 seeds for Milan, 3 seeds for Shanghai and YJMob; columns: Milan (2013, CDR), Shanghai (2014, sessions), YJMob (2023, 4.5 km)). Per-seed std where it affects adjacent rankings: DSWM

\pm

0.006 (Shanghai),

\pm

0.002 (YJMob); GA-MATR

\pm

0.058 (Shanghai),

\pm

0.039 (YJMob).

Fig. 14: World-model prediction visualization on Milan. From left to right: full-city context, ground-truth demand map, predicted demand map, and per-cell error. Predictions track the commute-driven hotspot migration; errors concentrate at cell boundaries during peak transitions.

Fig. 15: Shanghai dataset visualization. From left to right: full-region demand, ground truth, weekday profile, and weekend (OOD) profile. The weekday–weekend shift is a spatial redistribution of sparse sessions (85.3% zero cells), not a uniform intensity change.

Fig. 16: YJMob100K dataset visualization. From left to right: full-region demand, ground truth, weekday profile, and pseudo-weekend (OOD) profile. Demand semantics are distinct-user presence counts (a 5% sampled mobility proxy), as declared in Section V-A ; weekend labels are inferred from the autocorrelation-recovered week structure.

Explore similar work

Reinforcement Learning for Delivery Drone-Based Participatory Sensing in Dynamic Environments
Jul 21, 2026Xin Ouyang, Songxin Lei, Xusen Guo +3Dynamic Environments Urban Environments
The Hong Kong University of Science and Technology (Guangzhou) Guangzhou, China · Beijing Institute of Technology Beijing, China
Learning to Optimize UAV Path Planning for Data Sensing in Wireless Sensor Networks
Sep 15, 2026Sijie Ma, Zeyuan Ma, Weijia Cao +4Unmanned Aerial Vehicle Trajectories Path Planning
South China University of Technology · South China Normal University · Information Research Institute, CAS +2
Uncertainty-Aware World Model for Aerial Image-Goal Navigation
Aug 6, 2026Deyi Zhu, Haoyu Fan, Yinan Zhu +4Aerial Vision-Language Navigation Unmanned Aerial Vehicles
Tsinghua Shenzhen International Graduate School, Tsinghua University

Jul 21, 2026, cs.LGJ/K move · Enter open · S save

Reinforcement Learning for Delivery Drone-Based Participatory Sensing in Dynamic Environments

Xin Ouyang, Songxin Lei, Xusen Guo, Yutian Jiang+2

The Hong Kong University of Science and Technology (Guangzhou) Guangzhou, China · Beijing Institute of Technology Beijing, China

Dynamic Environments Urban Environments Crowdsourced Framework

Open paper AskPDF arXiv

Using Unmanned Aerial Vehicle (UAV) for urban sensing has emerged as a powerful paradigm to monitor the status of the city, e.g., air quality and noise levels, through agile aerial crowdsourcing. Despite this potential, existing UAV-based sensing approaches overlook environmental disturbances like wind that drastically impact drone velocity and energy efficiency. Consequently, directly applying existing methods to this joint delivery and sensing paradigm in dynamic environments faces two severe challenges: (1) scalability bottlenecks as fleet sizes expand; and (2) multi-timescale decision heterogeneity between macro task dispatching and micro velocity control. To tackle these, we formalize the problem as SensUAV and propose a Two TimeScale Reinforcement Learning framework (TSRL). Specifically, TSRL separates decision-making into two cooperative layers. At the macro level, a task-embedding sensing dispatcher handles scalability by separately encoding distinct task features and sequentially evaluating UAV suitability before task selection. At the micro level, a wind-aware velocity controller learns fine-grained velocity scheduling to adapt to dynamic environmental variations. Extensive experiments on real-world datasets demonstrate that TSRL significantly outperforms baselines, achieving average system profit improvements of 20.1% in Hangzhou and 46.6% in Shanghai.

Abstract

Explore similar work

Reinforcement Learning for Delivery Drone-Based Participatory Sensing in Dynamic Environments

Learning to Optimize UAV Path Planning for Data Sensing in Wireless Sensor Networks

Uncertainty-Aware World Model for Aerial Image-Goal Navigation