BEVLight controls traffic lights using drones with flexible views
BEVLight: Generalizable Traffic Signal Control from UAV Bird's-Eye Views
Machine Learning
Summary
Controlling traffic lights well is hard because many methods need detailed information from sensors or fixed cameras. The authors present BEVLight, a system that uses bird's-eye-view images taken from drones to decide traffic light signals. Their method understands different intersection designs and works well even on intersections it hasn't seen before. They trained BEVLight using a virtual platform where humans chose good traffic signals, and the system learned from these choices. Tests show it improves traffic flow and can adapt without extra tuning to new drone images.
What this means in practice
- •For smart city engineers: Use drone images to optimize traffic lights across different intersection types without specialized sensors or fixed cameras.
- •For traffic management operation teams: Deploy a control system that adapts traffic signal timing in real time from aerial visual data, reducing average vehicle wait times.
Authors
Maonan Wang, Yuxin Cai, Junming Liu, Zhengyan Huang, Yi Yu, Ding Wang, Yifeng Zhang, Yirong Chen
Abstract
Learning-based traffic signal control (TSC) has achieved promising performance, but most methods rely on structured traffic states that must be estimated from sensor observations in real deployments. Visual TSC learns signal decisions directly from images, yet existing methods depend on fixed roadside cameras and remain tied to specific intersection configurations and visual domains. UAVs offer on-demand bird's-eye-view (BEV) observations that jointly capture intersection topology and traffic dynamics, making them a promising sensing platform for visual TSC. We present BEVLight, a generalizable TSC framework that directly maps BEV observations to signal actions. BEVLight grounds visual features onto individual lanes and composes them along the junction topology into movement and phase representations, allowing a shared policy to score candidate phases across heterogeneous intersection layouts and signal plans. To train the policy, we develop a web-based virtual teleoperation platform and collect over 5K human control decisions together with their underlying traffic states. Simulator-provided traffic states first supervise lane- and movement-level representations, after which human phase selections and rankings supervise the phase scorer. Re-rendering the recorded trajectories under diverse appearances preserves both forms of supervision and encourages robustness to visual changes. Closed-loop experiments demonstrate zero-shot generalization across traffic demands, signal plans, and intersections: BEVLight reduces average waiting time by 7--9\% relative to perception-to-control visual baselines and matches or outperforms structured-state RL controllers that receive ground-truth traffic states, achieving the lowest waiting time in 12 of 16 scenarios. An offline study further shows that BEVLight transfers to real UAV observations without target-domain adaptation.