TransferLight: Zero-Shot Traffic Signal Control on any Road-Network

TL;DR

TransferLight tackles the challenge of generalizing traffic signal control across diverse road networks and traffic dynamics by removing rigid, geometry-specific encodings and degenerate reward shapes. It introduces a log-distance reward to provide spatially-aware prioritization, and couples this with a hierarchical, heterogeneous graph neural network that encodes state across lanes, movements, and phases. Domain randomization and a decentralized, weight-tied policy enable zero-shot generalization to unseen networks, with state transition priors providing proactive decision-making. Experimental results show TransferLight outperforms baselines on unseen networks and arterial progression scenarios and even demonstrates competitive performance on real-world cases, highlighting its potential for scalable, practical intelligent transportation systems.

Abstract

Traffic signal control plays a crucial role in urban mobility. However, existing methods often struggle to generalize beyond their training environments to unseen scenarios with varying traffic dynamics. We present TransferLight, a novel framework designed for robust generalization across road-networks, diverse traffic conditions and intersection geometries. At its core, we propose a log-distance reward function, offering spatially-aware signal prioritization while remaining adaptable to varied lane configurations - overcoming the limitations of traditional pressure-based rewards. Our hierarchical, heterogeneous, and directed graph neural network architecture effectively captures granular traffic dynamics, enabling transferability to arbitrary intersection layouts. Using a decentralized multi-agent approach, global rewards, and novel state transition priors, we develop a single, weight-tied policy that scales zero-shot to any road network without re-training. Through domain randomization during training, we additionally enhance generalization capabilities. Experimental results validate TransferLight's superior performance in unseen scenarios, advancing practical, generalizable intelligent transportation systems to meet evolving urban traffic demands.

Figures (11)

• Figure 1: Our proposed traffic signal controller learns a general policy for flexible phase prediction during training. Due to the weight-tied models, we can apply the learned model to any road-network during inference.
• Figure 2: Pressure (see \ref{['eq:pressure']}) is symmetric to vehicle position translations within the lane's coordinate space. Our more expressive measure breaks this symmetry.
• Figure 3: Our hierarchical state space encoding uses a position-encoded segment-density set on the lowest level. This information is embedded and aggregated to form movement representations, which then undergo another pass to the phase level. On the phase level, we have intra-level updates, otherwise information are passed down-to-top along the directed heterogenous graph structure.
• Figure 4: Test performances (moving averages) on Cologne8 over 3600 simulated time steps.
• Figure 5: Average Travel Time on Cologne3 over 3600 simulated time steps.