DeCoR v0.1.0

DeCoR v0.1.0

This release provides the code for DeCoR: Design and Control Co-Optimization for Urban Streets Using Reinforcement Learning.

DeCoR is a two-stage reinforcement learning framework for co-optimizing mid-block crosswalk placement and adaptive traffic signal control using pedestrian and vehicle flow observations. The repository includes SUMO-based simulation environments, PPO training code for the design and control agents, evaluation scripts, plotting utilities, and configuration files for reproducing the experiments.

Included

• Training and evaluation pipeline for DeCoR
• Pedestrian and vehicle dataset in the corridoor
• SUMO simulation environments for design and control
• PPO implementation for the design and control agents
• Configuration files

Reproducing results

SUMO 1.21, 1.22 is recommended for reproduction.

The README provides setup instructions for SUMO, Python 3.12, and uv, along with commands for training, TensorBoard monitoring, and evaluation.