Differentiable Hybrid Traffic Simulation

TL;DR

This paper presents a novel differentiable hybrid traffic simulator combining macroscopic and microscopic models, enabling gradient-based traffic control and optimization, which improves scalability and efficiency over existing methods.

Contribution

It introduces the first differentiable hybrid traffic simulator with a novel intermediate conversion component and analytical gradients for enhanced scalability and integration into neural networks.

Findings

Enables gradient computation across traffic models and lanes.

Accelerates traffic simulation with analytical gradients.

Improves scalability for traffic control and learning tasks.

Abstract

We introduce a novel differentiable hybrid traffic simulator, which simulates traffic using a hybrid model of both macroscopic and microscopic models and can be directly integrated into a neural network for traffic control and flow optimization. This is the first differentiable traffic simulator for macroscopic and hybrid models that can compute gradients for traffic states across time steps and inhomogeneous lanes. To compute the gradient flow between two types of traffic models in a hybrid framework, we present a novel intermediate conversion component that bridges the lanes in a differentiable manner as well. We also show that we can use analytical gradients to accelerate the overall process and enhance scalability. Thanks to these gradients, our simulator can provide more efficient and scalable solutions for complex learning and control problems posed in traffic engineering than other existing algorithms. Refer to https://sites.google.com/umd.edu/diff-hybrid-traffic-sim for our project.