A Computational Framework for Data-Driven Distributed Resilient Control of Traffic Corridors

Project Number


Project Summary

In this project, we propose to develop algorithms for distributed control and model parameter estimation for macroscopic traffic flow over freeway and arterial networks, with provable guarantees. There are several novel contributions of our proposed approach. First, we utilize the combination of System Optimum Dynamic Traffic Assignment (SO-DTA) and optimal control theory to understand the implications of the cost function and network ow dynamics on the sparsity of optimal control architecture. We also propose to develop a technique specific to traffic flow dynamics to compute sensitivity bounds on system trajectories under open-loop control derived from SO-DTA. Second, we propose to perform a rigorous analysis to characterize throughput, travel times and resilience properties for max pressure and proportionally fair traffic signal control policies under complex phase architectures, finite queue capacities, and dynamic route choice behavior. Third, we propose to extend our analysis of control strategies for freeway and arterial networks to consider output feedback policies under relatively sparse measurements. We also plan to leverage compressed sensing techniques to quantify the limitations on origin destination estimation from historical data. Our proposed approach relies on a combination of tools from traffic engineering, control theory, optimization, dynamical systems, and signal processing. Our analysis and algorithm development will be supplemented with case studies relevant to the Los Angeles areas, using a professional microscopic traffic simulator.

Project Status




Topic Area

Integrated Freight and Passenger Systems

P.I. Name & Address

Assistant Professor, Sonny Astani Department of Civil and Environmental Engineering; USC Viterbi School of Engineering
University of Southern California
3620 South Vermont Avenue
Kaprielian Hall (KAP) 254A
Los Angeles, CA 90089-2531
United States

Funding source: US DOT

Funding amount: $99,957

Start date: 1/1/2017

End date: 12/31/2017