Next Generation Intelligent Traffic Signals for the Multimodal, Shared, and Automated Future

The inefficient flow of vehicles on our roadways expend tremendous amounts of energy idling or reaccelerating their mass in stop-and-go traffic. This results in significant wasted energy, greenhouse gas emission, vehicle and roadway wear and tear, and economic inefficiency. According to the Centre for Economics and Business Research, the annual effect of traffic congestion in the U.S. includes 871 million gallons of wasted fuel, 8,610 kilotons of CO2 emitted, and 6.8 billion hours of time wasted. Despite this, 99% of U.S. intersections are controlled by a fixed timer, and incumbent adaptive traffic control technologies have only penetrated 1% of the marketplace over 35 years due to high cost and poor efficacy. We are proposing to develop the next generation Cooperative Intelligent Transport System (C-ITS) that can significantly reduce transportation network energy consumption by deploying the latest control algorithms in conjunction with emerging mobility concepts and sensing and learning technologies (i.e., Internet of Things, machine learning). The C-ITS system we envision is specifically designed to address the key hurdles, efficacy and cost, which are preventing wide- spread adoption of adaptive traffic signal capabilities and are therefore preventing system-level energy, environmental, and economic benefit. We validate and refine the core algorithms in live traffic intersections that determine real-time green splits, offsets, and cycle time to optimize network level objectives, such as throughput, energy consumption, and travel time. Traditional approaches typically use a combination of myopic intersection-level optimization and ad hoc techniques for coordination between intersections. On the other hand, our approach solves the problem at the system level, and hence enforces coordination, which also prepares the system’s level ground rules for emerging mobility concepts. A major thrust in our effort is to make these algorithms scalable. For this purpose, we adapt tools from information systems congestion control and distribution optimization to the transportation setting.