Stephen Smith, Founder and Chief Scientist at Rapid Flow Technologies, became the 15th recipient of the Robert S. Engelmore Memorial Award in 2018. This award, which is jointly sponsored by the Innovative Applications of Artificial Intelligence (IAAI) Conference and AI Magazine, is awarded annually for sustained research contributions to applied artificial intelligence and service to the AI community.
Awardees are invited to give the keynote talk at the annual IAAI Conference. Smith’s talk focused on the research that has produced the Surtrac intelligent traffic control system and is currently driving its integration with connected vehicle technology. An expanded version of this talk has just been published in the Spring 2020 issue of AI Magazine, the flagship publication of the Association for the Advancement of Artificial Intelligence (AAAI).
The article opens with a basic explanation of the differences between conventional signal systems, which rely on fixed, preprogrammed timing plans, and adaptive traffic signal systems, which sense approaching traffic and adjust timing plans accordingly.
Smith then goes on to discuss Surtrac, an adaptive traffic signal control system that combines concepts from artificial intelligence and traffic theory to generate timing plans in real-time that optimize for the actual traffic on the road. Surtrac takes a decentralized, online planning approach, wherein each intersection continuously senses approaching traffic and optimizes accordingly. Intersections then communicate their timing plans with their neighbors to achieve coordination at the traffic network level.
Surtrac was designed for complex urban road networks, and its initial deployment in Pittsburgh has seen immense success. Originally deployed at nine intersections, stops were reduced by over 30%, and wait times were reduced by 40%. The strength of these results led to several expansions and Surtrac now controls a network of over 50 intersections in Pittsburgh, with a plan to add 150 more in the coming years.
Smith’s article also covers work aimed at integrating Surtrac signal control with emerging connected vehicle technology, which is currently a major focus for Rapid Flow Technologies. Surtrac can communicate with vehicles equipped with the proper radio technology to learn their current locations, speeds, and the directions they are heading. The principal motivation for connected vehicle technology is traffic safety, but this work has shown that vehicle-to-intersection (V2I) communication can further improve the flow of traffic when used in conjunction with Surtrac.The full article in AI Magazine (based on Smith’s 2018 lecture) can be accessed here: Smart Infrastructure for Future Urban Mobility