MIT researchers create AI system that reduces urban traffic congestion by 30% in simulations.
MIT researchers have developed an AI system capable of optimizing urban traffic flow in real-time, showing potential to reduce congestion by 30% in city simulations.
The system uses deep reinforcement learning to analyze traffic patterns and adjust signal timings dynamically, taking into account factors like time of day, weather, and special events.
"Urban traffic congestion costs billions annually in lost productivity and emissions," said the lead researcher. "Our system could significantly improve traffic flow without requiring expensive infrastructure changes."
The AI system has been tested in simulation using data from several major US cities, with promising results. Researchers are now in discussions with city governments to conduct pilot programs.
The research was funded in part by the US Department of Transportation, which sees AI traffic optimization as a key component of smart city initiatives.
