WISP is an AI-oriented solution to improve traffic flow in large cities. Without additional infrastructure, the aim is to optimize traffic light switching times, resulting in better air quality, less noise pollution and considerable time savings.SUPPORT OUR PROJECT
WISP is an artificial intelligence solution applied to road traffic in large cities.
We believe that a dynamic traffic light management system can benefit millions of inhabitants in different ways. By adding dynamism to what already exists, we are fighting for :
Better air quality
Less noise pollution
Timesaving for workers and parents who don’t want to waste their time in their daily transportsOUR SIMULATION
ONE day in A ParisIAN REGION
Millions of car rides
Travel hours saved
tonnes of CO2 reducement
-WISP is a deepQ-Learning agent that selects the right sequence of traffic lights over multiple intersections to maximize traffic efficiency
The solution we have adapted has also been used for DeepBlue (1996), AlphaGo (2015) and OpenAI (2020).
"Who would have thought about it ? WISP could be one of the first concrete application of reinforcement learning that goes beyond the world of games"
AN AMBITIOUS PROJECT THAT WE CAN ONLY ENCOURAGE ! THIS ONE WOULD ALLOW US TO GET A DYNAMIC GRIP ON THE TRAFFIC OF OUR CITY
Department of Public Space and Living Environment
Vincennes City Hall
Today, WISP's AI is able to find complex relationships between road network congestion and traffic light phasing sequences, and applies its results in real-life situations in a dozen or so territories.
WISP is now working on adapting its agent to any road network.
WISP is constantly innovating, achieving new objectives such as predicting the remaining waiting time for each driver when stopped at a red light.DOWNLOAD OUR DECK