Smarter Traffic with Artificial Intelligence
Our goal is to eliminate traffic jams by creating traffic light intersections that learn from their environment. By leveraging machine learning, computer vision, and the Internet of Things, we can improve the congestion efficiency by creating smarter intersections.
Sensors and cameras placed strategically around the intersection feed environment information to an autonomous agent in charge of controlling the traffic lights, with the goal of minimizing the amount of time vehicles and pedestrians spend stopped at a red light.
By applying this method to many traffic light intersections around the city and connecting nearby autonomous agents, we hope to greatly improve the congestion efficiency of an entire city by employing traffic lights that learn from their surroundings.
When running on a simple timer, a traffic light will unnecessarily stop vehicles, potentially delaying many vehicles when there is not much traffic coming the other direction.
Autoflow not only adapts in real-time to current traffic, but learns the patterns and behavior of its specific intersection to be plan the optimal light-switching strategy.
When emergency vehicles such as ambulances and fire engines travel through intersections, they have to slow to ensure that they don't inadvertantly create another accident.
Using image recognition and machine learning, Autoflow understands that a high priority vehicle is coming through, and can switch the lights accordingly.
When traffic lights are out of sync, it causes cars to repeatedly accelerate and stop. This is the seed that grows into a large traffic jam given enough cars.
The smart traffic intersections of Autoflow communicate with eachother. One intersection will synchronize with a nearby neighbor, signalling there is more traffic coming.
Using reinforcement learning and feature extraction algorithms, each traffic light learns the patterns and behaviors of its specific intersection, tailoring the lights to optimally fit the needs of its own traffic and adapting in real-time to changing flow.
Simply by attaching a low-cost camera to the traffic light, the image recognition technology automatically determines not just the amount of cars at the light, but can determine if there are any high priority vehicles in the vicinity.
Autoflow is able to integrate telemetry from a variety of IoT sensor sources, creating an ever more detailed visualization of the intersection environment. Sharing data between IoT devices is the way toward a smarter city.