Robot vehicles can optimise traffic flow in cities, even when mixed with human-driven vehicles, improving traffic efficiency, safety and energy consumption, my colleagues and I discovered.
Robotic vehicles aren’t any longer a science fiction concept: cities around the globe have tested autonomous robot taxis since 2016Given the increasing presence of robotic vehicles on the road and the foreseeable long transition period from mixed traffic to totally autonomous traffic, my team and I asked ourselves whether robotic vehicles and their interaction with human-driven vehicles could alleviate today's well-known traffic problems.
I’m a pc scientist of artificial intelligence for transportation and smart cities. My colleagues and I hypothesized that because the variety of robotic vehicles in traffic increases, we are able to use AI to develop algorithms to regulate the complex mixed traffic system. These algorithms wouldn’t only be sure that all vehicles travel easily from point A to point B, but more importantly, they’d optimize all traffic by allowing robotic vehicles to influence human-driven vehicles.
To test our hypothesis, we used a branch of AI often called reinforcement learningwhere an intelligent agent learns to maximise cumulative rewards by interacting with its environment. By setting rewards for simulated robot vehicles to prioritize goals equivalent to traffic efficiency or energy consumption, our experiments show that we are able to effectively Manage mixed traffic at complex real-world intersections under real traffic conditions within the simulation.
Our algorithm teaches the robocars to optimize traffic flow by communicating with one another. The collective system of cars goals to make sure smooth traffic flow, regardless that each individual automotive decides when to enter an intersection based on its immediate surroundings. Since the robocars are distributed amongst cars driven by humans, all traffic is influenced by the algorithm.
We found that robotic vehicles only account for five% of traffic in our simulation. Traffic jams are eliminatedSurprisingly, our approach even shows that when robotic vehicles account for 60% of traffic, traffic efficiency Better than traffic controlled by traffic lights.
Wang et al
Why it will be important
Traffic is getting worse in all major cities around the globe. which results in significant economic and environmental costs. It is one in all the most important problems facing society today. Current methods of traffic control, equivalent to traffic lights, have limited effectiveness in reducing delays and congestion.
AI-controlled robotic vehicles offer a possible solution, but existing studies require universal connectivity and central control of all robotic vehicles, a scenario that’s unlikely to occur within the foreseeable future. The transition to totally autonomous traffic is more likely to happen regularlyleading to a chronic period of mixed traffic consisting of each robot and human-controlled vehicles.
This has led us to develop control algorithms that use robotic vehicles to understand the societal advantages of autonomous transport systems without requiring all and even the vast majority of vehicles to be autonomous.
What other research is being carried out
Recent studies have shown Potential of mixed traffic control in scenarios equivalent to ring roads, figure-eight roads, motorway bottlenecks and mergers, intersections and roundabouts. However, these scenarios are frequently not as complex as in the actual world and only involve a limited variety of vehicles that should be coordinated.
Our work is the primary to display the feasibility of mixed-traffic control by robotic vehicles at real-world, complex intersections. The ability to regulate traffic at these intersections is a vital step toward city-wide traffic control.
What's next
We plan to increase our framework to include additional driving behaviors for robot vehicles, equivalent to frequent lane changes. We also plan to check our approach at different intersection types, and we would love to check our approach under real-world conditions of vehicle-to-vehicle communication.
Our goal is effective and efficient management of mixed traffic at city level.
image credit : theconversation.com