With 25% of all accidents and around 33% of fatal accidents occurring at junctions there’s opportunity to improve traffic flow to make it safer. Add to that the time wasted by people sitting stationary in the vehicles at a red light, and the associated pollution, and a solution for this starts to look like it has huge wins.
Automated traffic lights in a form that we know them have been around since the late 1800s in America, although the first traffic light signal was installed in London in 1868. Despite using induction loops in the road to detect whether traffic is there, traffic lights themselves can still be very inefficient in some circumstances.
In 2012, Peter Stone, Professor of Computer Science at the University of Texas proposed a system called Autonomous Intersection Management where vehicles communicate with one another and an artificial intelligence system at the lights themselves to generate a traffic flow that doesn’t need to stop. In some places in the world, such as Meskel Square, Addis Abeba, there are no traffic lights but traffic makes its way through in a series of slots as shown in this video.
This video of Autonomous Intersection Management explains how it works:
More recently Massachusetts Institute of Technology (MIT) Senseable City Lab, the Swiss Institute of Technology (ETHZ) and the Italian National Research Council (CNR) developed a slot-based system called Light Traffic. The researchers estimate that, if implemented, vehicle delays would be virtually eliminated.
Traffic congestion will cost the UK economy over £300 billion in the 16 years from 2013-2030. Solving this problem has huge payoffs not only for this lost productivity, but also for the lives saved, the injuries lessened and the pollution stopped.
But here’s the problem: for it to work, all vehicles must be autonomous and that could be at least twenty years away given than the average age of a car in the UK is 7.7 years and we don’t even have fully autonomous cars yet! As it’s unlikely we’ll have fully autonomous motorcycles, it may be impossible, not even taking into consideration the massive cost to equip every set of traffic lights with the technology.
How will the system operate?
While we have the technology to create this situation right now, every vehicle would need to be fully autonomous, so let’s assume that’s the case and look at what kind of challenges would need to be overcome to make this functional. Let’s also assume that pedestrians and cyclists can, for the most part, be eliminated from the junction. The vehicles approaching the junction, and the artificial intelligence system at the junction, would need to know:
- How fast is the approaching vehicle travelling – enough notice would need to be given by an approaching vehicle so that its trajectory can be calculated. Of course, if all vehicle journeys are planned in advance, a central computer could manage this as it would know the entire route of a vehicle.
- What lane is the vehicle travelling in – from the direction of the vehicle some assumptions can be made, but if there are two lanes travelling straight ahead, which of those lanes is the vehicle in?
- Where is the vehicle going – if turning right, then the vehicle will be crossing the path of other traffic
- What type of vehicle is it – the vehicle’s corning ability must be taken into account as HGVs can’t corner as quickly as cars, and long vehicles will need to take a slightly different trajectory around the corner because of the swept path.
- What is the weather and temperature – icy roads are more slippery which affects cornering speed
- Is there an unexpected object in the junction such as a person, a dog, etc – this can be reported by any of the vehicles or by sensors in the junction itself
- How long is the vehicle – longer vehicles take longer to get through the gap between traffic
- Are any of the vehicles having mechanical troubles – a vehicle would need to report if it was breaking down approaching or in the intersection and didn’t have the power to get through
- Is the vehicle a genuine vehicle and not some kind of hack or spoof.
Darren has owned several companies in the automotive, advertising and education industries. He has run driving theory educational websites since 2010.