Scientists from the University of Cambridge have programmed a small fleet of tiny robotic cars to drive on a multi-lane track and noticed how the traffic flow varied when one of the cars halted.
When the cars did not drive cooperatively, cars behind the stopped car had to stop or decelerate and wait for a gap in the traffic, as would usually take place on a real road. A queue rapidly formed behind the stopped car and general traffic flow was slowed.
However, when the cars were interactive and driving cooperatively, as soon as one car halted in the inner lane, it transmitted a signal to all the other cars. Cars in the outer lane which were in direct proximity of the stopped car slowed down marginally so that cars in the inner lane were able to swiftly pass the stopped car without having to halt or decelerate considerably.
Moreover, when a human-controlled driver was positioned on the “road” with the autonomous cars and drove around the track in an aggressive manner, the other cars were able to give way to avoid the rash driver, enhancing safety.
The results, presented recently at the International Conference on Robotics and Automation (ICRA) in Montréal, will be beneficial for exploring how autonomous cars can interact with each other, and with cars regulated by human drivers, on real roads in the future.
“Autonomous cars could fix a lot of different problems associated with driving in cities, but there needs to be a way for them to work together,” said co-author Michael He, an undergraduate student at St John’s College, who designed the experiment’s algorithms.
“If different automotive manufacturers are all developing their own autonomous cars with their own software, those cars all need to communicate with each other effectively,” said co-author Nicholas Hyldmar, an undergraduate student at Downing College, who designed most of the experiment’s hardware.
The two students finished the research as part of an undergraduate research project in summer 2018, in the lab of Dr Amanda Prorok from Cambridge’s Department of Computer Science and Technology.
A number of current tests for many autonomous driverless cars are carried out digitally, or with scale models that are either too large or too costly to perform indoor experiments with convoys of cars.
Beginning with economical scale models of vehicles available in the market with realistic steering systems, the Cambridge scientists modified the cars with motion capture sensors and a Raspberry Pi, so that the cars could interact via Wi-Fi.
They then modified a lane-changing algorithm for autonomous cars to function with a fleet of cars. The original algorithm chooses when a car should move to the next lane, based on whether it is safe to do that and whether changing lanes would help the car move faster through traffic. The adapted algorithm enables cars to be packed more closely when changing lanes and adds a safety limitation to avoid crashes when speeds are low. A second algorithm enabled the cars to sense a projected car before it and make space.
They then examined the fleet in “cooperative” and “egocentric” driving modes, using both regular and aggressive driving behaviors, and monitored how the fleet reacted to a stopped car. In the regular mode, cooperative driving enhanced traffic flow by 35% over egocentric driving, while for aggressive driving, the enhancement was 45%. The scientists then tested how the fleet reacted to a single car regulated by a human via a joystick.
“Our design allows for a wide range of practical, low-cost experiments to be carried out on autonomous cars,” said Prorok. “For autonomous cars to be safely used on real roads, we need to know how they will interact with each other to improve safety and traffic flow.”
In future research, the scientists hope to use the fleet to test multi-car systems in more complex situations including roads with more lanes, intersections, and a broader range of vehicle types.