Follow
Subscribe

New Algorithm Can Make Driverless Cars Merge Like a Fish

Home > News > Content

【Summary】Researchers in Switzerland are recently putting forward a new algorithm, to allow autonomous cars to merge with human traffic like a school of fish.

Original Claire    Feb 05, 2017 11:00 AM PT
New Algorithm Can Make Driverless Cars Merge Like a Fish

The traffic condition can often times be complex and unpredictable. With more and more autonomous cars being developed in the next 5 to 10 years, there will be a transition period when human-driven cars share the public road with driverless vehicles. How can cars that are controlled by machines and humans peacefully drive on the road without conflicting with each other? Scientists are striving to make that happen. 

Researchers from Distributed Intelligent Systems and Algorithms Laboratory at the Ecole Polytechnique Federale de Lausanne (EPFL) in Switzerland are recently putting forward a new algorithm, to allow autonomous cars to merge with human traffic like a school of fish. Traditional way of organizing autonomous cars is often single lanes of driverless vehicles lining up in equal spaces and following a leader car. This algorithm allows autonomous cars to form multi-lanes of a platoon: once a human-driven car merges into highway, members in the platoon could adjust their position accordingly and make space for the car to enter the lane, just as flexible as a fish.

"These cars have different levels of intelligence on board, from fully-autonomous, to standard legacy vehicles. The goal of our project is really to essentially increase the coordination on the road, and eventually achieve reduced fuel consumption and additional safety." Alcherio Martinoli, head of the lab told EPFL News. 

He later told phys.org that they have been working on the algorithm for 10 years. Based on Vehicle to Vehicle (V2V) communication technology, each car is equipped with with LiDAR, radar and camera installations, enabling a far-reaching, 360-degree view of the road ahead and behind. More importantly, each vehicle is using wifi to share information and what they can "see" between each other, so that flexible positioning change is possible when a vehicle cuts in and interrupts the platoon. The algorithm, according to researchers, can be used to create platoons of any size, comprising of autonomous vehicles following the lead of a single (or multiple) human-driven cars.

The team had carried out a successful test-run last October, with two autonomous vehicles (a truck and a compact car) and one human-driven car running on a highway. The manual vehicle can communicate with the other two using V2V methods. Although currently connected cars are just emerging, the Swiss lab designed the algorithm based on a 2030 vision, when widespread use of V2V communication could enable the harmonious coexistence of driverless cars and human-driven vehicles that share one road. 

"For the first time we were able to validate what we had achieved in the simulation. And the number of vehicles in the convoy has no impact on the complexity of the control mechanism."Martinoli said. 

The project is under Autonet 2030, an EU funding program with the goal of incorporating autonomous vehicles into popular use. Before all cars on the road become fully-autonomous, a transitional way of coping with cars of different levels of intelligence is surely a core issue to be solved. 

Prev                  Next
Writer's other posts
Comments:
    Related Content