Autonomous vehicle reactions put to the test in real-life UK motorway scenarios

Autonomous vehicles reacted to potential collisions in ‘human-like’ ways thanks to new sensors and algorithms used in a research project.

Aimed at “radically” reducing the number of multi-vehicle crashes on motorways, the Multi-Car Collision Avoidance (MuCCA) R&D project used AI and vehicle-to-vehicle communications to instruct autonomous vehicles to cooperatively make decisions to avoid potential collisions.

Vehicles took part in exercises replicating real-life UK motorway scenarios on test tracks. When the on-board technology detected an incident, the cars shared information by radio links. On-board computers calculated the best manoeuvres to avoid the obstacles and then safely steered the agreed path to avoid an accident.

The MuCCA-equipped cars also avoided each other and removed the need to brake suddenly – which might have caused vehicles behind to drive into them.

Cranfield University developed the sensor and driving systems of the cars to make them more ‘human-like’ in their reactions.

Connected and autonomous vehicles (CAVs) are expected to become increasingly common within the next 10 years but there will be many non-autonomous human-driven vehicles. In the near-future, MuCCA-equipped vehicles could reportedly handle the added complexity of anticipating the likely behaviour of human drivers in an incident.

“The university is a pioneer in connected and autonomous vehicle engineering, and we were able to develop computer algorithms that help the cars to react in a more human-like way when avoiding collisions,” said Cranfield autonomous car engineer Ross Walker. “This can allow any potential accidents to be recognised in advance, and consequently avoided before they have a chance to begin developing.”

Autonomous car research fellow Icaro Bezerra-Viana from Cranfield said: “Computer simulations enabled us to model how human drivers behave on motorways, and how the proximity of surrounding cars influences their behaviour. The movement of the cars that surround a vehicle over the next few seconds can then be predicted in order to avoid a collision.”

The project was funded by Innovate UK and the Centre for Connected and Autonomous Vehicles (CCAV), and delivered by a consortium led by Applus Idiada with Cranfield University, Westfield Sports Cars, Cosworth, SBD Automotive and the Connected Places Catapult.

Original source article: Disclaimer: This article was not originally written by a member of the Safer Highways team.

25 views0 comments