Motorway pile-ups are costly – both in financial terms and in terms of human lives.

An Applus+ IDIADA-led consortium was awarded government funding towards the development and trial of technology that could radically reduce the number of multi-car collisions on motorways.

 

Multi-Car Collision Avoidance (MuCCA)

 

The Multi-Car Collision Avoidance (MuCCA) project will use Artificial Intelligence (AI) and vehicle-to-vehicle communications to help cars and eventually autonomous vehicles make cooperative decisions to avoid a potential accident.

The idea of this system is that a group of vehicles can use advanced sensing, processing and automated driving controls, together with vehicle to vehicle (V2V) messaging, to decide on a joint shared plan to avoid a collision. For example, one vehicle may steer out of the way to allow a second vehicle extra room to avoid an obstacle, allowing crash avoidance that no single human driver could achieve.

The project, led by automotive design and testing experts from the IDIADA Division, includes input from Cranfield University, Westfield Sports Cars, Cosworth, Secured by Design and the Transport Systems Catapult, will also develop data logging capabilities to create a record of the exact causes of accidents. A computer-simulated environment will also be created, in which the vehicles’ AI systems can practise complex crash scenarios before being trialled on real-world test tracks.

The division is using the IDAPT (IDIADA ADAS platform tool) control unit which combines high-performance GPU-based processing, V2V communications, video and satellite positioning with automotive safety, I/O and network protocols to expressly target CAV applications such as this.

The MuCCA project will be one of the first users of this tool. Considerable progress has already been made. A tailored software platform has already made V2V exchanges through the air and processed GPS signals, while the hardware design is being finalized for first prototypes to be built this year.

Key Innovations:

  • Cooperative decision and trajectory control for complex collision avoidance
  • Prediction of human-controlled vehicle paths
  • Shared sensor-agnostic world view
    • Match data from other connected vehicles to sensed vehicles
    • Merge sensor data from other vehicles
  • Develop insurance logging capability to support event reconstruction
  • Integrated simulation environment to evaluate complex crash scenarios
  • Cyber-security assessment for common requirements
  • Multi-vehicle system validation on a test track