With more than 8 million members, the RAC is one of the UK's most progressive motoring organisations, providing services for both private and business motorists. Whether it's roadside assistance, insurance, buying a used car, vehicle inspections and checks, legal services, or up-to-the-minute traffic and travel information—the RAC offers a solution for all motoring needs. The RAC is committed to making motoring easier, safer, more affordable, and more enjoyable for drivers and road users.

When the RAC decided to launch a “black box” based insurance telematics proposition, one of the key objectives was to reliably detect low-speed collisions and discriminate these events from more common driving events such as going over speed bumps or potholes.

In addition to this, we set ourselves the challenge of being able to convert this data into information that a non-technical user could understand by giving a plain English classification of the event that had been captured rather than numerous graphs.

This session gives an overview of how RAC undertook this process. It shows how MATLAB^® is a key part of the tool chain: from initial development and signal capture through feature extraction and machine learning, independent third-party verification, and ultimately deployment into a scalable cloud-based environment.