Challenge, approach, and impact

• Functionality overview

  Our goal was to develop an IT system for analyzing and assessing drivers' driving styles in terms of safety, flow, and economy along with a proprietary smart device. Thanks to sensors (camera, GPS receiver, accelerator), the system will be used to collect and then analyze telematics data, enabling the assessment of driving style and providing feedback to the driver.

• Backend system & Architecture

  The system analyzed complex data using multiple data sources and various sampling methods. We had to develop different algorithms and innovative approaches to make the system dependable and scalable without amassing the cost of the solution per user and the device.

• Machine Learning

  Our Deep Learning algorithms had to be designed to work on an edge device. It means that they had to be nimble enough to provide real-time video analysis but, at the same time, be precise enough to recognize dynamic situations on the road correctly.

• R&D

  There was no similar system in the market at the moment. Our team had to prepare a detailed roadmap with various success criteria and milestones to determine whether we were able to move to the next phase. There was also a lot of experimentation involved, especially with the Machine Learning part.

• IoT

  The smart camera was a major characteristic of our solution’s innovative character, so we had to build the prototype from scratch. As the global pandemic caused a supply shortage, a lot of ambiguity was involved in the process.