Situational awareness for video telematics systems
Recognition and understanding of traffic situations through distributed, self-learning camera sensors in motor vehicles
According to a study by the World Health Organization, around 1.25 million people die in traffic accidents worldwide every year [1]. Pedestrians, cyclists and motorcyclists in particular are exposed to high risks in traffic. It can further be assumed that the number of vehicles and road users will continue to increase, especially in cities. Last but not least, the COVID-19 pandemic has further accelerated the growth trend in the e-commerce sector and thus the growth in daily shipments of goods. This increases the need for commercial-use vehicles to deliver shipments with ever-shorter delivery times. When fleet vehicles registered today have advanced telematics systems, various basic data are usually collected on location, speed, and driving and resting times. However, additional information necessary to improve driving safety, such as driver assistance systems (ADAS), is not. Peregrine has developed a solution that brings AI-based technologies from the field of autonomous driving behind the windshield of every vehicle at a fair price.
In 2021, Peregrine applied to IBB’s ProFIT program to receive official funding and support from the European Regional Development Fund (ERDF). Since the end of 2021, the ERDF has been co-funding this project after IBB positively evaluated Peregrine’s application to research and develop distributed, self-learning camera sensors for context assessment in the mobility sector.
With this project, Peregrine aims to innovate in the field of video telematics by researching and testing new approaches to robotics and artificial intelligence. Through this innovation, Peregrine can help fleet operators use vehicle data in a GDPR-compliant way to reduce the risk of accidents for drivers and optimize fleet operations. Specifically, in this project, funded by the European Regional Development Fund, video information in (fleet) vehicles is analyzed using machine learning models, in an approach that is both GDPR-compliant, resource-efficient and hardware-agnostic. In addition to increasing traffic safety, the objectives are to optimize and document routes while strictly observing data protection.