Richard Kent, president of global sales at VisionTrack, said: “NARA proactively removes false positives and monitors driver behaviour without the need for human involvement. With traditional video telematics solutions, commercial fleets can be experiencing hundreds of triggered daily events, so this will enable them to deliver more efficient working whilst not compromising on road safety.”
NARA can be integrated with existing connected camera technology – whether VisionTrack or third-party hardware – and adds another layer of analysis to artificial intelligence (AI) vehicle cameras, installed with edge-based AI technology, that are often limited by the processing capacity of the device.
NARA represents a step forward for video telematics as it uses computer vision models with sensor fusion to assess footage of driving events, near misses and collisions.
This is said to ensure the review process is manageable and timely while eliminating human availability or error, so vehicle operators can make the best use of video telematics insight to better protect road users and help prevent collisions.
During the testing phase, a 1100-strong logistics fleet was found to be generating on average 2,000 priority videos a week, which would typically take someone over eight hours to review. NARA reduced the time needed to review events that require human validation to minutes per day. As a result, the company is now targeting more efficient risk management whilst supporting their road safety strategy.
Advanced object recognition uses deep learning algorithms to automatically identify different types of vehicles, cyclists and pedestrians. With high accuracy levels, it will be able to distinguish between collisions, near misses and false positives that can be generated by harsh driving, potholes or speed humps. The software will also include an occupant safety rating that uses a range of parameters to calculate the percentage probability of injury and immediately identify if a driver needs assistance.