Better Informed Insurance Decisions: V2V Machine Vision


You’ve heard of UBI. Well, what about UBI with glasses? At a recent Auto Insurance Report conference, the Tel Aviv-based tech firm Nexar gave a presentation on how collisions could be eliminated if we teach cars to talk to each other and, no less importantly, give them the ability to see.

The concept is called vehicle-to-vehicle (V2V) machine vision. By creating a cellular network by which cars can communicate with one another, then equipping cars with computer vision and machine learning, we can reduce human error and eliminate collisions by 80 percent or better, Nexar said.

How exactly does it work?

A car with machine vision on a V2V network can share data on its current trajectory. It can also share the trajectories of vehicles around it and receive updates from neighboring cars. When things get dicey, it can receive collision warnings.

Both functions, vision and communication, can play a role in that regard. For example, a forward collision warning can come from the car’s machine vision: its ability to see other cars. It can also come from the car’s V2V network: its ability to talk to other cars.

What does it mean for insurance?

Well, for starters – more detail on the circumstances surrounding a collision, greater insight on driver behavior, better risk profiles, and the ability to streamline claims via automation.

Granted, insurers who use UBI already have a lot of data in hand. But these additional technologies provide a visual and relational context to what telematics is already telling us. Using all three in conjunction, we can determine speed, acceleration, road type, driver actions like braking and cornering, time of day, distance between cars, traffic density, and more.

For insurers, the difference brings to bear on two main areas.

1. Claims

The combined data from UBI, machine vision and a V2V network makes it possible to generate automated collision reports in greater detail and context: speeding up claims processing, improving underwriting decisions, and making it harder than ever to commit fraud.

2. Risk assessment

It also brings driver scoring to a whole new level. Visual context can tell us how drivers respond to traffic signals, how their behavior changes in bad weather, and how fast they drive in relation to other cars. It brings other factors into the calculation too, like average headway distance. It also adds precision to the data we already have through UBI: for example, it detects not only hard braking, but medium braking, which is a better indicator of risk.

As a result? “We can calculate a more accurate driver score in a third of the time,” Nexar said.

Preventing collisions, streamlining claims, and improving driver scoring. If you’re looking for the next big thing for insurtech, you may have just found it. Want to know more about emerging auto insurance trends? Download our latest white paper here.