A new measure for safer vehicles, self-driving or not
In this Cars of the Future exclusive, the co-founders of Pasadena-based Streetscope, Mark Goodstein and David Muyres, explain how their Collision Hazard Measure enables self-driving stakeholders to accelerate deployment with confidence.
DM: “The UK is a leader in automated vehicle (AV) development and we’re in discussion with an array of companies and organisations about demonstrating our capability in the UK environment. That would be a great success point for communicating our value to customers worldwide.
“Our technology evaluates the safe moving of any vehicle, human or machine driven. We treat it as a black box and evaluate how safely it moves amongst hazardous objects in the street scene, using simulation or camera data. Then we create indexes that different industries can use, for example, insurers, vehicle manufacturers, regulators and planners.
“The insurance industry is the one we’re having most conversations with. They want to understand the hazards of new technology vehicles, and now they’re realising we can help with the human-driven side as well. They can use our data to price risk more effectively.
Informing safe self-driving deployment
“Vehicle developers need to answer basic questions like: Am I to safe to deploy yet? They currently don’t have an independent way to objectively measure how safely a vehicle moves, and we can provide that.
“Regulatory is very interesting and it’s nice they’re interested in using our measure to guide future development.
“Infrastructure planning companies can hire us to evaluate a future self-driving route. We can identify high hazard locations and make recommendations to mitigate issues.”
And you’re talking to vehicle verification bodies too?
MG: “Yes. We’re a start-up, not at scale yet, but all we need is kinematic data, the position of all objects each tick of the clock, from any traffic scene to calculate the hazard posed between the vehicle and all other objects. And we can get that from either simulations or using cameras as data sources. Then we score them based on the Collision Hazard Measure we’ve invented.
“We could use lidar and radar, but those geometric sensors are very expensive. Cameras are ubiquitous, so we’re using them and making a pretty good job of it.
“There’s a school of thought that you can use aggregated data from other drive events using cell phone based sensors, but they lack context. Why did they slam on the brakes? Why did they accelerate so aggressively? There is no correlation to risk.
“We’re trying to get these industries to recognise that the data they’re spending a lot of money on is insufficient, and they’d be better off using our Collision Hazard Measure.”
For further info, visit the Streetscope website.
