Source: dRISK announcement

London, U.K. — dRISK, a London and Pasadena-based startup, announced it has employed its edge case retraining tool to detect high-risk events for autonomous vehicles (AVs) as means to improve their operational safety.

Whereas semi-autonomous and autonomous vehicles (SAV or AV) currently do not always detect high-risk events in time to react to them (oncoming cars peeking into the lane from behind other vehicles, vehicles running red lights concealed by other cars), dRISK’s tools for retraining AVs to recognize edge cases represent a dramatic step forward in the ability to retrain autonomous vehicles to well outperform humans at even the trickiest driving scenarios.

Related post:
Ansys, Velodyne Join to Transform AV Safety

dRISK’s mission is to help make AVs dramatically safer as soon as possible. The new patented knowledge graph technology (analogous to Google’s knowledge graph of the internet, but in dRISK’s case a knowledge graph of real-world events) solves several problems which have plagued AV developers so far — encoding massively high-dimensional data from all the different relevant data sources into a tractable form, and then offering the full spectrum of edge cases to retrain on not just with what has already happened but will happen in the future.

dRISK delivers simulated and real-plus-simulated edge cases in semi-randomized, impossible-to-game training and test sequences, to achieve superior testing and retraining results for customers on real-life data.

Unlike traditional training and development techniques, in which AVs are trained to recognize primarily whole entire vehicles and pedestrians under advantageous lighting conditions, with dRISK’s edge cases AVs are trained to recognize just the predictors of high-risk events (such as the headlights of an oncoming car peeking into the lane amid low visibility).

AV systems trained this way can recognize high-risk events sooner, without a significant decrease in performance on low-risk events.

About dRISK

Using dRISK for retraining, Autonomous Vehicles can detect and contend with edge cases six times sooner and with two times greater accuracy. dRISK has built a taxonomy of edge cases derived from massive and heterogeneous data focused on high risk — millions of hours of CCTV footage trained on high-risk intersections, full-text accident reporting, and extensive expert input from both transportation specialists and NASA experts in failure mode analysis. Integration is easy, and data can be delivered exclusively for perception retraining on fully annotated simulated and real-life data, or for full-stack AV risk assessment with hardware in the loop. dRISK’s customers include AV developers and the world’s largest insurers and transport authorities. The entire announcement can be viewed by clicking HERE.