Automotive AppsConnected VehicleInfotainment

Zendrive launches Accident Detection SDK

Published: April 09, 2015| San Fransisco, CA

Zendrive announced that its Accident Detection SDK can now be used by developers for apps to alert emergency responders, tow truck drivers and families to car accidents. The records kept by Zendrive may also mitigate. 

Zendrive warns that 90% of accidents result from human error and 30% result from phone use. They claim that hardware solutions are limited to car diagnostics while Zendrive uses smartphone sensors and software to understand driver behavior.

Zendrive’s analytics uses a variety of signals, cell phone use, speed, swerves, hard stops, fast accelerations, fatigue, as well as weather, trip duration and time of day. It is the hopes of the company that Zendrive Accident Detection improve road safety.


The technology was developed by monitoring phones in crash-tests and on the road. The company tested and refined models in accidents and events such as hard brakes, phone drops that were not accidents. The software has a 100% accuracy rate with no false positives or false negatives.

BMW gave Zendrive access to its accident testing facilities. Zendrive Accident Detection will be available first for iOS iphone apps while an Android version will be available soon.

Beyond allowing developers to build emergency accident-response features into their apps, the technology can be incorporated into many other usecases. A roadside assistance app can build notifications for emergency assistance or towing. A ridesharing app can build notifications for emergency contacts in case of an accident. A driving education app can give both students and parents peace of mind by building in multiple types of alerts. A valet parking app can use the detection on the back end, to prove that their driver was driving cautiously and minimize customer and insurance liability in the event of a collision. Delivery fleets can send notifications to recipients so they know deliveries will be delayed.

The technology was refined by repeatedly putting phones into crash-test cars and normal cars on roads, and testing and refining the models on both “accidents” (crash-test collisions) and non-accident events (such as hard brakes, phone drops, etc), until the software had an extremely high accuracy rate. There were no instances of false positives or false negatives for rides over 30 mph, where the phone was either in a mount, in a pocket, or in a cup holder. 

Source: Zendrive


Related Articles