Autonomous driving development is focused on shifting human control to computers, in hopes of doing it more effectively and consistently. Auto supplier ZF is looking to build a platform that brings new intelligence to those systems, starting with a prototype that literally sees pedestrians before they’re visible to a human or camera eye.
The technology, called X2Safe, detects pedestrians that aren’t visible to drivers, cameras, or radar, and then communicates with vehicles, smart phones and smart watches.
Reimagining the Platform
“Pedestrians and cyclists account for nearly half of the 1.25 million deaths caused by traffic accidents worldwide every year,” explained Malgorzata Wiklinska, head of ZF’s think tank.
“The beauty of our platform is that it can empower pedestrians and drivers alike to intervene ahead of an imminent collision before they visually ‘see’ it.”
Key to the technology is ZF’s proprietary algorithm, which Wiklinska says took six months to develop and test. The algorithm individually analyzes the behavior of all road users, drawing on real-time readings of speed and direction, and then correlates it with probability models to assign level of risk.
Users can decide what notifications they want to receive, and the behavioral models learn and thereby get smarter over time. The algorithm can process situational dangers, such as roads with poor visibility, construction, or heavy use environments, such as bus stops and school crossings.
X2Safe is an example of the fundamental imagining of ZF’s “platform” of traditional mechanical components, and extending it to include the digital space around and beyond vehicles.
“The algorithm can be used car-to-X, X-to-car, and even in X-to-X applications,” said Wiklinska.
“The more people and vehicles linked within the interactive safety network, the more effective the accident avoidance.”
Looking for Innovation Partners
ZF is also actively looking for partners to innovate practical applications for X2Safe, choosing to pursue its development as it would a startup. Proof of concept tests could involve strategic alliances with technology partners, such as Google, as well as user communities, such as city transportation authorities, or school districts.
The company is also considering potential licensing of the algorithm for wearables, such as wristbands or tags that students could use. X2Safe is ready to play a role in autonomous driving applications as soon as model year 2020 vehicles, as it could be programmed to trigger a variety of responses, from warnings to emergency stops.