Terranet announces industry defining latency speeds at STARTUP AUTOBAHN

As its VoxelFlow™ sensor tech showcases ability to react to objects in less than 20 milliseconds, the Company will now focus on developing its unique AI technology for classification of objects

Terranet AB (Terranet/the Company),, developers of Advanced Driver Assistance Systems, announced today at STARTUP AUTOBAHN that its 3D motion awareness technology, VoxelFlow™, has achieved latency numbers unprecedented within the industry. In a joint presentation with Mercedes-Benz, Terranet successfully showcased VoxelFlow’s™ potential integration with Mercedes-Benz’s real-time map streaming technology LiveMap in addition to demonstrating VoxelFlow’s™ unmatched reaction time in a Mercedes-Benz’s Maybach S-Class Configurator.

After a successful completion of tests earlier this month, Terranet officially announced that VoxelFlow™ proved the ability to react to objects in less than 20 milliseconds, shaving precious time off the industry standard 300 millisecond expected from today’s ADAS tech companies. These impressive latency numbers are only the latest milestone from Terranet on its way to achieving a reaction time of up to 100 times faster than today’s systems.

Initially showcased at last February’s STARTUP AUTOBAHN, VoxelFlow™ consists of three event cameras and a continuous laser that enables AD/ADAS vehicles to quickly and more reliably perceive their surroundings in a continuous 3D voxel stream. The solid-state system is unique in its ability to detect, track and trace objects as they move through space in addition to specifying velocity, speed, direction and position, effectively locking onto obstacles as they enter the roadway. For LiveMap, the data generated is used to create a heat map of hazardous spots and safety threats in a real time safety navigation layer.

VoxelFlow™ specifically addresses the urban areas, where life-threatening accidents are most likely to occur, drastically decreasing a vehicle’s braking distance and significantly lowering the potential for roadway accidents. The ground-breaking latency numbers signify a new wave of advancements in autonomous driving and roadway safety by improving upon both autonomous vehicle and human reaction time.

Where current camera and laser technology is subject to delays in latency or deficiencies in image quality, VoxelFlow™ creates high-resolution, in real time events, moving beyond pixels to reimagine the world in a stream of voxels. Building upon VoxelFlow’s™ already impressive low latency, Terranet looks forward to sharing updates on the product’s accuracy later this year. Going beyond typical obstacle detection, VoxelFlow™ will leverage Terranet’s unique AI software to classify objects as they move through space, differentiating between stationary and mobile objects and even anticipating their trajectories.

“Our technology has the ability to identify and classify moving objects much faster than existing systems, it sets a new potential safety standard for what is possible when it comes to ADAS and autonomous driving technology, but is only the beginning for Terranet. Now we aim to utilize those revolutionary latency numbers in combination with our unique AI software stack for classification of objects” said Pär-Olof Johannesson, CEO of Terranet. “Ultimately, VoxelFlow™ is developed with roadway safety as a first priority. With the recent tests from the test-track, we have proven our technology is outperforming current ADAS systems in speed, the future of zero roadway accidents is even closer.”

This announcement follows a string of prominent news for Terranet, which includes a continued collaboration with Mercedes-Benz, joining NIVIDIA Inception, the hiring of former Mercedes-Benz Innovator Nihat Küçük as CTO and SVP, and a first-of-its-kind technical and business partnership with Audi-backed holoride.

SOURCE: Terranet 

Hi,

It looks like you're browsing from an enterprise account

Please confirm your login by clicking the button below. If you are not an enterprise customer, please click close.