Nvidia Self-driving car system
Nvidia self-driving car system unveiled at Los Angeles Auto Show, Nov. 19, 2016. Nvidia

When it comes to self-driving cars, safety remains a prime concern, especially after recent crashes involving such vehicles. Currently, such cars use cameras and sensors for object detection with only Tesla using radars in limited capacity for the same purpose.

Read: Uber Self-Driving Cars Pulled Off Roads After Crash In Arizona

An Israeli firm, Arbe Robotics, aims to make self-driving safer with its radar-based 4D mapping system called Ultres, which will provide self-driving vehicles with high-resolution imaging with 360-degree obstacle detection up to 300 meters. The company claims that the car would be able to act faster with relation to obstacles than humans in all weather conditions.

Ultres is the combination of the company’s patented hardware, signal processing and 3D modeling algorithm. According to the company, the system enables self-driving cars to have fully adaptive cruise control, accurate emergency braking and real-time blind-spot monitoring, which could make autonomous driving more convenient and safer.

Cruise control is a technology that puts automatic controls on the speed of the vehicle. In the context of self-driving cars, the steady speed is set and adapted according to driving conditions and feedback received by the car’s systems. Ultres is expected to make the system more accurate by providing real-time feedback, so that self-driving works better. Cars with the system will be able to detect the obstacle’s distance, velocity and direction in order to avoid a collision.

Read: Forget Self-Driving Cars, Autonomous Ships Are On The Way

The company claims that the system will provide high-resolution images to the car, which will be refreshed at the rate of 50 times an hour. It will then create a 3D image of the objects, track them, localize the car and share the information with the car's inbuilt system.

According to the Next Web, the company has already raised $2.5 million in private investment and might run tests in May. Its goal is to sell the system as a hardware and software package to car manufacturers as well as auto part sellers such as Denso and Bosch.