NVIDIA has declared what it claims is the first commercially accessible Level 2+ self-driving system in Las Vegas at the current CES 2019. Auto manufacturers can employ the tech, named as Drive AutoPilot, to include self-driving functions to their cars. In spite of its name, it is not actually a complete autonomous driving platform. Rather, it can offer cars with driver assistance features, such as lane change, highway merge, parking assist, lane split, and pedestrian detection.
The tech also includes personal mapping to cars, permitting them to recall where they have driven and makes an autonomous driving route even without HD map. Part from this, it can also offer cars the capability of monitoring drivers, enhanced in-cabin visualization of their computer vision system, and AI copilot abilities.
The Drive AutoPilot is fraction of Drive platform by NVIDIA for cars and is fueled by the firm’s Xavier system-on-a-processor. Xavier was particularly generated for AI car platforms, and as per NVIDIA, it is able of 30 Trillion processes each second employing just 30 Watts of energy.
On a related note, NVIDIA’s aim to fuel autonomous bots earlier started in earnest. The firm rolled out a Jetson AGX Xavier Module that offers intelligent machines and robots the processing power they require for their brains. You are not about to purchase one yourself since it is available in batches of 1,000 units for a price tag of $1,099 each. On the other hand, it can be essential for automatons and delivery bots that require a lot of customized performance with comparatively low use of power.
The Jetson Xavier chipset at the core of the module depends on no less than 6 chips to get its process done. There is a comparatively traditional 8-core ARM processor, but you will also discover two NVDLA deep learning processors, a Volta-supported GPU, and devoted video, image, and vision elements.