LG Innotek is collaborating with NVIDIA to build sensing solutions for physical AI and robotics, the company said. LG Innotek described the arrangement as a step toward a leading position in the physical AI market.
What was shown in Berlin
At GTC Berlin, held in the German capital on October 20, the company presented a 3-megapixel camera module and its Multi-modal Vision System, both designed to implement NVIDIA's Holoscan Sensor Bridge technology for real-time sensor connectivity to a GPU. The company said the bridge is an open-source, Ethernet-connected fabric for sensors that sends the information they gather to an AI computer without passing through the CPU. LG Innotek's convergence sensing work covers cameras, LiDAR and radar, the company said.
What the two plan next
The companies intend to extend the collaboration to physical AI sensing solutions including vision sensing and robotic hands for humanoid robots, which can be integrated with NVIDIA's next-generation platform using the same bridge technology. LG Innotek said the solutions are to be unveiled at next year's CES and GTC. An official from the company said the collaboration would broaden its engagement with global physical AI companies using NVIDIA's platform and expand opportunities to supply camera and sensing solutions equipped with the technology.
The wider supplier picture
LG Innotek said it established a strategic partnership with the US LiDAR company Aeva last year, and this year began collaborations with the autonomous driving software company Applied Intuition and with Kakao Mobility. In robotics, it said it is developing robotic components with Boston Dynamics and has started mass production of camera modules for North American robotics customers. CEO Moon Hyuksoo said the NVIDIA collaboration was recognition of the company's technological competitiveness and an opportunity to expand its sensing solutions globally.
