LG Innotek has unveiled sensing products developed for robots at Nvidia’s GTC Berlin technology event in Germany. The disclosure brings together the Korean component maker’s camera and autonomous-driving sensor capabilities with Nvidia’s push to give physical AI systems a more detailed view of their surroundings.

The products include a 3-megapixel camera module and a multimodal vision system that combines cameras, lidar and radar. The combination is designed to detect not only an object’s shape but also its distance and movement, allowing a robot to build a three-dimensional understanding of its environment.

The system applies Nvidia’s Holoscan Sensor Bridge technology to send sensor data directly to an AI computer without passing through a CPU. LG Innotek says this architecture can reduce processing delays and help robots respond faster as conditions change. The evidence does not establish a contract value, production timetable or completed joint-development result.

The companies plan to expand their cooperation from sensing systems that provide humanoids with vision to robotic hands that grasp and manipulate objects. LG Innotek has separately worked with Boston Dynamics on robot vision sensing and formed a strategic partnership with lidar company Aeva. It is positioning sensing, communication modules and semiconductor substrates as parts of a broader physical AI component business.