TI accelerates the next generation of physical AI with NVIDIA
TI today announced it is accelerating the safe deployment of humanoid robots into the real world with NVIDIA. By combining TI’s real-time motor control, sensing, radar and power technologies with NVIDIA’s advanced robotics compute, ethernet based sensing and simulation technologies, robotics developers can validate perception, actuation and safety earlier and more accurately. TI connects NVIDIA physical AI compute to real-world applications with deterministic control, sensing, power, and safety at every joint and subsystem. This partnership will help developers move faster from virtual development to production-ready, scalable and safety-compliant systems.
As part of this collaboration, TI designed a sensor fusion solution by integrating its mmWave radar technology with NVIDIA Jetson Thor using NVIDIA Holoscan Sensor Bridge to enable low-latency, 3D perception and safety awareness for humanoid robots.
TI’s mmWave radar sensor, IWR6243, connected via ethernet to NVIDIA Jetson Thor enables scalable low-latency, 3D perception and safety awareness for physical AI applications. By fusing camera and radar data, the solution improves object detection, localisation, and tracking while reducing false positives for confident, real-time decision-making in humanoid robots.
This solution enables human-like perception that works reliably in challenging conditions – from low light and bright glare to fog and dust indoors and outdoors – and addresses a critical safety gap that has limited real-world deployment of humanoid robots. For example, while cameras may not reliably detect glass doors or reflective surfaces, radar provides consistent detection of these transparent obstacles, enabling smooth navigation in places like office buildings, hospitals and retail environments.
