STMicroelectronics has partnered with Leopard Imaging to introduce an all‑in‑one multimodal vision module designed to accelerate the development of humanoid and advanced robotic systems.

The new module integrates 2D imaging, 3D depth sensing, and human‑like motion perception, while offering native compatibility with NVIDIA Jetson platforms and the NVIDIA Isaac open robot development ecosystem.

Developed to meet the size, weight, and power constraints of humanoid robots, the vision module combines ST’s imaging sensors, 3D scene‑mapping technologies, and motion sensing capabilities with the NVIDIA Holoscan Sensor Bridge. This architecture enables multi‑gigabit, plug‑and‑play connectivity over Ethernet, supporting real‑time sensor data ingestion into NVIDIA Jetson systems.

Propels Physical AI With Human-Like Awareness

The solution is fully supported by the NVIDIA Isaac open robot development platform, which provides access to AI models, simulation frameworks, and software libraries. According to STMicroelectronics, this tight integration simplifies vision system design and helps robot developers accelerate development cycles, particularly when transitioning from simulation environments to real‑world deployment.

Marco Angelici, Vice‑President of Marketing and Application for Analog Power MEMS and Sensors at STMicroelectronics, noted that humanoid robotics is moving beyond research and demonstration projects into real‑world applications across manufacturing, automotive factories, logistics, warehousing, retail, and customer service. He added that the collaboration with Leopard Imaging brings ST’s sensors and actuators directly into the NVIDIA robotics ecosystem, supporting the deployment of physical AI systems with human‑like awareness.

Meanwhile, Bill Pu, CEO of Leopard Imaging, emphasized that direct access to ST sensors and actuators within the NVIDIA ecosystem has enabled the company to standardize and streamline data acquisition and logging through the Holoscan Sensor Bridge interface. This approach allows robot manufacturers to leverage NVIDIA Isaac tools to accelerate training workflows and reduce the gap between simulation and real‑world performance.

Powered by the NVIDIA Holoscan Sensor Bridge, the multimodal vision module is positioned to support next‑generation robotic vision systems requiring high‑bandwidth, synchronized sensor data and scalable AI development frameworks.

17 March 2026