Accelerating Robotic Vision with NVIDIA Jetson-Compatible Multi-Sensor Module from STMicroelectronics and Leopard Imaging

[Ahmet Ö.]

## Accelerating Robotic Vision with NVIDIA Jetson-Compatible Multi-Sensor Module from STMicroelectronics and Leopard Imaging

STMicroelectronics and Leopard Imaging have introduced a multimodal vision module that provides 2D imaging, 3D depth sensing, and human-like motion perception for humanoid and advanced robotic systems. This module offers full integration with the NVIDIA Jetson and NVIDIA Isaac development platform via NVIDIA Holoscan Sensor Bridge technology, simplifying and accelerating the design of robotic vision systems.

### About the Project

The module integrates ST's imaging, 3D scene mapping, and motion sensing capabilities with NVIDIA's robotics ecosystem. This supports the implementation of advanced vision applications within the size, weight, and power constraints for human-like robots. Marco Angelici, Vice President of Marketing and Applications at STMicroelectronics, stated that the technology will increase the use of humanoid robots in many sectors, from manufacturing to logistics and retail.

Bill Pu, CEO of Leopard Imaging, emphasized that this module provides direct access to ST sensors, standardizing data collection and analysis processes for humanoid robots and accelerating the development process.

### Integration and Software Features

Powered by the NVIDIA Holoscan Sensor Bridge, this module works seamlessly with NVIDIA Jetson over Ethernet for real-time sensor data acquisition. Additionally, open AI models, simulation tools, and developer libraries are integrated into the NVIDIA Isaac platform. The module offers a simulation environment that includes AI algorithms, sample applications, and sensor models prepared for mobile robots.

### Technical Specifications

• Image Sensor: ST VB1940, 5.1 megapixels, RGB-IR, automotive-grade, combination of rolling shutter and global shutter modes
• Alternative Sensor: V943 from ST's BrightSense family, monochrome or RGB-IR versions, both die and packaged sensor options
• Motion Sensing: LSM6DSV16X, 6-axis inertial measurement unit (IMU), machine learning core for local AI, low-power sensor fusion, and electrostatic sensor
• 3D Depth Sensing: VL53L9CX dToF LiDAR module, accurate ranging up to 9 meters, 54x42 zone resolution, 55°x42° field of view, 1° angular resolution, performance up to 100 frames per second

### About ST and Leopard Imaging

STMicroelectronics is a global leader in semiconductor technologies, focusing on environmental sustainability goals. Leopard Imaging, headquartered in Silicon Valley, offers innovative AI-powered imaging solutions and serves sectors such as automation, robotics, and healthcare.