Advantech's "MIB Series" Enhances Robotics and Autonomous Vision with NVIDIA® Jetson Thor™ Compatible GMSL Cameras

Advantech Co., Ltd. announced the verified integration of GMSL cameras with its MIB series powered by NVIDIA Jetson Thor, accelerating the development of autonomous robotics.
新製品NQ 82/100出典:PR Times

📋 Article Processing Timeline

  • 📰 Published: April 23, 2026 at 19:00
  • 🔍 Collected: April 23, 2026 at 10:31
  • 🤖 AI Analyzed: April 24, 2026 at 03:29 (16h 57m after Collected)
Advantech Co., Ltd (Headquarters: Taito-ku, Tokyo, hereafter Advantech) officially announced that GMSL (Gigabit Multimedia Serial Link) camera integration has been verified in the high-performance board-level MIB series equipped with NVIDIA Jetson T4000, Jetson T5000, and IGX T5000 modules.

▶ Advantech Co., Ltd – Press Release

Press Release: English

Advantech's "MIB Series" Enhances Robotics and Autonomous Vision with NVIDIA® Jetson Thor™ Compatible GMSL Cameras

On March 26, 2026, Advantech Co., Ltd (Taiwan, TWSE: 2395, hereafter Advantech) officially announced that GMSL (Gigabit Multimedia Serial Link) camera integration has been verified in its high-performance board-level MIB series, which includes the MIB-741, MIB-742, and MIB-735 equipped with NVIDIA Jetson T4000, Jetson T5000, and IGX T5000 modules. Through this verification, the stable and highly reliable operation of GMSL cameras has been confirmed across Advantech's entire product portfolio equipped with NVIDIA® Jetson Thor™, allowing robotics and autonomous vision developers to confidently build AI recognition systems ready for real-world deployment.

Verification of GMSL Camera Operation in Practical Robotics

GMSL technology is widely adopted in the automotive and robotics sectors due to its high bandwidth, long-distance capabilities, and low latency. Advantech's verification confirms that the MIB series is capable of real-time image acquisition and AI processing in autonomous robots, AMRs, robotic arms, and AI inspection platforms. In this verification, signal quality, BSP/driver support, and continuous operation were evaluated to ensure high reliability in practical operational environments. Developers can choose from 2D, stereo, and 3D cameras provided by the following partner companies:

- e-con System
- Leopard Imaging
- ORBBEC
- oToBrite
- RealSense
- SENSING
- StereoLabs

Download the list of verified compatible GMSL cameras

Integration Verification on MIB Platforms Powered by NVIDIA® Jetson Thor™

The MIB series supports a wide range of NVIDIA AI infrastructure, including Jetson T4000, Jetson T5000, IGX T5000, and IGX T7000, providing scalable edge AI performance from compact inference to safety-critical AI workloads.

Developers and system integrators can leverage NVIDIA NIM™ microservices for robotics, NVIDIA Isaac™ foundation models and runtime libraries, NVIDIA Metropolis models and development tools for vision AI, design templates, as well as NVIDIA Halos for Outside-in Safety and NVIDIA Holoscan for real-time sensor processing, which expands the perception capabilities of autonomous robots.

To enhance high-bandwidth sensor integration, the MIB series supports the NVIDIA Holoscan Sensor Bridge (HSB), enabling low-latency communication between sensors and systems via Ethernet, as well as direct data transfer to GPU memory. This architecture synchronously captures multi-sensors such as cameras, LiDAR, radar, and IMU, enabling real-time robotics perception and decision-making through precise time synchronization and efficient sensor integration.

Accelerating Time-to-Market for Autonomous Vision

By combining industrial-grade system design, verified GMSL vision, and the advanced AI performance of NVIDIA Jetson Thor, Advantech is driving the practical application of autonomous vision in edge environments.

The company provides an environment that reduces the burden of system integration and development risks, allowing developers to focus on AI model design and application development. This platform achieves highly reliable multi-camera vision for recognition, navigation, and situational awareness, as well as the low-latency image acquisition required for real-time AI inference and sensor integration. By reducing engineering efforts with verified hardware and software, it provides flexi