Neousys NVIDIA® Jetson™ Rugged Computers

Neousys' NVIDIA® Jetson™ Rugged Computer solutions are engineered to meet the ever-evolving demands of edge AI applications. From edge inspection and roadside systems to AMR, in-vehicle IVA, and ADAS, these platforms feature rugged designs, versatile onboard camera connectivity, and diverse form factors. Each product is purpose-built to deliver optimal performance and reliability across various deployment environments.

Scenario

In the past, a CPU-based architecture was used for automation processes with multiple stations along the production line, from cleaning, inspection, sorting, to packing. To enhance overall automation efficiency and precision, integrating additional AI performance into the existing CPU-based automation process is a new trend.

Challenges

The challenges of incorporating AI computing into traditional CPU-based 19" rack mount or expansion box pc include:

• Significant AI performance for real-time image processing from multiple cameras
• Inability to supply sufficient power to power-hungry GPU cards
• Issues with heat dissipation from GPU cards
• Insufficient expansion slots for installing GPU cards and frame grabber cards

Neousys' Offering

Neousys' industrial-grade intelligent frame grabber PCIe-NX150 series, and flattop heatsink edge AI computer NRU-150-FT series, are lightweight solutions designed to enhance AI capabilities to existing CPU-based machine vision and inspection systems. They are equipped with full-bandwidth 6x USB3 or 4x PoE+ GbE ports for camera connectivity, 100 TOPS AI performance, while ensuring rugged, wide-temperature operation.

Scenario

With urban infrastructures becoming more AI-enabled, such as intelligent traffic, V2X communication, public space IVA, etc., the need to deploy roadside edge AI computers for real-time image capturing and processing has become more prevalent.

Challenges

Deploying edge AI computers in roadside cabinets needs to overcome the following challenges:

• Significant AI performance for real-time image processing from multiple cameras
• Extreme weather temperatures ranging from freezing winters to scorching summers
• Heat dissipation inside cabinets
• Roadside environments causing dust to accumulate inside computers

Neousys' Offering

The NRU-220S, NRU-52S+, NRU-154PoE-FT stand out as fanless AI NVRs with built-in connectivity for multiple PoE+/ PoE++ IP cameras. Engineered with a wide operating temperature range, these devices facilitate intelligent video analytics for roadside applications. Furthermore, they include an isolated RS-484 interface for seamless infrastructure communication.

Scenario

There are more and more autonomous Mobile Robots (AMRs) applications ranging from smart agriculture, autonomous warehouse, last-mile delivery. Multiple sensors and sufficient AI performance are essential to enable the AMR to interact with its surroundings.

Challenges

One integrating a computing unit into a robot may encounter the following challenges:

• Multiple camera and sensor connectivities
• Significant AI performance for real-time image processing from multiple cameras
• Limited space for deployment
• Heat dissipation
• Powered by limited battery capacity
• Considerations for shock and vibration
• Mobile communication capability

Neousys' Offering

The NRU series delivers outstanding AI performance for up to six GMSL cameras, all while maintaining an ultra-low power consumption of just 15 to 30 watts, enabling extended operation for AMRs, even on battery power. It supports both 2D and 3D GMSL2 automotive cameras, providing clear vision in dynamic lighting conditions and precise distance measurement for obstacle avoidance. With a wide voltage input range of 8 to 35V and 4G LTE/5G NR connectivity for seamless communication, this rugged, compact AI computer is designed to tackle outdoor challenges, making it the perfect choice for AMR applications.

Scenario

Intelligent video analytics in vehicles play a crucial role in implementing collision avoidance, recognition, and detection, thereby enhancing in-vehicle ADAS capabilities and overall safety. These applications extend to areas such as off-road vehicle ADAS, teleoperation, and railway intrusion detection.

Challenges

To ensure reliable implementation on vehicles, the computing systems need to overcome:

• Multiple camera and sensor connectivities
• Significant AI performance for real-time image processing from multiple cameras
• Limited space for deployment
• Heat dissipation
• Powered by limited battery capacity
• Constant shock and vibration
• Mobile communication capability
• Demanding environments like mining, agriculture, or construction sites

Neousys' Offering

The NRU series features IP66-rated waterproof, rugged designs that simplify in-vehicle deployments, supporting an 8V to 35V DC input, ignition power control, GMSL2, IP cameras, sensors, LiDARs, and more. The NRU-230V-AWP takes it further with an onboard automotive MCU and an independent M12 port for CAN bus and digital output, enabling real-time monitoring and predictive maintenance, reducing downtime and enhancing functional safety for autonomous vehicles.