ADLINK Announces AGX Xavier Industrial Module

ADLINK's AVA family of AI image analysis platforms (AVA) has added the AVA-RAGX, a compact, rugged AI solution for both in-vehicle and trackside applications, powered by NVIDIA's newest AI-enabled single board computer module, the Jetson AGX Xavier. solution for both in-vehicle and rail-side applications, with tough reliability and superior performance in harsh environments. ADLINK's AVA-RAGX and the AVA platforms in the same series offer a high degree of flexibility for railway application solutions, with optimised configurations based on application requirements. Applicable applications include foreign object intrusion detection, platform video surveillance, in-vehicle passenger safety and railway hazard detection.

AI is accelerating the digital transformation of the railway transportation industry, effectively improving operational efficiency, providing smarter and safer travelling services, and creating new business opportunities,’ said Fuyao Gao, General Manager of ADLINK's Network Communications and Public Construction Business Unit. ADLINK provides advanced AI computing solutions that allow railway practitioners to focus on the development of differentiated terminal application products and the corresponding transformation. The launch of ADLINK's AVA-RAGX ruggedised AI image analysis platform demonstrates our focus on and support for the railway transportation industry and proves that we, as an NVIDIA Elite Partner, are able to effectively accelerate product development, gain faster access to the new generation of NVIDIA GPUs, and receive advanced technical support.’

ADLINK is one of the few companies in the world that is also an NVIDIA Quadro Embedded Partner, an OEM Preferred Partner, and a Jetson Elite Partner. These accolades show that ADLINK has been recognised by NVIDIA as a partner in creating rugged embedded onboard GPU solutions. In addition, ADLINK's deep capabilities and partnerships provide access to NVIDIA's advanced technology and support for innovative, open standards-based heterogeneous computing solutions for next-generation AI and machine learning applications.

ADLINK AVA-RAGX is a rugged, fanless AI platform designed for flexible deployment in harsh operating environments. Certified to EN 50155 standards, it combines AI, machine learning, edge computing and IoT technologies, making it an ideal platform for real-time imaging and visual analytics applications. Equipped with 8-core NVIDIA Carmel CPUs and 512-core NVIDIA Volta GPUs, it is essential for today's increasingly complex railway traffic operations and demonstrates ADLINK's expertise in developing next-generation railway systems.

With compact dimensions of 288 x 190 x 72 mm, the ADLINK AVA-RAGX is designed specifically for space-constrained on-board environments, making it an ideal choice for small form factor chassis solutions for rail edge applications. It offers a full complement of I/O ports, including four M12 ruggedised Gigabit Ethernet ports and wireless network interfaces (5G/LTE and Wi-Fi), as well as four digital inputs and four outputs with 1.5kV isolation for voltages between 24VDC and 110VDC.

When integrated with advanced camera systems, radar, optical radar, GNSSC and AI software, the AVA-RAGX platform provides accurate information about a train's surroundings and location. Unique to the railway industry is the need for trains to have long braking distances, and by connecting the AVA-RAGX to the appropriate sensors, it can detect hazardous conditions that will occur in the distance, increasing warning time. In addition, hazard detection systems can improve operational efficiency and safety, and help prevent accidents, especially when trains are operating at night or in poor weather conditions with reduced visibility.

When the AVA-RAGX platform is installed in the detection train, with the track equipment fault detection system processing in real time the images obtained from the pantograph, down to the tracks and other wayside equipment, and then through the accelerated visual computing of NVIDIA GPUs, the whole system can effectively identify potential equipment faults at train speeds of up to 120 kilometres per hour.