Hitachi Rail Enhances Real-Time Railway Analysis with High-Speed AI Utilizing NVIDIA Technologies.
Nvidia IGX Platform
The international transportation firm is using NVIDIA IGX, the industrial edge computing platform that is enterprise-ready, to enhance train operations and save maintenance and energy expenses.
With the goal of lowering maintenance costs for rail operators, cutting down on train idle time, and enhancing passenger transit dependability, Hitachi Rail, a multinational transportation firm that powers railway networks in over 50 countries, is integrating NVIDIA AI technology.
HMAX
In order to handle sensor and camera data in real time, the business is integrating NVIDIA IGX, an industrial-grade, enterprise-level platform, into its new HMAX platform. This platform offers high-bandwidth sensor processing, sophisticated AI computing, functional safety features, and enterprise security.
The HMAX platform will help Hitachi Rail customers identify train tracks in need of repair, follow the deterioration of overhead power lines, and evaluate the condition of trains and signaling apparatus more swiftly by eliminating the lag time between data collection and processing.
According to Hitachi Rail, preventive maintenance is around seven times less expensive than emergency repairs made when infrastructure breaks down without warning. By lowering fuel expenses at train depots by up to 40%, its current AI monitoring systems are already reducing energy usage by up to 15%, service delays by up to 20%, and train maintenance costs by up to 15%.
The business hopes to further boost these savings with real-time analysis utilizing NVIDIA IGX and NVIDIA Holoscan platform for sensor processing.
AI Engine Runs in Real Time with NVIDIA IGX
Hitachi Rail has utilized NVIDIA IGX and the NVIDIA AI Enterprise software platform to develop new accelerated AI applications to assist operators in monitoring train fleets and infrastructure. These applications build upon the company’s current collection of HMAX applications, which are currently operating on data from 8,000 train cars on 2,000 trains. To make the creation and implementation of production-grade AI applications more efficient, NVIDIA AI Enterprise provides tools, pretrained models, and application frameworks.
These apps, which will be made accessible shortly via the HMAX platform, will enable the company’s global clientele to handle massive amounts of data that are sent from train sensors, imported from third-party software that the client currently uses, or pulled from current systems.
For instance, every Hitachi Rail train in the United Kingdom contains sensors that provide information on about 50,000 data points as often as every fifth of a second. When a train or rail line component requires repair, AI technology that keeps up with this data flow can notify train operators in a timely manner. A chatbot interface is another way that operators may get the AI insights, which make it simple for them to see patterns and chances to improve maintenance schedules among other things.
Creating Long-Term Benefits
The potential extend beyond rail and train surveillance
Hitachi Rail can monitor power lines above to detect deteriorating electric wires and assist in preventing disruptive breakdowns by installing cameras atop trains. Processing one day’s worth of video footage recorded by the train often takes up to ten days. Data may be analyzed in real time at the edge using NVIDIA-accelerated sensor processing, returning only pertinent information to operational control centers for analysis and decision-making.
Nvidia IGX Orin
NVIDIA IGX Orin AI platform for industrial edge that is ready for enterprise use.
The NVIDIA IGX platform combines enterprise-grade hardware, software, and support to provide edge AI capabilities suitable for industrial use. With its robust AI computing, high-bandwidth sensor processing, enterprise security, and functional safety, it is specifically designed for use in industrial and medical settings. Together with NVIDIA AI Enterprise and a maximum of ten years of maintenance, the platform allows you to safely and securely provide AI to facilitate cooperation between humans and machines.
Better Performance in Industrial-Grade Systems with NVIDIA IGX Orin
With two production versions, you can build NVIDIA-certified complete systems while getting blazing-fast performance in the size and power envelope required for the edge. The world’s most powerful AI computer for energy-efficient autonomous machines, the IGX Orin 700, comes in a MicroATX board kit with an NVIDIA ConnectX-7 smart network interface card for accurate timing and fast data streaming. A System-on-Module (SOM) for specialized carrier board solutions with optional functional safety measures is the IGX Orin 500.
AI Enterprise IGX: Software of the Enterprise Class for AI Mission-Critical
This enterprise-grade edge AI software solution is designed specifically for NVIDIA IGX Orin-powered mission-critical applications. IGX OS, AI frameworks, SDKs, libraries, and long-term corporate support are all part of the software stack that it offers unparalleled performance, security, stability, and enterprise support. Best-in-class security, safety, and support are therefore guaranteed. The software stack incorporates a board management controller that comes with IGX for remote software upgrades and system administration.
Nvidia IGX Orin Developer Kit
The NVIDIA IGX Orin Developer Kit accelerates edge AI application development with a chassis and power supply for high-performance, industrial-grade edge AI.
IGX Orin Developer Kit | |
---|---|
AI Performance | Up to 1705 TOPs with optional RTX 6000 Ada GPU (Sparse) |
SOM (System on Module) | GPU: 2,048-core NVIDIA Ampere architecture with 64 Tensor Cores CPU: 12-core Arm Cortex-A78AE v8.2 |
NVIDIA ConnectX-7 | NVIDIA ConnectX-7 2x 100GbE 32-lane Gen 5 PCIe switch (x8 upstream, x16 Downstream, x8 Downstream) |
Safety MCU (sMCU) | Infineon Aurix TC397 |
NVIDIA BMC (Baseboard Management Controller) Module | Aspeed AST2600 Microchip ERoT |
GPU Max Frequency | 1.185GHz |
CPU Max Frequency | 1.971GHz |