Highly Reliable Fanless Computers with MXM GPU
Artificial Intelligence : A new force to integrate within train rolling stock
Rail has gradually stepped into the application areas of AI edge computing. A train is in a state of vibration for a long time and involves ensuring passenger safety at all times. All on board electronic equipment must adapt to differing environments including shock, vibration, temperature, altitude, humidity, electrical properties and fire protection to meet stringent rail safety requirements. As a result, DFI have brought about their highly reliability embedded computer, the RC300-CS, specifically designed for rail applications where both reliability and high performance for AIoT edge computing is pre-requisite.
Driving assistance or autonomous driving is either a growing trend, both of which require high-performance computers to help AI calculation.
The Pantograph is (still) the current main source of power for rail transportation and it needs to be consistently monitored for its running condition, the data of which can be assessed by AI.
To ensure passengers in train coaches are evenly distributed, AI can accurately help the train driver relocate passengers within carriages to ensure no crowding and passenger safety is retained at all times.
Obstacles on the track of any type can affect running efficiencies – or even cause accidents. AI imagery analysis can detect dangers early and issue warnings to the train operator to avoid potential risk.
On board coach or carriage monitoring is an essential requirement in ensuring the safety of passengers. Analysis of passenger behaviour and unidentified left luggage or belongings through AI creates a more reassuring (and safer) carrier service.
Digital signage within the train carriage provides passengers with the ability to stay up to date where real time monitoring of rail traffic conditions is concerned and provides the additional revenues within rail transit through commercial advertising.
A Total Solution at the EDGE
Refined for Railway
The RC300-CS not only has a powerful performance, a rich mixture I/O interfaces, GPU expandability brought about by MXM, a wide voltage input and cellular 5G networking capabilities, but it has also been certified in line with European Union rail certification standards including electronic equipment EN50155, fire protection EN45545 and vibrations & shock (EN61373) to provide reliability, stability and scalability for the rail industry applications.
Powerful Performance Adds AI Horsepower
The RC300-CS utilises Intel’s 8th/9th generation desktop Core processors and can also be equipped with an Nvidia GeForce / Quadro MXM graphics cards along with a maximum of 2560 CUDA cores and 8TFlops, which is powerful enough to meet the needs of autonomous driving. The RC300-CS has the ability to run in a harsh operating environment of -25°C~70°C with a wide input voltage of 9-36V (or 77V to 137.5V) – to meet the needs where a ambient temperature differences and unstable voltages are prevalent. There are 4 x 15W PoE ports on M12 connectors and 2kV isolation protection DI/DO which provide for the deployment & connectivity of peripheral equipment within the train. Excellent heat dissipation efficiencies within the RC300-CS enables it to stably operate even when its deployed under the drivers or a passenger’s seat where there is little or no airflow.
Expandability For Always Connected
To ensure uninterrupted cellular network communication during train operation, the RC300-CS provides multiple M.2 & mini PCIe slots which can provide connectivity for LTE, 5G, WiFi/BT, GNSS, CANbus and MVB (Multifunction Vehicle Bus). Standard I/O interfaces are readily available including an externally swappable bay installed with 4 x 2.5-inch SSDs providing for high data storage capacities and retention of data integrity brought about by disk arrays.
- Securable I/O connectors to attach cables securely
- Up to date 5G-NR module support for high bandwidth & low latency train to ground communication
- High performance with efficient thermal dissipation capability
- Support of MXM graphics card for AI acceleration to provide better reliability and thermal dissipation.
- Intel® Gen 8th/9th Coffee Lake Core i7/i5/i3 processors
- Dual channel DDR4-2400/2666MHz SODIMMs, up to 64GB
- Wide temperature operation of -25°C ~ 70°C (configuration dependent)
- Multiple expansion interfaces for cellular or any module on Mini-PCIe
- (x3) or M.2 (x2) slots, such as CAN bus, LTE, 5G-NR, Wi-Fi, GNSS
- 4 x 802.3af PoE at 15W in RJ-45 or M12 X-coded connectors
- 4 x swappable 2.5” SSD supporting RAID 0/1/5/10
- 8-bit DI and 8-bit DO with 2KV isolation
- Up to 4 x RS232/422/485, 4 x USB3.0, 2 x USB2.0 ports
- Three display interfaces: VGA, HDMI, DP++
- +9~36VDC or +77V~137.5VDC power input range
- EN50155 and EN45545 certified.
Crystal TsengDirector of System Product (DFI)
“The purpose of RC300-CS is to be an in-vehicle system with comprehensive functions,” said Crystal, Director of System Products at DFI.
“It will be used for real-time information analysis regarding the vehicle through various sensors and cameras, and then generates useful information for peripheral systems, drivers or passengers, all the designs are meant to meet the demands required for a rail-based in-vehicle system, with high integration of various configurations.”
In-vehicle – VC300-CS
- 8th/9th Generation Intel® Core™ Processors
- Railway regulation: EN50155
- 4 M12 X-coded PoE ports at 15.4W
- AI accelerated: up to 115W MXM GPU
- Expansion capabilities: 3 mini-PCIe, 1 M.2 M key, 1 M.2 B key
(5G/LTE/WiFi/GNSS/CAN bus/MVB and function modules)
- Operating temperature up to -25 to 70°C (fanless)
Contact & Inquiry
We like to make life easier ….
Our technical & commercial team members will always provide you with valuable but impartial advice around the products and services that BVM supplies. With their collective backgrounds, they’ll provide you with the benefit of their knowledge and experiences when & where you need it. We’ll always help you in the first instance and get back to you when additional information is required.