Understanding the driving environment requires a large amount of data that is collected by different sensors throughout the train and then processed by the vehicle's autopilot system.

The RC300-CS acts as an autonomous driving artificial intelligence computer system. Accompanied by high computing power and deep learning algorithms, RC300-CS makes real-time decisions by analyzing continuous data. Makes real-time decisions by analyzing continuous data collected from LIDAR, radar and cameras.

This artificial intelligence technology can make faster and more accurate decisions than humans and assist the ATO's signaling system. and assist the ATO signaling system system to support driverless train operations.

With the images/video collected from IP cameras, RC300-CS can analyze passengers' behavior with pre-trained models to track any person and any
abnormal activities and then alert the staff to take necessary actions. The video can be saved locally or remotely. The RC300-CS provides space for up to four solid state drives and supports RAID 0/1/5/10 to perform a reliable storage solution.

Video analytics technology also helps monitor the opening/closing of control train doors and directs passenger flow to emptier cars. Providing a better experience for passengers.

Fanless computers are very quiet, clean, reliable and flexible as to their location.

Fans cause noise.
Fans bring the air circulation, dust and dirt.
Fanless is stable without potential risks.
No maintenance needed.

The RC300-CS's power consumption is up to 250W with specific settings, it operates at up to 70°C without active fan.

• Intel OpenVINO toolkit
• 3rd party AI
• NVIDIA GPU-based AI frameworks

Cellular modules can be installed in MiniPCIe or M.2 slots, scenarios include:

• CAN bus or MVB module
  for device communications between carriages.
• Train-to-ground communication
• GNSS module helps real-time and high precision GNSS positioning
• Additional I/O ports for peripherals

The RC300-CS is particularly planned for the railway industry which needs solid and trusted products to deal with the EMC, vibration, ESD and control surge protection issues.

Thermal design of DFI’s RC300-CS meets the requirement of its performance. As an AIoT edge computer with NVIDIA GPGPU module installed, the RC300‑CS bolsters operation up to 70°C encompassing with 58W P2000 module and 60°C with 115W RTX2070 module without active fan.

This benefits SI to fathom thermal problem whereas the unit is introduced in the limited space with constrained stream like under the driver chair or passenger seat.