AI Vehicle Computer For Machine Learning, Computer Vision, Sensor Fusion Supports Two CAN Bus Interfaces
Syslogic Industrial Computing introduced its AI railway computer RSL A3 for AI-assisted railroad applications. It comes with two CAN Bus interfaces.
The device utilizes the recent Nvidia Jetson AGX Xavier industrial System on Module (SOM), created for challenging environmental conditions. The module features a 512-core Nvidia Volta GPU with 64 Tensor cores, two Nvidia deep learning accelerators, two vision accelerators, and an eight-core Nvidia Carmel Arm CPU. The computer’s hardware base supports decision-making immediately within the railway vehicle without the need for a cloud connection. As a result, an AI performance of 32 Tera operations per second (TOPS) is feasible. Consequently, the system is suitable for AI applications such as video analysis, computer vision, inferencing, or machine learning. Special applications include automated driving, collision avoidance, signal and hazard identification, predictive maintenance, and visual inspection of rails.
The computer is tailored to the railway market by combining the Nvidia module with Syslogic’s carrier board and robust housing. The embedded electronic components are designed for 24/7 railroad operation in an extended temperature range from -40 °C to +70 °C. The circuit boards are protected against condensation through a conformal coating. In addition, screw-on M12 connectors and a galvanically isolated voltage input guarantee stable operation under extreme conditions.
The available hardware interfaces include two CAN Bus ports accessible via 9-pin Dsub connectors. The galvanically-isolated CAN Bus interfaces support the listen-only mode, i.e., the transmission function of the transceiver is disabled. In addition, there are sockets for micro-SD cards and mini PCIe, a display port, two USB 3.1, USB 2.0, and two 10/100/1000-Mbit/s Ethernet ports are offered as well.
The computer passed the EN 50155 railroad certification, complies with the EN 45545-2 HL 3 (hazard level) fire protection standard, and is shock and vibration resistant following EN 61373. Furthermore, the device offers the pre-installed Nvidia Jetpack SDK (Software Development Kit), providing tools and libraries for software development. The supported operating system is Linux for Tegra (L4T).
- NVIDIA Jetson JETSON AGX XAVIER
- 512-core NVIDIA Volta™ GPU with 64 Tensor Cores
- AI Performance: 32 TOPs
- 8-Core ARM v8.2 64-bit NVIDIA Carmel CPU
- 32GB 256-bit LPDDR4 RAM soldered on board
- 32GB eMMC 5.1 Flash Storage on board
- microSD Card socket
- CFast socket with retention frame
- Graphic interface DisplayPort 1.2
- 2x USB version 3.0 (Type A )
- Internal USB version 2.0 OTG behind the cover (micro USB Type AB)
- Ethernet 10/100/1000Mbit (M12 female x-coded)
- Ethernet 10/100/1000Mbit (I210-IT) (M12 female x-coded)
- 2x Active/passive-CAN ESD protected, isolated (DSUB9)
- Mini PCIe socket
- I2C bus
Raspberry Pi 4 With PICAN Dual CAN Bus HAT
The Raspberry Pi 4 offers ground-breaking improvements in processor speed, multimedia performance, memory, and connectivity compared to the prior-generation boards while preserving backward compatibility. In addition, the Raspberry Pi 4 provides desktop performance comparable to entry-level x86 PC systems.
The Raspberry Pi 4 comes in three onboard RAM options for further performance benefits: 2 GB, 4 GB, or 8 GB RAM.
The PiCAN2 DUO board provides two independent CAN Bus ports for the Raspberry Pi 4. It uses the Microchip MCP2515 CAN controller with MCP2551 CAN transceiver. Connections are made via a 4-way screw terminal. Furthermore, the board comes with a 5VDC 3A SMPS (Switch Mode Power Supply) to power the Pi and the PICAN board.