Parametric brings AI-based analysis directly into the vehicle with two products that work together: streamCTX, a streaming device for analog and digital measurement data with extremely low latency, and the MEC4000 rail computer with Neural Network Accelerator.
streamCTX - sensor data in real time
streamCTX records analog and digital measurement data, timestamps it and streams it with extremely low latency to an external device. In the opposite direction, it receives commands to control its digital outputs, and an API is available for telemetry. The streamed data includes:
- 4 digital inputs (level, frequency, counter), 100 SPS
- 8 analog input channels with IEPE, 16 bit, 20 kSPS, ±2.5 V / ±5 V or ±10 V, 100 blocks per second
- 3-axis acceleration up to ±4 g, 100 SPS
- 3-axis angular rate up to 500 dps, 100 SPS
- Mechanical shock up to 32g, 100 SPS
The robust IP40 sheet-metal housing is built for installation in vehicles and meets EN50155, with an operating temperature range of -40 to 85 °C and power supply variants for all rail and vehicle supplies.
MEC4000 - AI at the edge
The MEC4000 is a compact, fanless edge computer with an integrated Neural Processing Unit, LTE connectivity (optional 5G / WiFi 6), a high-precision GPS receiver and a managed Gigabit Ethernet switch. Software runs in containers (Docker/OCI compatible), so AI models and your own applications run directly in the vehicle, certified for railway environments according to EN50155.
Physical Edge AI
With streamCTX as the data source and the MEC4000 as the AI platform, sensor signals are evaluated where they arise. Results can trigger actions in real time, for example through the digital outputs of streamCTX. The receiving device can also be another existing system, so streamCTX fits into your current architecture.
Learn more on the product pages for streamCTX and MEC4000 or get in touch through our contact form.