Platform

Arches

High-performance edge computing platform powered by NVIDIA Jetson, optimized for AI/ML workloads, computer vision, and real-time processing across autonomous systems and intelligent edge applications.

Reference Hardware Designs

In-house developed hardware platform based on NVIDIA Jetson SoM, with custom carrier boards, integrated sensors, and proven interfaces. All designs are open source and available on GitHub.

Base Platform

  • • NVIDIA Jetson Nano / Xavier NX / AGX Xavier / Orin
  • • ARM Cortex-A57/A78AE CPU clusters
  • • NVIDIA GPU with CUDA cores (128-2048)
  • • 4GB to 64GB LPDDR4/5 RAM
  • • Gigabit Ethernet, WiFi 6, Bluetooth 5.1

Peripherals & Interfaces

  • • MIPI CSI-2 camera interfaces (up to 6x)
  • • PCIe Gen 4, USB 3.2, HDMI 2.1
  • • CAN, I2C, SPI, UART, GPIO
  • • M.2 Key M for NVMe SSD
  • • DisplayPort, eDP display support

Operating Systems

Custom OS builds with NVIDIA JetPack, optimized kernel, pre-configured drivers, OTA support, and comprehensive SDK. All software is open source on GitHub.

Embedded Linux (Yocto)

  • • Custom Yocto-based JetPack distribution
  • • Linux kernel 5.15+ with NVIDIA drivers
  • • UEFI/U-Boot bootloader with secure boot
  • • Custom BSP for Jetson modules
  • • CUDA, cuDNN, TensorRT integration
  • • OTA updates (Mender/OSTree)
  • • Diagnostic and profiling tools

Pre-configured Drivers:

GPIO, I2C, SPI, UART, CAN, PCIe, USB, Ethernet, WiFi, Bluetooth, Camera (V4L2), Display, GPU (CUDA), DLA, Video Encode/Decode (NVENC/NVDEC), ISP

SDK & Tools:

JetPack SDK, DeepStream SDK, CUDA toolkit, VPI (Vision Programming Interface), TensorRT, cuDNN, Application SDK, ESDK for kernel development

FreeRTOS

  • • FreeRTOS on ARM Cortex-R5 safety cores
  • • Safety-certified bootloader (ISO 26262)
  • • Real-time task scheduling
  • • Inter-processor communication (IPC)
  • • Hardware safety monitors and lockstep
  • • OTA updates (USB/Network based)

Peripheral Drivers:

GPIO, I2C, SPI, CAN, Timer, PWM, ADC, Watchdog, Safety islands, DMA

Middleware:

CAN stack, Safety monitors, Power management, IPC framework, Secure storage

Middleware & Frameworks

Communication

  • DDS - Real-time data distribution
  • ROS 2 - Robotics middleware
  • MQTT - IoT messaging
  • gRPC - High-performance RPC
  • ZeroMQ - Message queuing

AI/ML & Vision

  • TensorRT - GPU inference
  • DeepStream SDK - Video analytics
  • OpenCV - Computer vision
  • CUDA - GPU computing
  • PyTorch/TensorFlow - ML frameworks

Development

  • Rust - Safe systems programming
  • Docker - Containerization
  • Kubernetes - Orchestration
  • GStreamer - Multimedia pipeline
  • ROS 2 Nav2 - Navigation stack

Application Areas

Autonomous Vehicles

ADAS, sensor fusion, path planning, object detection, lane keeping, parking assist

Robotics

AMR, industrial robots, warehouse automation, collaborative robots, manipulation

Smart Cameras

AI-powered surveillance, facial recognition, behavior analysis, traffic monitoring

Industrial Inspection

Defect detection, quality control, optical inspection, measurement systems

Medical Imaging

Real-time image processing, diagnostic tools, surgical robots, AI diagnosis

Edge AI Servers

Multi-stream video analytics, inference at edge, distributed AI, smart cities