AI enabled Sensor Fusion Kit Lite

The AI-enabled Sensor Fusion Kit Lite (AI-SFK Lite) is an integrated, performance-oriented, Camera Vision and mmWave RADAR fusion platform that meets the compelling needs of modern AI, Deep Learning, and Video Analytics applications.

OVERVIEW

The NVIDIA Jetson SoM based AI-SFK Lite (AI-enabled Sensor Fusion Kit Lite) from Mistral is an integrated, performance oriented, Camera Vision and mmWave RADAR fusion platform that meet the compelling needs of modern AI, Deep Learning, and Video Analytics applications.

The AI-SFK Lite (Camera Vision + mmWave RADAR) is uniquely designed to provide range precision along with superior imaging capabilities for applications such as smart retail, industrial 4.0, robotics, smart building, smart city, ADAS, autonomous vehicle, among others. The camera & mmWave RADAR complement each other for object detection, classification, range, velocity and elevation parameters. The platform provides vision analytics and RADAR data fusion in real-time.

The AI-SFK Lite Kit consists of a Mistral’s Neuron Base Board (NB-Basic) mated with NVIDIA Jetson Xavier NX SoM along with Mistral’s 77GHz mmWave RADAR Module and a Camera Module (CSI-based 8 MP / 4K colour sensor) supporting up to 21 fps.

The board also supports NVIDIA JetPack SDK, which enables the development of AI applications with accelerated libraries supporting all major AI frameworks as well as computer vision, graphics and multimedia among others.

In addition to providing engineering services for product/system development, Mistral also provides customization services around SFK Lite for Custom camera integration, Other sensor integration like thermal, LiDAR, etc, Application Software, ML/DL algorithms development and integration, Remote firmware upgrade, Audio and video, Cloud and mobile apps, PoE, Power optimization, Mechanical enclosure customization, FCC/CE certification support, Chirp Profile Tuning support and AI-SFK Lite with 60GHz RADAR module.

TECHNICAL SPECIFICATIONS

  • 384-Core NVIDIA Volta GPU with 48 Tensor Cores
  • 6-Core NVIDIA Carmel ARM v8.2 64-bit CPU
  • 2x NVDIA Engines with 7-Way VLIW Vision Processor
  • 8 GB 128-bit LPDDR4x, 16 GB eMMC 5.1 and 10/100/1000 Base-T Ethernet
  • CSI2-2 Lane Camera Interface
  • Sony’s iMX219 from Leopard Imaging
  • FMCW Transceiver with Integrated PLL, Transmitter, Receiver, Baseband and ADC
  • 76 to 81GHz Coverage with 4GHz available Bandwidth
  • M.2 M key based 250GB SSD
  • GigE, Wi-Fi 802.11 a/b/g/n/ac and BT 5.0
  • Supports Ubuntu 18.04 Linux OS natively
  • OpenCV, VisionWorks, TensorRT, LibArgus, CUDA, CuDNN, ROS1 Melodic