Hardware - NXP NavQplus

 

NavQPlus mission / payload companion computer


  • Open reference design, NXP EVK SOM, with customization also available from 3rd party.
  • Machine Learning using 2.3TOPS NPU accelerator, eIQ ML  Software development environment w/TFLite, ArmNN, ONNX
  • Yocto Linux, Ubuntu POC** and ROS2 enablement.
  • Vision – Dual MIPI camera w/ ISPs, hardware codec accelerators
  • WiFi 5 / BTLE 5.0
  • Dual Ethernet 100BaseT1 “2-Wire” + IX industrial 1GB
  • 2x USB-C with up to 20V power input + external power input
  • 2x CAN (TJA1463)
  • SE050 EdgeLock secure element with NFC interface
  • RTC with tamper timestamping
  • PCIe expansion
  • Supports up to three displays simultaneously MIPI,LVDS,HDMI (w/CEC)
  • Other uses include as a QGC ground station or smart remote controller, or machine interface terminal.

** Customized Ubuntu Core also available from Canonical  

Available from NXP.com
Documentation on Gitbook



Popular posts from this blog

Successful test of ML based object detection

Image
Successful test of ML based object detection  The WRD team was out recently testing ML based object detection on real hardware. Supported by DLRG team we had a windsurfer and a kite surfer placed in the water. Task of our NXP Auriol drone equipped with SIYI ZT-30 gimbal camera and NXP NavQ+ (i.MX8M-plus) with internal NPU was to detect the objects in water and deliver a live video stream of both RGB and IR camera to the ground control van. Object detection worked very well. Only challenge was the limited bandwith of the LTE connection because we streamed three (3) RTSP videos live (IR, RGB and AI-ML output). Special thanx for acting as "rescue targets" floating in the water for several hours to Christian Lautenbach from kiteguide-spo.de  and Kevin Peklar. Looking forward to the next successful test.
Read more

The idea behind WaterRescueDrones project

Image
The idea behind WaterRescueDrones project  A SAR, SearchAndRescue mission consists of two main activities. Search and Rescue.. While at rescue you know where a person or object is located and the task is to get it back the search is most of the times the most time and effort consuming part. Often there is just a hint in which area someone/something is located. In the sea drift and wind is playing a big role. The longer a search takes the more an object can float away from it's initial location. Therefore being able to cover a wide area within a reasonable short time is key in a search mission. Being airborne gives several advantages: You see more (wider area) compared to from a small boat where u see just see ~30m wide. You can cover a bigger area within same time because of higher travel speed. It's less risky for the search personnel. Because even in the wadden sea there are waves with several meter height possible. The search could be done with a helicopter. But a drone is m...
Read more

Support to fire brigade at a wildfire

Image
Support to fire brigade at a wildfire On May 6th we were preparing a flight test at the beach of St.Peter-Ording to collect training data for the onboard AI. What started as a routine flight test day turned into a real emergency scenario soon. Right before takeoff to our first planned flight we got requested to support the local fire brigades at a wildfire at the salt marshes at the south beach. We did 4 flights with over 20min each providing live video feed from our onboard RGB and IR camera into the ground control van helping the operations team on a live situational awareness. Related media relase https://www.ndr.de/nachrichten/schleswig-holstein/Salzwiesenbrand-in-SPO-hinterlaesst-verheerenden-Schaden,regionheidenews2968.html (German) RGB camera IR camera On the ground
Read more