Data House

The Data House is a dedicated research facility for the deployment of state-of-the-art computer vision and provision of rea-time geolocation services. The Data House, positioned adjacent to the N4 freeway, is ideally situated to study the true distribution and composition of vehicles in addition to providing a clear line-of-sight to the sky for RTK (Real-Time Kinematic) GNSS (Global Navigation Satellite System) signal reception.

 

 

RTK GNSS

RTK GNSS relies on a permanent, stationary receiver (base station / NTRIP caster) on the ground with accurately defined coordinates to transmit correction data. This correction data - compensating for atmospheric delays, space weather, variation in temperature, relative humidity and gravimetric variations - enables the corresponding rover or field unit to achieve accuracies (relative to the antenna) of 14 mm. The permanent antenna (Tallysman TW1889 supporting GPS L1/L2, GLONASS G1/G2/G3, Galileo E1/E5b, BeiDou B1/B2) saw first light on the 19th of October 2020, marking the start of permanent RTK coverage on both Innovation Africa and the surrounding region. Additional testing, verification and surveying of the antenna is ongoing for the remainder of 2021, with permanent operations expected to commence in March 2021.

 

 

At the time of writing (19th of March 2021), the coordinates of the antenna as determined by the CSRS-PPP service is provided below, with an associated error of 0.003 m, 0.004 m and 0.015 m for the latitude, longitude and height respectively (Link to the latest CSRS-PPP report).

  •  25°44'30.42295" S, 28°15'29.92321" E, 1373.057 m (LLH)
  • -25.741784153, 28.258312003, 1373.057m (LLH)

The correction source operates over a SNIP-RTK2GO interface, with the correction source NTRIP stream accessible using the following URL: http://rtk2go.com:2101/SNIP::MOUNTPT?NAME=engineering4

 

 

Active Satellite Tracking Information

 

For more information on the hardware and NTRIP caster service, please contact André Broekman ([email protected]).

 

Traffic Quantization

A low-cost, open solution is currently being trailed at the Data House, consisting of a small developer kit computer (Nvidia Jetson TX2) which runs an off-the-shelf neural network for identifying and classifying vehicle traffic. A live video feed in the form of a webcam is used to provide the data stream for the software. The hardware is powerful enough to process the data in real-time through the neural network. with the statistics stored both on site and in out dedicated Innovation Africa data hub, integrating the stream with other complimentary sensors on the LoRaWAN network distributed across the campus.

 

- Author Andre Broekman
Published by André Broekman

Copyright © University of Pretoria 2022. All rights reserved.

FAQ's Email Us Virtual Campus Share Cookie Preferences