Project Ngulia

What we do

50 years ago, several hundreds of thousands of rhinos roamed the African savannahs. Today, due to poaching the remaining 20,000 are living in national parks and sanctuaries, protected by rangers. In many parks, there are as many rangers as rhinos. Globally, over the last ten years, the elephant population was reduced by 30 percent and more than 1,200 rhinos per year have been slaughtered by poachers since 2012.

In Kenya, the black rhino population is facing extinction. In 1970s there were 20,000 black rhinos. Today there are only about 750. Wildlife crimes – including trafficking in elephant ivory and rhino horns – have skyrocketed, threatening the economic viability of many African countries that rely on wildlife tourism. The current situation has significant implications for Kenya’s development as about 10 percent of Kenya’s GDP is associated with the tourism industry that relies on popular wildlife attractions as elephants and rhinos.

To meet these daunting challenges, Project Ngulia develops cost effective technology and innovation to be deployed in wildlife sanctuaries and national parks as part of a holistic wildlife security and monitoring solution.

Screenshot of the mobile app from the testing deployment at Kolmården Wildlife Park.

Phase 1: The Command, Control and Communications (C3) system

From 2015 to the present, the KWS rangers at Ngulia Rhino Sanctuary and the Project Ngulia team have jointly developed the WildProtect Mobile App and Dashboard that form the Command, Control and Communications (C3) system that digitises the workflow of the rangers in Ngulia.

WildProtect Mobile App

Before our app was released in 2015, the rangers reported their patrol efforts and observations orally every month. Only urgent reports were communicated using their radio. Since then, our app provides the ranger with a complete reporting tool with the following main menus:

  • Animal reports, for both indirect and direct sightings of different animals, but with a detailed report form for our flagship species, rhinos and mountain bongos;
  • Security alerts, with pre-defined one-click reports for fence breaks, footprints, gun shots and fires;
  • Information reports, not falling into the above categories. The rangers use this for reporting natural water holes or other features of interest;
  • Patrol logging, that saves the route patrolled by the ranger team, as well as meta-information such as patrol time and distance.

The rangers can for security reasons only see their own reports for the last few days. When a commander logs into the app, he can see all ranger reports. Furthermore, the commander gets a notification in real time whenever a security alert is logged, so he can take immediate action.

WildProtect Web Dashboard

The Dashboard is a web interface accessible from desktop computers, tablets as well as smartphones. The Dashboard collates and visualizes all information collected by the rangers out in the field, namely:

  • Direct rhino or bongo and other animal sighting reports;
  • Indirect rhino or bongo and other animal sighting reports;
  • Security alerts;
  • Information Alerts;
  • Ranger patrol information i.e. date, time, distance and GPS tracks

Screenshot of the web dashboard from the testing deployment at Kolmården Wildlife Park.

All the information is accessible both in a chronological list view and a map view. The Dashboard is also an administration tool, where new users and monitoring areas can be added. Ranger performance can also be monitored via the ranger performance statistics, which displays each ranger’s efforts over the time period, that is their patrol distance and coverage in are shown, as well as the number of reports they have made. It also features a report scheduler, where the commander can request full reports at regular times, either as readable pdf file or as a csv file ready to be imported for further analysis or input into a management information system.

The dashboard is cloud based so no information is saved on the local device, and all information in the database and in the browser is encrypted. There are several user categories, each with differing permissions, so a user can only see the information they have the authorization to access.

Screenshot of the web dashboard from the testing deployment at Kolmården Wildlife Park.

Field Control Center

We have developed another module for setting up a field control center in the KWS Offices.

The Field Control Center is designed to be set up in a radio room or other facility that is manned 24/7. The system consists of a:

Display monitor – a large screen (24”) that displays a live view of the web dashboard.

Control center module – This is a small box (housing a Raspberry Pi) that is connected with HDMI and USB cables (for power). The module is also connected to the display monitor, and a local WiFi access point. This box gives access to the live view of the Dashboard, showing a summary of the recent reports and activities.

Our robust field station (solar panel not included in image), featuring space for batteries, charge controller, voltage controllers, USB hub for charging, modem to cellular network, WiFi access point, LoRa gateway and GPS.

Phase 2: The Virtual Watch-Tower

When we initiated Project Ngulia in early 2014, we interviewed the rangers and commanders in the Ngulia Rhino Sanctuary and asked what kind of tools they were lacking and were highly desirable from their point of view. The most common answer we received was a watch tower, their most conceivable option to get the ability to get above the vegetation to both see and hear better. Our ambition is something much more effective to provide them with “super-eyes” and “super-ears”. Sensors in the field can automate many of their duties, and like their own observations, all the information from the sensors will be completely integrated into the App and Dashboard, making their work more efficient and effective.

We translated the watch tower wish into the concept of a Virtual Watch Tower (VWT), which will integrate one or more of the innovations below.

Innovations

  • Intelligent tags for monitoring wildlife with extended lifetime. One of the key performance indexes in rhino monitoring is the critical sighting time (CSI). If the time since a certain rhino was seen and identified exceeds the specified CSI, the rangers have to actively search and find this individual. Reliable identification requires seeing all ear notches. This is a tricky and time-consuming operation. Some rhinos are equipped with radio transmitters. Our concept tag has besides a much longer lifetime also the benefit of performing motion analysis. The current version can count the number of steps over time. Since new software can the uploaded over the air, future biological studies can add more features, such as correlating horn angle to well-being and activity. For the purpose of security and battery lifetime, information will only be transmitted in proximity to one of our VWTs. The first version of this tag was developed during 2020 in collaboration with the LiU project Agtech2030. Patent granted: SE542278. A method and a system for monitoring rhinoceroses. Fredrik Gustafsson. 2017-07-05

  • Connected camera traps. The camera traps that are used extensively today by KWS in all of their rhino sites are conventional trail cameras. They need battery replacement every week, when also the SD card is exchanged and brought back to the office, where a researcher analysis the images. We have deployed a few of a new generation of commercial trail cameras that have a SIM card slot and antenna and connection to cellular network. They also have external solar panel, eliminating the need for battery replacement. The innovation is that these cameras provide real-time images to the Dashboard, in a similar way to the rangers’ manual reports. This also makes it useful for security purposes, where the camera can be placed along tracks where poachers are likely to approach.

  • Intelligent camera traps. The connected camera traps still need visual inspection and manual classification. Security applications would require continuous attendance. We have during 2020 by the work by Sara Olsson and Amanda Tyden developed machine learning algorithms for automatic detection and classification of humans and the largest savannah animals. This camera is one part of the VWT.
  • Shot detection and shooter localization. During the fall of 2020, a project was launched to develop a microphone array that can detect gun shots and also triangulate the position of the shooter. This is another component of the VWT.

  • Environmental monitoring. A central theme in the LiU project Agtech2030 is monitoring of environmental parameters such as soil moisture, water quality and weather stations. Electric fence monitoring also fits in here. These are the kind of sensors that would complement the rangers reports.

  • Robust field station. This innovation was initiated by Take Anstoot at our partner Modio, who was responsible for our ten installations of the solar charging stations in Ngulia. He found that the rangers quite often misuse the installations by incorrect charging. The new concept is to integrate all required components into a closed box where only USB connections are available for charging. The concept was generalized into two directions. A module for small households, leading to the spinoff company Solarbora AB, and the robust field station also including all kind of radio equipment that we need for our virtual watchtower. It can be seen as a base station for a set of VWT’s.
Top