This document is Copyright © 2013 by Lars Boegild Thomsen - all rights reserved
This document describe a system for tracking livestock (or any other things) location with real near real time reporting  even in locations with no commercial mobile coverage.
The system is based on establishing a wireless network based on relatively low frequency 900 MHz transmitters. This network will be established by putting up a number of cheap and self contained base stations that together will form a so-called mesh network.
The base stations, which will be solar powered, will have a range of up to 10 km line of sight - perhaps even more under ideal conditions. It is not necessary to “connect” the individual base stations. As long as a base station can “see” at least one (but preferably more) other base stations, the network will arrange and manage itself. For optimal reliability the individual nodes in a mesh network should be able to communicate with more than one neighbor, so that multiple paths are available through the network.
The individual animals will be equipped with a collar that attaches a small waterproof box which will contain a GPS receiver, a wireless transmitter/receiver and a rechargeable battery. Also attached to the collar will be a solar power cell, which will be dimensioned so that the battery will be kept fully charged.
The animal tracking units will check their exact location and report this location back to a central server at fixed intervals - for example every 5 minutes.
- In order to save power, the animal tracking devices will not transmit a continuous stream of location date. Rather they will sleep and wake up at regular intervals to update the location. These intervals are expected to be rather short - for example every 5 minutes or shorter depending on available power
- 900 MHz will have a considerable better range than devices operating in the more common 2.4 GHz bands at the cost of lower bandwidth - which for this application is entirely acceptable.
A number of systems for tracking livestock already exists. The following sections describe the existing options.
RFID Based Systems
These systems rely on a cheap and small tag that will be attached to each animal. The RFID can be passive (not powered) or they can be active (battery powered). Even the powered tags will have a very limited range - typically around 10 meters. In other words, these systems can only be utilized when livestock pass through a pre-defined checkpoint - for example when being loaded on a truck.
Offline GPS Systems
These systems work with a GPS receiver attached to each animal. The GPS receiver will log locations over time, but these data will later have to be offloaded by some means (cable, USB stick) and transferred to a central system. VFR Based GPS Systems
Like the offline system, these systems work with a GPS receiver attached to each animal. They do however offer means to offload the data by radio transmissions. Typically these radio transmissions will require receiver equipment to be close to the animal being tracked as the VFR based radios offer limited range (typically significantly less than 1 km).
GSM Based GPS Systems
These systems includes a GSM modem that will transfer the GPS data to a centralized server. Unfortunately they are dependant on existing GSM networks and they are quite costly as some kind of data subscription is necessary for each animal being tracked.
Satellite Based GPS Systems
These systems - like the GSM based systems, include a transmitter that will transmit the recorded GPS data, via a satellite (Iridium and Globalstar), to a central system. Unlike GSM based systems a global coverage is possible but it is extremely costly both in terms of equipment and in terms of satellite usage.
Triangulation Based Unknown Backhaul
One system  uses distributed base stations to triangulate the location of animals. They claim a 7 km spacing between base stations with a central server doing the number crunching, but they fail to explain how data is transferred from the towers to the central server (cables? microwave links? gms?).
Wireless base stations will be placed at regular intervals in a triangular pattern. The distance between each tower can be up to 10 km on flat unobstructed terrain, but this distance might be lower in hilly terrain or in terrain with obstructions such as buildings.
The base stations themselves can be physically mounted in a number of different ways. The simplest approach would be a simple wooden or metal stick hammered into the ground. And alternative would be mounting them as high as possible in trees as long as the devices will still be hit by sunlight to keep the batteries charged.
Animals within this area that are wearing a tracking collar can be individually tracked in near real time.
The base stations and the wearable tracking devices are architecturally identical:
A solar power cell will keep a rechargeable battery charged. The battery will supply power to a GPS Receiver, a CPU (Central Processing Unit) and a Wireless Transmitter & Receiver.
The devices - both base stations and the tracking devices - will organise themselves in a so-called mesh network. This network will be self organising and self healing, in other words, should a node drop out, this might (in case of a base station) create a temporary black spot, but it will not disrupt the overall network operation as long as the network nodes can find paths around the disabled node. Similar, the tracking devices arrange themselves as part of this network.
- Or it might indeed be slightly longer. In fact the manufacturer of the wireless transmitter & receiver is promising 9 miles line of sight, equivalent of 14 km.
- In reality the area will be smaller due to the rounded corners.
- The largest cattle station in Australia, Anna Creek, cover an area of roughly 24000 km2 - roughly equivalent of the Republic of Macedonia or Belize. This area could be completely covered with less than 220 base stations.
Base Station Layout
180 km x 180 km = 32400 km2 using 289 base stations in square layout.
The system described in this document is currently at a prototype state where the functionality can be proven. In order to make the individual devices affordable further funding will be necessary to finance final product development.