More information on the cattle tracker please

We are an environmental organization in New Zealand (NIWA) looking to track sheep movements and are wondering whether the cattle tracker unit you have developed may be useful for us.

We want to track at least 80 sheep and need cm range accuracy (< 1m at least). We will be logging once every 5 - 15 minutes for a duration of approximately one month.

So we need to know:
resolution/spatial accuracy of the results and how this accuracy is achieved (dead reckoning?) and what sensors are employed to achieve this (accelerometer/magnetometer updated by GPS?)

Can the LoRa be switched off if not required.

Any information you can supply is much appreciated.



Hi @katesfb,

The board is capable of getting a fix <1m accuracy.
The software at the moment isn’t, due to battery optimization we look at a minimum satellites after this we turn off the gps module.
The software is opensource and so you can make your own changes to achieve the <1m accuracy.

In New Zealand you have a lot of sun hours, so I expect this will be possible.
I advice you to make a power calculation if this is possible.

On the board we have a magnetometer and accelerometer.
The GPS fix can be time based, every x minutes. And/or by movement trigger from the accelerometer.

The RN2XX3 module can be put into deep sleep.

Let me know if you need more information.

Best regards,

Richard… In support of what Jan has posted, I can share some testing I’ve been doing with the Sodaq tracker. In this case, the tracker was placed away from any metal objects and was not moved. As you can see, there is significant variation in position reporting, with the majority of reports showing less than 25 meters of displacement from the actual position.

Note that this testing was done with the default position parameters Jan mentioned. I have not tried going for higher accuracy yet, although we would like to extend the recurring interval to perhaps 30 or 60 minutes and enable movement reporting.

And for what it’s worth: We increased the fix time from the default 120sec to 180sec. With reports every 30 minutes overnight (12 hours, almost every report was identical).

This might not do anything. If you had a fix within 120 sec, you will also have it within 180 sec.

The GPS Fix Timeout (sec) (gft=): 120 is a timeout.

A GPS fix takes every 4 hours a refresh time of 30 seconds to download the required data. After this it is just a few seconds of getting a correct fix if you are in an open field.
It can take a bit longer when you are surrounded by buildings.

I hope it’s clear, let me know if you have questions.

Best regards,

Ahh - I misunderstood the parameter. So there is no configuration change that would improve location accuracy? That is, to get the cm level accuracy is just a function of openness of the ground?

The only parameter that would improve the accuracy at the moment is the minimum satellite count.
Minimum sat count (sat=): 4

1 Like

What about using a Kalman filter. The Arduino IDE has a library for it. Cheers.

Richard, you may want to check out Smart Paddock with your tracking requirements