discussion / Biologging  / 28 January 2024

Seeking advice for remotely retrieving high frequency acceleration data 

Hi everyone! I'm currently working with a small team to develop and deploy custom affordable trackers for crocodiles, and I'm wondering if anyone has advice regarding data transmission. 

Our devices will be deployed on free-ranging American crocodiles in Costa Rica this spring and we are uncovering the many challenges involved with remotely monitoring these animals (especially on a budget), with the greatest of all being data communication. My priority is to conduct a behavioral study using GPS and triaxial acceleration / IMU sensor data, which can quickly turn into a large data load that is simple enough to store but difficult and expensive to send via satellite transmission. Right now, we plan that all data will be continuously logged and then off-loaded a few times a day whenever satellite communication is possible. 

We are working on 2 prototypes currently, one with an Argos module and one with Iridium. There are pros and cons to both, including limitations on how much data can be sent with Argos and the sheer price of Iridium subscriptions. We are looking for any guidance or input as we continue to moving towards the final versions of our devices. If anyone has any ideas or experiences with these issues, especially using satellite transmission for high frequency acceleration data, I would greatly appreciate hearing it! 

 

Some additional details:

We cannot recapture the animals or invest in the equipment and labor that would be required to retrieve an automatically released device, therefore any logged data must be transmitted. Additionally, no acceleration ethogram has to my knowledge been created for crocodilians yet so we will be training and validating a behavior classification model using captive alligators, however, this may not be ready before our study on wild crocodiles begins, meaning I'm unsure if downsampling or on-board preprocessing of the sensor data would be an option but I'm open to ideas.

 

Thank you to anyone who can help!

All the best,

 

Maggie




Hi Maggie!

This sounds like a really interesting - and challenging project!

UHF is used for tranfer of acceleration data (along with other stuff like magnetometry and GNSS locations) by systems developed by companies like Milsar and TechnoSmart

Ceres Wild uses satellite transfer over the GlobalStar but no raw acceleration data is transmitted.

I would fear the bandwidth would be too restrictive if you insist in using satellite based transfer of accelerometry.

Cheers,

Lars

This sounds like a lot of fun!

The lowest cost satellite comm I know of is https://swarm.space/.  It's pretty cool tech, but the data rates are quite low.

Do you have any idea how much data you need to move per day on average?

Perhaps LoRa would be an option?  Here is an example of a LoRa device https://www.adafruit.com/product/4284

LoRa is not satellite, but it has quite a long range which may suffice?  It is also pretty low data rate, but it is a good deal better than Swarm.  The best thing about LoRa is that the end-point can be very low power, so battery life can be quite good.

If you know somewhere the crocs frequent, you might also be able to work something out with Bluetooth LE?  Both LoRa and Bluetooth would require the installation of at least one base station.  Lora's base station could be substantially further way.

I have a few suggestions Maggie. Hopefully one or more will be useful. 

Preprocessing - Compressing the data before transmission will reduce file size and reduce transmission time.

Reduce data volume - I guess the animals are only active for some periods and a lot of the time are just "cruising along". So a useful strategy could be to only start recording when a certain threshold of acceleration occurs. And if you also buffer the data for say 10 seconds, you will get info about what's leading up to the activity. And adding a GPS module you will get some info on where the activity occurred plus movements prior. This will of course increase the data volume but only recording the most recent location will probably suffice as the location of the previous activity will tell you how far the animal moved.

Use LoRa instead of satellite - LoRa is well suited to this type of application. I saw in 

that the home range of crocodiles is relatively small so although a forested area will limit LoRa performance, the range should be sufficient. You could also use multiple base stations if needed. 

Of course eventually the data that has been recorded at the LoRa base station(s) will need to be collected. You could add a long range radio to the base station, or just collect the data 'manually'.

Hope this helps