Calculating Wingbeat Frequency From Accelerometer Data

Travis Bayer

@TravisBayer

Movement Ecologist studying the where, why, and how animals move throughout our world. Current work is investigating the movement and energetics of large neotropical bats in a changing environment with ephemeral resources.

Groups

Data management and processing tools

Does anyone have any experience calculating WBF from ACC data? I'm trying to accomplish this in R. For the most part, I'm getting back pretty accurate number when going in to manually check its accuracy, but there are portions of the data where it will randomly generate very low WBF for short bouts of flight. I would chalk this up to gliding, but some of these low values encompass the whole flight duration.

This seems to only occur for short flight durations though (~4-6 seconds). To provide a bit more information, his flight data has been collected from bats.

Any R package or function suggestions or other software you have used to calculate this metric?

Thanks for any help!

Travis

Lars Holst Hansen

@Lars_Holst_Hansen

Aarhus University

Biologist and Research Technician working with ecosystem monitoring and research at Zackenberg Research Station in Greenland

23 February 2024 10:32am

No experience in this myself. Have you looked at references on swimming (diving) birds? I believe wingbeats have been derived from ACC data in studies of penguins and guillemots etc.

Cheers,

Lars

Matthew Stanton

@M_Stanton

Western Sydney University

Zoologist

23 February 2024 1:00pm

Sounds like an interesting problem and I would love to see the code you have come up with. Mega-chiropteran bats in my part of the world, usually get airborne by falling to get airspeed, which I'm guessing would look quite different on your accelerometer.

Is your accelerometer triaxial and body mounted?

Are you assuming flight because of body orientation?

Are you calculating VeDBA as part of your beat detection?

Is it possible your study subjects are hovering to feed for short flights which would require a steady head and body thus reducing the amplitude of movement or maybe increasing the wingbeats faster than your accelerometer sample rate?

Sorry for all the questions.

Travis Bayer

23 February 2024 1:00pm

I'm very close to solving the problem. Just waiting for a function to run on a fairly large dataset to see the results. I will share the repository link with you when it gets accomplished!

The species I'm working with roost atop cave ceiling as also drop from there to get airborne!

Yes, they are triaxial (Technosmart) and body mounted right on their backs.

So far, I have created thresholds for different metrics derived from the accelerometer data. Essentially, I sectioned out a bunch of ACC data where I am positive flight is occurring, and did the exact same with roosting, and crawling around/scratching(activity while roosting). From there, I plotted the distribution of all the metrics to see which metrics have unique distributions that were significantly different than roosting/activity.

Using those distributions, I created thresholds for the important metrics in which all flight behavior was either above or below a certain value for that metric. This got me to being able to construct a decision tree based on these metrics which had pretty solid accuracy.

The downside is a small chunk of flight from the beginning and end of flight bouts are not being included in the behavior classification. I noticed that their wbf during those small chunk are indicative of flight and am going to try and add wbf as the last decision on the tree to improve the accuracy of it.

VeDBA is also being included and calculated and based on the values for the thresholds I have created for flight it should not matter how high their head is, rather how low it is, when x y and z thresholds are also met. If that makes sense.

Hope I answered most of your questions!

Courtney Shuert

@CourtneyShuert | she/her

I am a behavioural ecologist and eco-physiologist interested in individual differences in marine mammals and other predators

26 February 2024 4:00pm

Hi Travis,

I had similar problems for calculating traveling on land in seals ('steps', if you will) for short bouts of repetitive behaviours. Part of the problem is possibly with the window size that you are using to calculate WBF. Many of the algorithms have a window size (how many time points/samples you include) and some degree of overlap (how much of the adjacent data you are considering). If the bout is too small relative to the window size and overlap, the frequency detected during non-flying periods adjacent to those points will 'bleed' into the calculation (also referred to as spectral leakage) and you may not be able to accurately pull out the signal of wingbeats during these short periods. I would look at some of the swimming papers as Lars mentions above for resources and probably do a better job than me of explaining :)

Are there other features from the accelerometry data that could be used to pull out these very short bouts of behaviour to classify them?