Hello WildLabs Community,
My name is Luca and I represent Bearware. We are developing an integrated solution designed to mitigate Human-Wildlife Conflict (HWC) between brown bears and rural communities in the Carpathian Mountains.
We are sharing our current acoustic modulation strategy to discuss its viability within the conservation technology community. While literature confirms ursid sensitivity to high frequencies, static ultrasonic devices often demonstrate limited long-term efficacy due to habituation. We are prototyping a multi-modal acoustic approach to address this limitation.
Project Scientific Overview
The core of our project is a Fixed Perimeter Unit utilizing YOLOv8 on thermal streams to trigger a non-lethal acoustic deterrent. Unlike standard sirens or static ultrasonic repellers, our system implements a Dynamic Frequency Modulation protocol.
The Acoustic Hypothesis: Alternating Infrasound & Ultrasound
Our working hypothesis is that alternating between two extremes of the acoustic spectrum—ultrasound and infrasound—will disrupt the habituation process more effectively than randomizing a single frequency band.
Ultrasonic Component (~21 kHz):
We target the ~21 kHz range, which sits at the upper threshold of human hearing but is well within the sensitivity range of bears. This is intended to cause auditory discomfort without affecting local human populations.
Literature (e.g., Miller, 1978) indicates that while ultrasound can be initially effective (showing ~69% deterrence in polar bears), it is prone to habituation if the signal remains static.
Infrasonic Component (<20 Hz):
We are experimenting with high-pressure infrasound bursts interspersed with ultrasonic pulses. Studies on other mammals suggest that infrasound can induce physical sensations (such as dizziness or anxiety) distinct from standard auditory stimuli.
By introducing a physical sensation via infrasound, we aim to create an avoidance buffer that is more resistant to habituation than simple auditory annoyance.
Duty Cycle & Modulation:
The system does not emit a continuous tone. Upon positive AI detection, it triggers an alternating cycle (e.g., 20s Ultrasound / 10s Infrasound). This intermittency is designed to prevent the animal from predicting the stimulus pattern.
Safety & Acoustic Modeling
A critical constraint is ensuring safety for human settlements bordering the forest. We have modeled the propagation using the Inverse Square Law to optimize speaker placement.
We aim for a sound pressure level (SPL) of 60–100 dB within the effective deterrence zone (10–20m from the perimeter).
Using attenuation calculations, we ensure that the SPL drops below 80 dB (the EU safety threshold for continuous exposure) at the distance of the nearest human dwelling. This modeling allows for high-intensity bursts in the deterrence zone while maintaining a safe acoustic footprint for the community.
Discussion Points
We are analyzing the following specific areas and welcome technical insights or comparative data:
While infrasound effects are documented in elephants and cetaceans, data on bears is limited. We are interested in any existing field data regarding ursid behavioral responses to <20Hz or >21kHz signals.
Generating effective SPL at <20 Hz requires significant energy and driver excursion. We are evaluating whether the behavioral impact justifies the power trade-off compared to standard audible sirens in field deployments.
We are looking for empirical data on whether rapid switching between high and low frequencies (the "glitch" effect) significantly reduces habituation rates compared to simple silence intervals.
Thank you for your time and insights.
22 November 2025 3:43pm
Super! Is there anything further to read about the project so far ?
12 December 2025 6:32pm
Have you considered impact on other wildlife with your approach?
15 January 2026 5:22pm
I know very little about bears, but I do have quite a lot of experience related to deterrent sounds. Here I think the key feature is that the sounds must be more than irritating, uncomfortable, or just scary, otherwise they will ultimately habituate to them. Gas cannons are a classic example. They are very scary at first but after a while birds may even sit on the gas cannon and just jump up when it goes off.
I think the really key feature here is that the animals must be really scared of what might happen if they intrude into the no-go area. This can be created by sounds that stimulate the mind but the source remains unknown. i.e. it creates anxiety. I don't think animals can habituate to a scary or threatening unknown. I believe animals would rather just avoid an area than be subjected to increasing anxiety because of something always being there - always being heard - but never seen. It is the intangible.
I am not sure what would be scary to a brown bear, but it could be a bigger bear, or sounds of many humans and guns without any evidence of them. We have had great success over many consecutive years with virtual fencing for baboons using this technique.
PS These sounds should only be broadcast when repeat problem bears approach. In our case we only broadcast sounds when the resident baboons were known to be approaching the area. This avoids the problem of affecting other wildlife, or just too much sound which would likely lead to habituation again.
16 January 2026 8:19am
I think the key here is to involve a human in the loop. To avoid habituation you should avoid using the sound unless really necessary. Addition, in order to conform to the goal of trying to live with animals, if an animal is not causing harm or a threat, then why can't it share some space ?
So that means notification and review, one should ideally be able to review the situation before illiciting a response. Then you can choose to leave the bear alone if he's not cause harm and indeed enjoy the spectacle. That's of course more technically challenging but we should be striving for the best also out of fairness for the animals.
Kim Hendrikse