BioAcoustic DeepTech

“We hear what others don’t see.”

Contact

Patent P.455746 · Edge AI · Living Lab

BioAcoustic

Off-grid: terrain without always-on power and without always-on network at every sensor.

A small vector instead of raw sound — decisions at the edge, year-round.

LTE or Starlink can be the uplink on the Master. The essence sits one layer down: mesh, LPWAN, battery life and inference across a large area — without raw signal at every point.

Node autonomy

~35×

longer on one cell — proof in the facts

Payload on the air

49 B

instead of ~128 kB WAV — compare in the facts

State of the art today

  • Modem + raw signal at every point
  • Battery and network die before a decision arrives
  • Heavy inference in the cloud — where there is no power

The BioAcoustic recipe

  • Worker: a feature vector, not WAV
  • Master: spectrogram + local inference
  • Hub-and-spoke (mesh): Master + up to 255 Workers · ~4 km · ~5,027 ha

Early anomaly prediction off-grid. One recipe, separate data per domain.

P.455746 Edge AI 2.5 yrs Living Lab

Technical effect · P.455746

State of the art vs BioAcoustic

Simulation from the patent filing (5000 mAh battery, 10 s cycle every 30 min). Range: LoRaWAN SF12 geometry (1 gateway).

Power autonomy

LTE + WAV 0 days
49 B + LPWAN 0 days

~35× longer run on one cell — the off-grid key.

Transmission payload

10 s WAV 0 B
Edge vector 0 B

131 072 B (128 kB) → 49 B. The spectrogram is assembled on the receive side.

Radio range

Wi‑Fi (order) 0 m
LoRa SF12 0 m

Radius ~4 km · up to 255 nodes (Workers) · ~5 027 ha per gateway (Master).

Energy use drop

725 → 20.4 mAh / day

0%

Network volume (100 nodes / month)

~17.6 GB → ~6.7 MB

99.96%

Source for autonomy and payload figures: patent filing P.455746, Table 6.

First deployment · TRL 4

BeeEcho AI

2.5 years Living Lab · patent P.455746

One platform that unifies fragmented attempts at hive acoustics into effective, cost-accessible colony monitoring and prediction — off-grid, year-round.

Pollination at stake

~EUR 153 bn

annual value of pollination services in the food economy (Gallai et al., 2009)

Colony losses today

~55% · ~44%

USA / year · EU winter — order-of-magnitude industry figures

Market today

  • Islands: labs, sensors, LTE — each separate
  • From autumn feeding to spring — a blind period; a stethoscope instead of knowledge
  • Raw audio kills the battery before a decision

With BeeEcho AI

  • Prediction of swarming, anomalies and disease — before the eye sees it
  • Year-round monitoring, including under winter cover
  • Targeted trips — less time, cost and carbon footprint

The beekeeper is not a radiologist reading spectrograms — they get a simple recommendation. 2.5 years Living Lab. Product: beecho.eu — in preparation.

TRL 4 P.455746 2.5 yrs Living Lab

Domain 2 · Opportunity

Same patent. Separate herd data · research underway.

Precision livestock and veterinary

Early herd-health listening without cables at every animal and without streaming the raw barn — where wall power and always-on network do not scale across the whole site.

PRRS stake (USA)

~USD 1.2 bn / year

production losses in swine (Holtkamp / Iowa State — 2024 update, 2016–2020 period)

Acoustic biomarker

Cough · breath

Precision Livestock Farming consensus: an early signal of respiratory disease — before the clinic and a site visit catch the outbreak

Market today

  • Islands: socket-powered monitors + continuous audio → cloud
  • Camera and the caregiver’s eye — delay and subjectivity
  • Raw barn audio does not scale beyond a building with power

With P.455746 · opportunity

  • The Worker rejects fan roar at the edge — a vector on the air, not WAV
  • 49 B about an outbreak / respiratory anomaly — spectrogram and inference on the Master
  • Mesh where there is no socket at every point: pasture, temporary housing, wide perimeter

The vet and stockperson get an outbreak signal, not a recording of the whole hall. Research is underway. We seek domain partners for pilot deployments.

Talk about a pilot
Opportunity P.455746 49 B

Domain 3 · Opportunity

Same patent. Separate forest data · research underway.

Precision forestry

Stand security and tree condition — where there is no always-on power at every point. A small vector instead of raw audio: an early signal of event and stress, not an archive of the clearing.

Illegal logging

~USD 30–100 bn / year

order of magnitude in the global wood trade (INTERPOL / UNEP — Green Carbon, Black Trade)

Coverage of one kit

~5,027 ha

1 Master + up to 255 Workers · ~4 km radius (SF12) — from the patent filing

Two signatures

Chainsaw · root cavitation

chainsaw — stand security; root cavitation — research path for potential drought and tree disease (AE / tree hydraulics literature)

Market today

  • Camera trap: a photo of the fact after the event for some anomalies — not analysis, not prediction
  • Patrol and satellite — delay, blind spots, cost of scale across km²
  • Always-on power and raw stream — do not scale across a forest complex

With P.455746 · opportunity

  • Security: 49 B about a logging attempt (chainsaw) — event alert, not WAV from the clearing
  • Research / condition: root cavitation — early vector of hydraulic stress (potential drought, tree disease)
  • Mesh: 1 Master + up to 255 Workers · ~4 km · ~5,027 ha — spectrogram and inference on the Master

Foresters and agencies get an early signal — of cutting and of tree stress — not a recording of the whole complex. Research is underway. We seek domain partners for pilot deployments.

Talk about a pilot
Opportunity P.455746 49 B

Domain 4 · Opportunity

Same patent. Separate infrastructure data · research underway.

Critical infrastructure — Oil & Gas

Integrity of the corridor and an early interference signal — no standing watch and no socket at every section. A small vector instead of raw vibration recording.

Incidents (USA)

~1.7 / day

order of magnitude of the long-run average of PHMSA reports (~628 / year)

Corridor segment

~8 km

1 Master + up to 255 Workers · ~4 km radio radius → ~8 km of route; at full staffing — order of ~31 m per Worker

Two signatures

Pump noise · delta

Silence Gateway: pumping noise vs impact (interference). Research path: leak / stress (NDT) · geophone / probe on the pipe.

Market today

  • Underground spills often found by a third party or field consequence — not early analysis (CONCAWE)
  • Patrol, flyover, methane satellite — delay or a different class of problem
  • Raw vibration recording along the whole route — a wall of battery and radio

With P.455746 · opportunity

  • Security: 49 B about an impact anomaly (delta) — alert, not a corridor stream
  • Research / integrity: stress / leak vector — research with an NDT partner
  • Mesh along the corridor: 1 Master + up to 255 Workers · ~8 km of segment — spectrogram on the Master

Operators and agencies get an early section signal — of interference and pipe stress — not an archive of the whole route. Research underway. We seek domain partners for pilot deployments.

Talk about a pilot
Opportunity P.455746 49 B

Domain 5 · Opportunity

Same patent. Separate ocean data · research underway.

Oceanography

Listening to the ocean without a continuous buoy stream — where battery and satellite link are hard limits. A small vector instead of a raw wave: an early ecosystem-stress signal, not an archive from the sea.

Hard coral cover

−8.9%

decline 2023→2024 (GCRMN Status of Coral Reefs 2025 — largest single-year drop on record)

Payload from the buoy

49 B

a vector instead of raw WAV — the condition for buoy telemetry (satellite / long-wave)

Two signatures

Reef · marine mammals

reef acoustic landscape (loss of sound biodiversity); migration paths. Hydrophone — same patent, different sensing physics.

Market today

  • Campaigns and cruises — rare, expensive listening
  • Continuous recording + satellite — a wall of energy and link cost
  • Bleaching / loss visible late — the effect, not an early signal

With P.455746 · opportunity

  • Condition: 49 B about a reef soundscape anomaly — not WAV from the buoy
  • Research / migration: mammal-signature vector — research with a domain partner
  • The buoy sleeps; Master on shore or vessel builds the spectrogram and infers

Scientists and conservation agencies get an early signal from the sea — not a full ocean recording. Research underway. We seek domain partners for pilot deployments.

Talk about a pilot
Opportunity P.455746 49 B

Domain 6 · Opportunity

Same patent. Separate field data · research underway.

Large-area agriculture

Early listening for invasion and feeding in the field — before camera and drone see the damage. A small vector instead of raw wind noise: an insect-biomass signal, not an archive of hectares.

Invasive insects (global)

≥USD 70 bn / year

economic losses estimated by FAO (plant pests / invasive insects)

Coverage of one kit

~5,027 ha

1 Master + up to 255 Workers · ~4 km radius (SF12) — from the patent filing

Two signatures

Insects · wind

acoustic biomass of invasion / feeding; Pulse Freq (~7 B in the vector) — isolating biological bands from open-field wind noise.

Market today

  • Camera and drone: damage on the leaf / in the stand — delay
  • Patrol and the agronomist's eye — blind spots on a large field
  • Raw recording + modem at the edge — does not scale energetically

With P.455746 · opportunity

  • Invasion / feeding: 49 B about an insect-biomass anomaly — not WAV from the field
  • Pulse Freq: wind filtering at the edge — research path with a domain partner
  • Mesh off-grid: 1 Master + up to 255 Workers · ~4 km · ~5,027 ha

Agronomists and plant-protection agencies get an early field signal — not a recording of the whole stand. Research underway. We seek domain partners for pilot deployments.

Talk about a pilot
Opportunity P.455746 49 B

Domain 7 · Opportunity

Same patent. Separate structural data · research underway.

Structural Health Monitoring

Structural integrity without wrapping the asset in kilometres of cable — a small vector instead of raw WAV from the pier: a fatigue spectrogram, not a vibration archive.

Bridges in poor condition (USA)

6.8%

share of bridges rated poor (ASCE Infrastructure Report Card 2025); ~168 million trips / day across such bridges

Payload from the node

49 B

vector instead of raw WAV — fatigue spectrogram at the Master (L2: asynchronous reconstruction of a crack image)

Two signatures

Concrete · cables

concrete fatigue, steel-cable cracking, structural resonance — research path with an NDT / SHM partner

Market today

  • Periodic inspection — lag behind degradation
  • Kilometres of diagnostic cable — cost and a single failure point
  • Raw stream from the pier — does not scale energetically

With P.455746 · opportunity

  • 49 B about a vibration anomaly — not WAV from the pier
  • Master: fatigue spectrogram (L2) — crack image from a small vector
  • Nodes without a fixed cabinet at every point — mesh where cable makes no sense

Engineers and asset owners get an early fatigue signal — not a recording of the whole structure. Research underway. We seek domain partners for pilot deployments.

Talk about a pilot
Opportunity P.455746 49 B

Domain 8 · Opportunity

Same patent. Separate rock-mass data · research underway.

Geology and mining

Listening underground where every byte and amp-hour costs more than on the surface. A small vector instead of a raw seismic record in the shaft.

Rockbursts (Poland)

5 fatalities · 2024

fatalities from rockbursts in mining (WUG — 2024 incident table)

Dynamics indicator

Pulse Freq · ~7 B

impact anomalies in the vector — without loading the mine network with a raw stream

Two signatures

Microtremor · gas

pre-burst microtremors and gas-related signals — research path with a mining partner

Market today

  • Sparse seismic campaigns — gaps in time
  • Raw shaft recording — energy and radio wall
  • Lag behind the rock-mass event

With P.455746 · opportunity

  • Pulse Freq / 49 B about an impact anomaly — not WAV from the working
  • Inference at the Master — without loading the mine network with a raw stream
  • Research with a domain partner — separate rock-mass data

Safety services and engineers get an early rock-mass signal — not a seismic archive of the shaft. Research underway. We seek domain partners for pilot deployments.

Talk about a pilot
Opportunity P.455746 49 B

Domain 9 · Opportunity

Same patent. Separate wildlife data · research underway.

Wildlife protection

An early threat signal off-grid without always-on power at every point — before a camera trap catches the fact after the event.

Coverage of one kit

~5,027 ha

1 Master + up to 255 Workers · ~4 km radius (SF12) — from the patent filing

Illegal trade (global)

~4,000 species

species covered by seizure records of illegal trade 2015–2021 (UNODC World Wildlife Crime Report 2024)

Three signatures

Shot · 4×4 · helicopter

events no eye is watching at 3 a.m. — research path with a conservation partner

With P.455746 · opportunity

  • 49 B alert for an event (shot / vehicle / helicopter) — not WAV from the reserve
  • Mesh off-grid: 1 Master + up to 255 Workers · ~4 km · ~5,027 ha
  • Master at the edge builds the spectrogram — research with a domain partner

Market today

  • Camera trap: a photo of the fact after the event — not analysis, not prediction
  • Delay and blind spots across a large area
  • Power and infrastructure that stay visible

Protection services get an early threat signal — not a recording of the whole reserve. Research underway. We seek domain partners for pilot deployments.

Talk about a pilot
Opportunity P.455746 49 B

Domain 10 · Opportunity

Same patent. Separate track data · research underway.

Rail infrastructure

Track pulse without a power cabinet at every sleeper. A small vector instead of raw listening along the whole line.

Broken rails (USA)

Leading cause

broken rails / welds — a leading cause of freight derailments in FRA cause analyses (USA)

Line segment

~8 km

1 Master + up to 255 Workers · ~4 km radio radius → ~8 km of track (range canon from the patent filing)

Three signatures

Axle · rail · ballast

axle wear, rail micro-cracks, ballast settlement — research path with a rail partner

Market today

  • Periodic inspection — gaps between diagnostic runs
  • Passing-train noise masks the defect
  • Raw stream + traction integration — cost and energy

With P.455746 · opportunity

  • 49 B about a segment anomaly — not WAV from the whole line
  • Nodes on sleepers — without a traction cabinet at every point
  • Master builds a segment-state map (~8 km) — research with a domain partner