Technology
The BB-S architecture: from sound to operational data.
Four stages
- High-fidelity microphones
- Audible, ultrasound and infrasound coverage
- Microsecond-level synchronisation
- Protected 24/7 recording
- Protected copy of the original data on cloud infrastructure
- Acoustic mapping of the environment
- Analysis of variations and anomalies
- Sound source separation
- Environment-specific learning models
- Configurable notifications and outputs
- Sharing of metadata, indicators and alerts
- Integration with existing platforms and systems
Field validation
The BB-S V2 prototype operated 24 hours a day, 7 days a week, for 9 months in a real operating environment.
26,000+hours
Environmental acoustic data recorded.
70 million+
Audio files generated in the proprietary dataset.
17,000+hours
Environments already processed by the AI models developed by Sonogram.
This activity validated the continuous operation of the BB-R architecture, data management and segmentation, the creation of the dataset and the operation of the first AI models over extended periods. Quantitative metrics for industrial anomaly detection - including false positives, precision/recall and lead time - are the subject of the field validation with the Early Adopters.
Maturity and composition
It starts with acoustic memory. Intelligence is added when the data is mature.
TRL 6/7IRL 5
- BB-R and BB-MIC
- Acquisition and recording units.
- BB-AI
- Intelligent analysis and anomaly detection modules, which can be activated progressively.
- Evolutionary approach
- It starts with acoustic memory, the baseline is built, and intelligence is added when the data is mature.
- Integration
- BB-S is designed to integrate with predictive maintenance tools, CMMS, supervisory systems, fleet management and the customer's dashboards, without replacing them.