The health of your equipment can be measured by the sound it makes. This fact, exploited by technicians for decades, inspired our journey to develop our edge audio analytics system that will record and learn from the machine sounds. The result is a robust and performant monitoring system that will alert you when sudden events or creeping changes require your attention.
Our Edge Audio Analytics solution is capable of continuously recording and analysing up to 16 channels of 48kHz each on the edge hardware, and is equipped with a cloud agnostic control system utilising MQTT as the communication protocol.
On the edge, we run tens of models to capture the audio features and understand them. These models can be simple, like calculating the total loudness from a microphone, or advanced machine learning models used to classify certain audio patterns of interest. Only the relevant information, in terms of time series and audio snippets are uploaded to the cloud, reducing the cloud cost by several orders of magnitude.
Our Edge Audio Analytics solution was developed in partnership with Å Energi, StepSolutions, with partial funding from Innovasjon Norge.
Expert Analytics was part of the Distributed Artificial Intelligence Systems (DAIS) initiative, which is a pan-European project aimed at bringing faster and more secure data processing solutions through the development of edge AI software and hardware components. This large consortium was funded by the European Union’s Horizon 2020 research and innovation program.
Check our website dedicated for the Edge Audio Analytics solution!