1. Pilot project
- Introduction and cooperation setup
(1 – 2 weels)
- First sound recordings collection
(1 – 2 weeks)
- Recordings analysis, neural network model training (2 – 4 weeks)
- Pilot project results presentation
- Operation environment solution design & testing (1 – 2 weeks)
- Neural network model update
(1 – 2 weeks)
- Operation environment implementation
(1 – 8 weeks)
- Monitoring & updates
(1 – 8 weeks)
1. Introduction and cooperation setup
We will present you with our technology and solution. We will briefly discuss its advantages for your business, the main benefits, constraints, technical options and the implementation process.
As a result of the discussion, we will prepare the cooperation offer tailored to the specific needs of your business. The summary of the expected benefits, costs and required collaboration will be provided.
2. First sound recordings collection
As part of the pilot project, our technicians will come to your factory to collect the “training” recordings of your machines. We need to record your machines in both proper and broken states to achieve the best results.
The ambient noise in your operation environment will be filtered out by using our very sensitive contact microphones. Optionally, we may also utilize other sensor data sources than sound. Our solution is, among other things, able to process the temperature, pressure or humidity data to improve the final accuracy of the analytical model.
3. Recordings analysis, neural network model training
By processing all the gathered sound recordings in powerfull cloud servers, we will train the neural network model. After being trained, the model will be able to tell you whether your machine works properly, is broken or will probably break in the near feature.
Not only we will help you to cut your unplanned downtime, maintenance time & costs but our technology can also lower the number of malfunction products shipped to your customers. That would definitely boost your brand reputation and avoid costly fines.
4. Pilot project results presentation
After finishing the pilot project we will comprehensively summarize the results and accuracy of our model and answer all your questions.
If our solution meets your expectations, we will discuss the plan of its operation environment deployment and the details of the final contract.
5. Operation environment solution design & testing
We are working hard to prepare plan of the deployment of our solution to the operation environment. All the important decisions are consulted with your experts.
The designed solution will have zero negative impact on your operation environment and won’t burden your employees. There is no need for an expensive integration projects as we do collect all required data by our IoT devices.
6. Neural network model update
The neural network model learns from the data. More data is available, better results it would achieve. As part of the installation, we do focus also on extending the number of recordings. So we can recognize more different issues with higher accuracy.
Your team will access on-line tool for audio data tagging. So each recording provides a valuable input for the algorithm improvement.
7. Operation environment implementation
The approved and tested solution will be deployed to your operation environment. The solution works online and offline. We do provide our IoT devices running the analytical software, so there is no need to for the intensive data transfer to cloud.
We can also adapt the system to work on edge IoT devices of your preference. We are also working with partners to build a low-power solution using batteries.
8. Monitoring & updates
We continuously monitor our solution to ensure trouble-free operation. In order to improve the accuracy of our solution, we can regularly update the neural network model with the newly collected data.
Your machines would be like a living organism. Automatically updating and continuously learning from the shared experiences.