Week 1
Task list:
Update the hardware list for the IoT devices in your possession. https://general.gqc.com/people/Pavan/Research/IoT%20devices%20List
Get the devices fired up.
- Node → S3 root → Ubuntu machine → TDEngine cloud
- Node → S3 root → S3 USB-to-HTTP → TDEngine cloud
Review the status of the time-series DB and associated documentation. DeepVibe related TDEngine documentation can be found here
Review Grafana and associated documentation. Documentation can be found here
Review the documentation for creating alerts in Grafana. Documentation can be found here
Review existing code and git status. Make sure everything is committed. DeepVibe related code is committed and in-sync with github.com.
Create a design for extracting acceleration history and running CEEMDAN on it on the desktop. We will then migrate it to the cloud at a later time.
Test decoding an acceleration measurement for a selected timestamp and write it back to a table,
acceleration_selectedin TDEngine cloud.Also populate
fft_selectedbased on theacceleration_selectedDefine new methods of analysis for the time-series data. Found the LoRA-WaN based research paper.
Implement Harmonic generation as a menu config entry in the sensor node. We should be able to customize the
- harmonic 1 frequency and amplitude
- harmonic 2 amplitude
- harmonic 3 ... 5
If the amplitude is set to zero, the harmonic should not come up
Couldn't figure out the rectify the relationship between timeseries amplitude and FFT amplitudes. Putting that on a hold for now.
- d
check the CEEMDAN paper on how the results are used.
Setup a testbed for vibration analysis. Sudhir to procure the parts.
Create a high-level diagram for AgPoint dev setup using mermaid.js. Use left to right template.
- Change the bullet points according to the diagrams.
Create an overview diagram for deepVibe using mermaid.
- There is a bug in mermaid.js which renders overlapped nodes. Added the diagram and a note about that.