Developer Marcial / Soil Moisture with LoRa
This is your Edge Impulse project. From here you acquire new training data, design impulses and train models.
About this project
Getting Predictions Back from the Edge
Deploying a small MCU device to the edge is one thing. When the device is part of a larger ecosystem, getting the machine learning predictions and other data back to a central evaluatuion center is crucial. Here a long range radio is investigated that uses the LoRa radio protocol.
The overall project runs a machine learning model developed with Edge Impulse Studio on a Sony Spresense microcontroller. The machine learning inference predictions are sent over LoRa to a LoRaWAN gateway. The gateway connects to The Things Network (TTN). A custom app in TTN uses a webhook to send data to ThingsSpeak for charting and public review.
This project uses a Sony Spresense MCU programmed with the Arduino IDE and an Arduino library deployed from this Studio. The execution time for the moisture data to run thru the Edge Impulse DSP and Classifier is sub-millisecond!
The code and overall project details are here
Download block output
| Title | Type | Size | |
|---|---|---|---|
| Flatten training data | NPY file | 63 windows | |
| Flatten training labels | NPY file | 63 windows | |
| Classifier model | TensorFlow Lite (float32) | 3 KB | |
| Classifier model | TensorFlow Lite (int8 quantized) | 2 KB | |
| Classifier model | TensorFlow SavedModel | 9 KB | |
| Classifier model | Keras h5 model | 3 KB | |
| Anomaly detection model | JSON | 7 KB |
Clone project
Summary
Data collected
32sProject info
| Project ID | 167207 |
| Project version | 5 |
| License | Apache 2.0 |
