StressLess IoT

StressLess is an IoT product aimed at monitoring workers' physiological stress levels in real-time to improve their performance and reduce the risk of injury. The system collects biometric data from workers, analyzes the data to determine stress levels, and provides customized training recommendations based on the workers' physiological stress levels.

StressLess IoT

Database Structure

The database contains records for the following worker types:

Manual Workers (MW)
Forklift Operators (FO)
Inventory Managers (IM)
Maintenance Technicians (MT)

Worker IDs are represented as follows:

MW1 - Manual Worker 01
MW2 - Manual Worker 02
MW3 - Manual Worker 03
MW4 - Manual Worker 04
MW5 - Manual Worker 05
...
FO1 - Forklift Operator 01
FO2 - Forklift Operator 02
FO3 - Forklift Operator 03
...
IM1 - Inventory Manager 01
IM2 - Inventory Manager 02
...
MT1 - Maintenance Technician 01
MT2 - Maintenance Technician 02

System Overview

The system consists of the following main components:

IoT devices to collect biometric data from workers
Azure IoT Hub to securely manage and transfer the data
Azure Cosmos DB to store the collected data
A machine learning model to analyze the data and predict stress levels
A user-friendly interface to display the training recommendations

Initial Setup

Before running the main scripts, ensure that all the documents in the Cosmos DB have their "Processed" flag set to false. This allows the RandomForest algorithm to start with a clean state and set the flag to true when the model has been trained.

Execution Order

To achieve optimal results, execute the scripts in the following order:

01_reset_processed_flag.py - Set the "Processed" flag to false for all documents in Cosmos DB.
02_send_telemetry_to_cosmosdb.py - Sends biometric/telemetry data to Cosmos DB.
03_stress_detection_and_recommendations.py - Contains the RandomForestClassifier model to detect stress levels and the recommendation algorithm to send responses to the IoT Hub for workers to access.

By following this order, you ensure that the system is properly set up and the RandomForest model can accurately detect stress levels and provide appropriate training recommendations.

Model Training and Predictions

The RandomForest model is initially trained using data with the "Processed" flag set to false. Once the model is trained, the system directly performs predictions on any new incoming data without updating the model. The "Processed" flag is not changed for the new data, and it is used only for making predictions.

To control whether the model should be retrained with new data or not, a variable named model_trained is introduced in the 03_stress_detection_and_recommendations.py script. This variable tracks whether the model has already been trained or not. Once the model has been trained using the initial data with the "Processed" flag set to false, the variable will be updated, and the model will not be retrained with the new incoming data.

Installation

Clone the repository:

git clone https://github.com/VSM97/Akkodis-IOT-StressLess.git

Install the required Python packages:
```
pip install -r requirements.txt
```

Set up the environment variables for the Azure IoT Hub and Azure Cosmos DB:

export DEVICE_CONNECTION_STRING=<your_device_connection_string>
export COSMOS_ENDPOINT=<your_cosmos_db_endpoint>
export COSMOS_KEY=<your_cosmos_db_key>
export COSMOS_DATABASE_NAME=<your_cosmos_db_database_name>
export COSMOS_CONTAINER_NAME=<your_cosmos_db_container_name>

Usage

If there is already data in the Cosmos DB, run the script to reset the "Processed" flag:
```
python 01_reset_processed_flag.py
```
Run the IoT device simulator script:
```
python 02_send_telemetry_to_cosmosdb.py
```

Run the main analysis and recommendation script:

python 03_stress_detection_and_recommendations.py

The system will start collecting data from the simulated IoT devices, analyze the data, and provide training recommendations based on the workers' physiological stress levels.

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Please make sure to update tests as appropriate.

License

This project is a property of Akkodis and its usage is prohibited without explicit permission. For inquiries and usage permissions, please contact Akkodis. Unauthorized use or distribution of this project may result in legal consequences.

VSM97/Akkodis-IoT-StressLess