Search Results (2)

Stress has become a common mental health issue in modern society, causing individuals to experience acute behavioral changes. Exposure to prolonged stress without proper prevention and treatment may cause severe damage to one’s physiological and psychological health. Researchers around the world have been working to find and create solutions for early stress detection using machine learning (ML). This paper investigates the possibility of utilizing Tiny Machine Learning (TinyML) in developing a wearable device, comparable to a smartwatch, that is equipped with both physiological and psychological data detection system to enable edge computing and give immediate feedback for stress prediction. The main challenge of this study was to fit a trained ML model into the microcontroller’s limited memory without compromising the model’s accuracy. A TinyML-based framework using a Raspberry Pi Pico RP2040 on a customized board equipped with several health sensors was proposed to predict stress levels by utilizing accelerations, body temperature, heart rate, and electrodermal activity from a public health dataset. Moreover, a few selected machine learning models underwent hyperparameter tuning before a porting library was used to translate them from Python to C/C++ for deployment. This approach led to an optimized XGBoost model with 86.0% accuracy and only 1.12 MB in size, hence perfectly fitting into the 2 MB constraint of RP2040. The prediction of stress on the edge device was then tested and validated using a separate sub-dataset. This trained model on TinyML can also be used to obtain an immediate reading from the calibrated health sensors for real-time stress predictions. Full article

The importance of the usage of renewable energy sources in powering wireless sensor nodes in IoT and sensor networks grows together with the increasing number of utilized sensor nodes. Considering the other types of renewable energy sources, solar power differs as the most suitable one and emerges as the major source for powering sensor nodes. Thus, the consideration of using sensor nodes and collected sensor data for estimating solar panel performances and therefore solar power potential can improve the efforts in this direction. This paper presents the methodology for implementing edge intelligence on wireless sensor nodes for solar panel output voltage estimation and forecasting. The methodology covers the usage of the Python Scikit-learn package and micromlgen library for the implementation of edge intelligence on Arduino clone-based sensor nodes, particularly the development boards based on the ESP8266 chips. Scikit-learn is used for analyzing the efficiency of various regressors on collected solar data. The micromlgen library is then used for implementing those regressors on Arduino and clone nodes. The prediction of solar panel voltage generation is based on a single-sensor reading—UV or BH1750 light sensor. The Random Forest and Decision Tree regressors are implemented on the ESP8266-based development board—Wemos D1 R2. The estimation accuracy of the RF model is an MSE of approximately 0.10, MAE of 0.07 for UV and 0.04 for BH1750, and an R² of approximately 0.93 for both UV and BH1750 light sensors. The Decision Tree model has a lower accuracy with an MSE between 0.13 and 0.14, MAE of 0.07 for UV and 0.04 for BH1750, and R² of 0.90 and 0.89 for the UV and BH1750 sensors, respectively. The methodology and its efficiency are presented and discussed in this paper. Full article