Empirical Studies on Wearable WSNs in Human Activity Recognition

Our research study, "Wearable WSNs in Human Activity Recognition," describes a complete method for detecting human activity using wireless sensor networks and wearable devices. Gathering and processing data, extracting characteristics, and applying them are three techniques to identify movement. We test with several WSN-equipped persons. Our portable WSNs include accelerometers, gyroscopes, and heart rate monitors to collect and analyze data. Continuously filtering and altering raw sensor data prepares it for usage. Feature extraction is essential to interpret mixed sensor data. No more MAD or Fast Fourier Transform. We identify human behavior using Random Forest, K-NN, and LSTM Network techniques. Our WSNHAR approach outperforms others in "Comparison of Recognition Metrics" and "Computational Efficiency and Energy Consumption" charts. Memory, F1-score, precision, accuracy, and processing time indicate system speed. The recommended approach computes swiftly and uses little power, making it suitable for energy-efficient programs.