This February 2021 report in the Sensors Journal deals with human activity recognition (HAR) using smartphone sensor data. To explore this strategy, a generic framework is proposed based Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) networks. The experimental results indicate that the proposed CNN-LSTM network performs well in activity recognition, enhancing the average accuracy by up to 2.24% as compared to prior state-of-the-art approaches.
Keywords: Architecture/Engineering, Association, Audio/Video, Communications, Controls/Sensors, Intelligent Building, Internet of Things (IoT), Cybersecurity/Privacy, Protocols/Standards, Smart Cities, Training, Voice & Speech Recognition, activity recognition