A Novel Human Activity Recognition Model

With the continuous improvement of the living standard, people have changed their concept of disease treatment to health management. However, most of the current health management software makes recommendations based on users' static information, with a low frequency of updates. The effect of targeted suggestions becomes weak with the passage of time, and it is hard for their recommendation effect to be satisfactory. Based on the use of smartphones for recognizing human activities on a real-time basis, a novel 'CNN+GRU' model is proposed in this paper†, utilizing both convolutional neural networks (CNNs) and gated recurrent units (GRUs). 'CNN+GRU' is able to extract the features in sensor data more accurately and improve the recognition speed. The proposed model was evaluated on a public data set, where it achieved an average accuracy of 91.27%, thus outperforming other models participating in the performance comparison experiments. In summary, the 'CNN+GRU' model can effectively recognize mobile users' activities based on the sensor data collected by their smartphones.