A Single Accelerometer-based Robust Human Activity Recognition via Wavelet Features and Ensemble Feature Selection

Human activity recognition (HAR) based on sensors has been widely used in many fields. Instead of using multi-sensor system which is not convenient in practical applications and requires high computational cost, this paper utilizes a single wearable accelerometer to collect human activity information. In order to improve the recognition performance of the whole system and select the features that are most relevant to the wearing position of sensor, the wavelet decomposition-based features and a novel feature selection method are introduced. Considering the limitation of single filter feature selection method, this paper proposes an ensemble-based filter feature selection (EFFS) approach to optimize the feature set. Experiment results show that the wavelet decomposition-based features can increase the discrimination of activities and significantly and improve the activity recognition accuracy. Compared with other four popular feature selection methods, the proposed EFFS approach provides higher accuracy with fewer features.