Outlier Detection in Wearable Sensor Data for Human Activity Recognition (HAR) Based on DRNNs

Abstract

Wearable sensors provide a user-friendly and non-intrusive mechanism to extract user-relateddata that paves the way to the development of personalized applications. Within those applications, humanactivity recognition (HAR) plays an important role in the characterization of the user context. Outlierdetection methods focus on finding anomalous data samples that are likely to have been generated by adifferent mechanism. This paper combines outlier detection and HAR by introducing a novel algorithmthat is able both to detect information from secondary activities inside the main activity and to extract datasegments of a particular sub-activity from a different activity. Several machine learning algorithms havebeen previously used in the area of HAR based on the analysis of the time sequences generated by wearablesensors. Deep recurrent neural networks (DRNNs) have proven to be optimally adapted to the sequentialcharacteristics of wearable sensor data in previous studies. A DRNN-based algorithm is proposed in thispaper for outlier detection in HAR. The results are validated both for intra- and inter-subject cases and bothfor outlier detection and sub-activity recognition using two different datasets. A first dataset comprising4 major activities (walking, running, climbing up, and down) from 15 users is used to train and validatethe proposal. Intra-subject outlier detection is able to detect all major outliers in the walking activity in thisdataset, while inter-subject outlier detection only fails for one participant executing the activity in a peculiarway. Sub-activity detection has been validated by finding out and extracting walking segments present inthe other three activities in this dataset. A second dataset using four different users, a different setting anddifferent sensor devices is used to assess the generalization of results.