Multiphase Identification Algorithm for Fall Recording Systems Using a Single Wearable Inertial Sensor

A Deep Convolutional Neural Network-XGB for Direction and Severity Aware Fall Detection and Activity Recognition

Activity and Fall detection have been a topic of keen interest in the field of ambient assisted living system research. Such systems make use of different sensing mechanisms to monitor human motion and aim to ascertain the activity being performed for health monitoring and other purposes. Towards this end, in addition to activity recognition, fall detection is an especially important task as falls can lead to injuries and sometimes even death. This work presents a fall detection and activity recognition system that not only considers various activities of daily living but also considers detection of falls while taking into consideration the direction and severity. Inertial Measurement Unit (accelerometer and gyroscope) data from the SisFall dataset is first windowed into non-overlapping segments of duration 3 s. After suitable data augmentation, it is then passed on to a Convolutional Neural Network (CNN) for feature extraction with an eXtreme Gradient Boosting (XGB) last stage for classification into the various output classes. The experiments show that the gradient boosted CNN performs better than other comparable techniques, achieving an unweighted average recall of 88%.

A Hierarchical Approach to Activity Recognition and Fall Detection Using Wavelets and Adaptive Pooling

Human activity recognition has been a key study topic in the development of cyber physical systems and assisted living applications. In particular, inertial sensor based systems have become increasingly popular because they do not restrict users’ movement and are also relatively simple to implement compared to other approaches. In this paper, we present a hierarchical classification framework based on wavelets and adaptive pooling for activity recognition and fall detection predicting fall direction and severity. To accomplish this, windowed segments were extracted from each recording of inertial measurements from the SisFall dataset. A combination of wavelet based feature extraction and adaptive pooling was used before a classification framework was applied to determine the output class. Furthermore, tests were performed to determine the best observation window size and the sensor modality to use. Based on the experiments the best window size was found to be 3 s and the best sensor modality was found to be a combination of accelerometer and gyroscope measurements. These were used to perform activity recognition and fall detection with a resulting weighted F1 score of 94.67%. This framework is novel in terms of the approach to the human activity recognition and fall detection problem as it provides a scheme that is computationally less intensive while providing promising results and therefore can contribute to edge deployment of such systems.