Adaptive Activity Recognition with Dynamic Heterogeneous Sensor Fusion

Abstract

In spite of extensive research in the last decade, activity recognition still faces many challenges for real-world applications. On one hand, when attempting to recognize various activities, different sensors play different on different activity classes. This heterogeneity raises the necessity of learning the optimal combination of sensor modalities for each activity. On the other hand, users may consistently or occasionally annotate activities. To boost recognition accuracy, we need to incorporate the user input and incrementally adjust the model. To tackle these challenges, we propose an adaptive activity recognition with dynamic heterogeneous sensor fusion framework. We dynamically fuse various modalities to characterize different activities. The model is consistently updated upon arrival of newly labeled data. To evaluate the effectiveness of the proposed framework, we incorporate it into popular feature transformation algorithms, e.g., Linear Discriminant Analysis, Marginal Fisher's Analysis, and Maximum Mutual Information in the proposed framework. Finally, we carry out experiments on a real-world dataset collected over two weeks. The result demonstrates the practical implication of our framework and its advantage over existing approaches.

Multi-sensor fusion in body sensor networks: State-of-the-art and research challenges

“…Next step is to select the optimal set of features using feature selection methods such as windowing technique, kernel discriminant analysis [122], minimal redundancy maximal relevance heuristic [126], Correlation based Feature Selection [125,121]. Afterwards, the selected features are fed into supervised classifiers including Naive Bayes [122], SVM [124,126], Decision Trees [121], Neural Networks [15], k-Nearest Neighbor [125] or unsupervised clustering methods such as LDA, MFA [127] depending on the availability of the labels for each class of activity to detect the activities including static physical activities like sitting, standing, lying, watching TV [122,15], and dynamic like walking and up stairs [122], down stairs, running [125], ramp up and down [121]. Table 6 summarizes our analysis of the literature on feature-level sensor fusion in activity recognition systems; for the sake of clarity, we reported only the most representative (in terms of citations and novelty) analyzed works.…”

AttriNet

“…In the field of mobile, wearable, and pervasive computing, extensive research has been conducted to recognize human activities [2,4,7,8,[23][24][25][31][32][33][34][35]. One line of research in this field starts with Bao et al [2], who placed accelerometers on different body positions to recognize daily activities such as "walking," "sitting," and "watching TV."…”

A Big Network Traffic Data Fusion Approach Based on Fisher and Deep Auto-Encoder

“…A few examples which introduce Fisher into data fusion are as follows. Zeng et al [20] proposed a sensor fusion framework based adaptive activity recognition and dynamic heterogeneous, and they incorporated it into popular feature transformation algorithms, e.g., marginal Fisher's analysis, and maximum mutual information in the proposed framework. Chen et al [21] introduced the finite mixture of Von Mises-Fisher (VMF) distribution for observations that are invariant to actions of a spherical symmetry group.…”