Enhancing Human Activity Recognition in Smart Homes with Self-Supervised Learning and Self-Attention

Abstract

Deep learning models have gained prominence in human activity recognition using ambient sensors, particularly for telemonitoring older adults’ daily activities in real-world scenarios. However, collecting large volumes of annotated sensor data presents a formidable challenge, given the time-consuming and costly nature of traditional manual annotation methods, especially for extensive projects. In response to this challenge, we propose a novel AttCLHAR model rooted in the self-supervised learning framework SimCLR and augmented with a self-attention mechanism. This model is designed for human activity recognition utilizing ambient sensor data, tailored explicitly for scenarios with limited or no annotations. AttCLHAR encompasses unsupervised pre-training and fine-tuning phases, sharing a common encoder module with two convolutional layers and a long short-term memory (LSTM) layer. The output is further connected to a self-attention layer, allowing the model to selectively focus on different input sequence segments. The incorporation of sharpness-aware minimization (SAM) aims to enhance model generalization by penalizing loss sharpness. The pre-training phase focuses on learning representative features from abundant unlabeled data, capturing both spatial and temporal dependencies in the sensor data. It facilitates the extraction of informative features for subsequent fine-tuning tasks. We extensively evaluated the AttCLHAR model using three CASAS smart home datasets (Aruba-1, Aruba-2, and Milan). We compared its performance against the SimCLR framework, SimCLR with SAM, and SimCLR with the self-attention layer. The experimental results demonstrate the superior performance of our approach, especially in semi-supervised and transfer learning scenarios. It outperforms existing models, marking a significant advancement in using self-supervised learning to extract valuable insights from unlabeled ambient sensor data in real-world environments.

AtomGID: An Atomic Gesture Identifier for Qualitative Spatial Reasoning

“…Our laboratory, the Laboratoire d'Intelligence Ambiante pour la Reconnaissance d'Activité (LIARA), has been at the forefront of research and development in the domain of smart homes aimed at facilitating aging in place for almost two decades [12,21,23,25,29]. Over the years, our research has focused on creating intelligent environments that seamlessly integrate pervasive sensing technologies with user-friendly interfaces to enhance the quality of life of older adults.…”

“…Moreover, these models often lack the flexibility to adapt to the diverse profiles and behaviors of individual residents [10]. Recognizing these limitations, a substantial portion of the research community, including our team, has shifted towards data-driven approaches [21]. However, in our specific context, the utilization of existing machine learning algorithms poses challenges.…”

“…On the contrary, unsupervised methods do not require a labeled dataset for the learning phase [20]. These models are scarce in the literature and can only recognize coarsely grained ADLs [21]. A good tradeoff is the recent semi-supervised learning models [22].…”

Self-Supervised Open-Set Speaker Recognition with Laguerre–Voronoi Descriptors

“…Microphones integrated with speaker recognition technology serve as intelligent sensors within the Internet of Things ecosystem, enabling them to respond selectively to commands provided by authorized users. Self-supervised learning (SSL) is the process of learning representation from unlabeled data [ 1 , 2 ]. For speaker recognition, SSL offers numerous advantages over supervised learning.…”