Abstract:Human activity recognition (HAR) technology based on wearables has received increasing attention in recent years. The traditional methods have used hand-crafted features to recognize human activities, resulting in shallow feature extraction. With the development of deep learning, an increasing number of researchers have focused on studying deep learning methods. To achieve higher recognition accuracy, the majority of the current HAR research involves multi-source and multi-modal sensors (MMS) data. However, due to the limitations in the receptive fields of single-dimensional convolutional kernels, these networks are still infeasible for extracting spatiotemporal features. In this study, a multi-dimensional parallel convolutional connected (MPCC) deep learning network based on MMS data for HAR is proposed that fully utilizes the advantages of multi-dimensional convolutional kernels. Moreover, multi-scale residual convolutional squeeze-and-excitation (MRCSE) modules are proposed to enrich the diversity of feature information by combining squeeze-and-excitation (SE) blocks. A daily home activity (DHA) dataset is constructed based on the requirements for HAR in certain scenarios, such as smart home, and we conduct experiments on the optimal combination of sensor locations on the DHA dataset according to a weighted F1 (FW)-score. Both 10-fold and leave-one-subject-out (LOSO) cross-validations (CVs) are used to evaluate the performance of the proposed network. The MPCC-MRCSE network achieves FW-scores of 98.33% and 95.42% on the physical activity monitoring for aging people (PAMAP2) and the OPPORTUNITY datasets using 10-fold CVs, respectively, and achieves FW-scores of 81.47% on the PAMAP2 when applying a LOSO CV.

Multi-source Transfer Learning for Human Activity Recognition in Smart Homes

Transferable Representation Learning for Human Activity Recognition in Smart Homes

A holistic multi-source transfer learning approach using wearable sensors for personalized daily activity recognition

Human Activity Recognition Based on Transformer in Smart Home.

Transfer Learning for Human Activity Recognition using Representational Analysis of Neural Networks

Privacy-preserving Activity Recognition Using Multimodal Sensors in Smart Office

Source Domain Selection for Cross-House Human Activity Recognition with Ambient Sensors.

Transfer Learning in Human Activity Recognition: A Survey

Human activity recognition based on multienvironment sensor data

Cross-Domain HAR: Few Shot Transfer Learning for Human Activity Recognition

Multi-User Activity Recognition In A Smart Home

A Survey of User-Centred Approaches for Smart Home Transfer Learning and New User Home Automation Adaptation

Human Activity Recognition (HAR) in Smart Homes

Know Thy Neighbors: A Graph Based Approach for Effective Sensor-Based Human Activity Recognition in Smart Homes

Optimal search mapping among sensors in heterogeneous smart homes

A Multi-dimensional Parallel Convolutional Connected Network Based on Multi-source and Multi-modal Sensor Data for Human Activity Recognition

Multi-resolution Fusion Convolutional Network for Open Set Human Activity Recognition

Ftmar: A Fusion Transformer Network for Multi-Resident Activity Recognition

Layout Agnostic Human Activity Recognition in Smart Homes through Textual Descriptions Of Sensor Triggers (TDOST)

Transfer Learning Model Knowledge Across Multi-sensors Locations over Body Sensor Network

Using Graphs to Perform Effective Sensor-Based Human Activity Recognition in Smart Homes