A Deep Learning Approach to on-Node Sensor Data Analytics for Mobile or Wearable Devices

Abstract

The increasing popularity of wearable devices in recent years means that a diverse range of physiological and functional data can now be captured continuously for applications in sports, wellbeing, and healthcare. This wealth of information requires efficient methods of classification and analysis where deep learning is a promising technique for large-scale data analytics. While deep learning has been successful in implementations that utilize high-performance computing platforms, its use on low-power wearable devices is limited by resource constraints. In this paper, we propose a deep learning methodology, which combines features learned from inertial sensor data together with complementary information from a set of shallow features to enable accurate and real-time activity classification. The design of this combined method aims to overcome some of the limitations present in a typical deep learning framework where on-node computation is required. To optimize the proposed method for real-time on-node computation, spectral domain preprocessing is used before the data are passed onto the deep learning framework. The classification accuracy of our proposed deep learning approach is evaluated against state-of-the-art methods using both laboratory and real world activity datasets. Our results show the validity of the approach on different human activity datasets, outperforming other methods, including the two methods used within our combined pipeline. We also demonstrate that the computation times for the proposed method are consistent with the constraints of real-time on-node processing on smartphones and a wearable sensor platform.

Multi-Channels LSTM Networks for Fence Activity Classification

“…We believe that this model could be improved, for example combine raw data and man-made features, [18]. In real world scenario, inertial sensors are used with Wireless Sensor Network (WSN), [11]- [13].…”

Cyber–Physiochemical Interfaces

“…The healthcare data such as EEG, ECG, and EMG, are often difficult to mine effectively by using traditional data analysis approaches, due to the long‐term time dependency, varying length, and irregular sampling of the sensor data . Enabled by automated feature extraction and end‐to‐end learning, various deep learning algorithms have been developed to process the time‐series sensor data (Figure b–f) . In particular, RNNs with varying length sequences and long‐range dependencies on training data are good at extracting knowledge from time‐series data.…”

“…b) Schematic workflow of deep learning for time‐series data. Reproduced with permission . Copyright 2017, The Authors.…”

“…Although deep learning provides a promising approach in analyzing these complex sensor data, most current deep learning algorithms cannot run directly in the end of sensor devices . Instead, deep learning is usually implemented on high performance servers or personal computers, as there are thousands and millions internal parameters need to be adjusted in the model, which is difficult to run in the sensor end .…”

Learning analytics for IoE based educational model using deep learning techniques: architecture, challenges and applications

“…Mirella Atherton et al (Tempelaar et al, 2018) conducted the study aimed at finding the correlation between the online study approach and its positive effects on attainments of the learners. Dirk Tempelar et al (Ravi et al, 2017) provided an empirical study aimed at showcasing how "learning decomposition data" can be used to deliver (Walker, 2012) proposed a methodology using the deep learning techniques to capture sensor data and perform analysis to predict real time activity classification. Furthermore in order to refine the approach, spectral domain pre-processing was incorporated in their approach.…”