A Fast and Robust Deep Convolutional Neural Networks for Complex Human Activity Recognition Using Smartphone.

Wen Qi,Andrea Aliverti,Chenguang Yang,Giancarlo Ferrigno,Elena De Momi,Hang Su

doi:10.3390/s19173731

Abstract

As a significant role in healthcare and sports applications, human activity recognition (HAR) techniques are capable of monitoring humans’ daily behavior. It has spurred the demand for intelligent sensors and has been giving rise to the explosive growth of wearable and mobile devices. They provide the most availability of human activity data (big data). Powerful algorithms are required to analyze these heterogeneous and high-dimension streaming data efficiently. This paper proposes a novel fast and robust deep convolutional neural network structure (FR-DCNN) for human activity recognition (HAR) using a smartphone. It enhances the effectiveness and extends the information of the collected raw data from the inertial measurement unit (IMU) sensors by integrating a series of signal processing algorithms and a signal selection module. It enables a fast computational method for building the DCNN classifier by adding a data compression module. Experimental results on the sampled 12 complex activities dataset show that the proposed FR-DCNN model is the best method for fast computation and high accuracy recognition. The FR-DCNN model only needs 0.0029 s to predict activity in an online way with 95.27% accuracy. Meanwhile, it only takes 88 s (average) to establish the DCNN classifier on the compressed dataset with less precision loss 94.18%.

Highlights

With the widespread usage of portable and wearable smart devices, human activity recognition (HAR) becomes an active and comfortable research field
To improve the computational speed and classification accuracy, we propose a new fast and robust structure based on deep convolutional neural networks (DCNN), called FR-DCNN model
To evaluate the claims of the merits of the proposed FR-DCNN model in Section 1, we design several experiments to compare with the state-of-the-art deep learning (DL) algorithms and using different inputs

Summary

Introduction

With the widespread usage of portable and wearable smart devices, human activity recognition (HAR) becomes an active and comfortable research field. Continuous HAR systems are developed as part of a framework to monitor long-term human behaviors, such as ambient assisted the living, sports injury detection, and surveillance [2]. A HAR system is expected to report on people’s daily activities outside a hospital setting becomes an essential tool for healthcare interventions evaluation and clinical decision-making [3]. The wearable sensor technology can potentially facilitate many applications such as rehabilitation instruction, motion evaluation, activity reminder, and fall detection [4]. HAR helps bridge the gap between the low-level sensor and the high-level human-centric applications. Among the various available sensing components, the smartphone is widely used due to its low intrusiveness, convenience, and high adherence. Smartphones offer too much convenient for monitoring human’s physical and physiological

Objectives

Methods

Results

Conclusion