An Ensemble of Condition Based Classifiers for Device Independent Detailed Human Activity Recognition Using Smartphones †

Abstract

Human activity recognition is increasingly used for medical, surveillance and entertainment applications. For better monitoring, these applications require identification of detailed activity like sitting on chair/floor, brisk/slow walking, running, etc. This paper proposes a ubiquitous solution to detailed activity recognition through the use of smartphone sensors. Use of smartphones for activity recognition poses challenges such as device independence and various usage behavior in terms of where the smartphone is kept. Only a few works address one or more of these challenges. Consequently, in this paper, we present a detailed activity recognition framework for identifying both static and dynamic activities addressing the above-mentioned challenges. The framework supports cases where (i) dataset contains data from accelerometer; and the (ii) dataset contains data from both accelerometer and gyroscope sensor of smartphones. The framework forms an ensemble of the condition based classifiers to address the variance due to different hardware configuration and usage behavior in terms of where the smartphone is kept (right pants pocket, shirt pockets or right hand).The framework is implemented and tested on real data set collected from 10 users with five different device configurations. It is observed that, with our proposed approach, 94% recognition accuracy can be achieved.

Step by Step Towards Effective Human Activity Recognition: A Balance between Energy Consumption and Latency in Health and Wellbeing Applications

“…The first stage of HAR is data collection and filtering. The data collection process begins by defining a set of activities to be recognized, and then recording data from the sensors during a defined activity set [16], or simply taking over data from a publicly available HAR dataset [17].…”

“…Besides different sliding window sizes [64], some researches can also choose to introduce window overlapping [65]. The most commonly used fixed-size sliding window overlapping size is 1 s (used in [17,21,42,66]).…”

Classification of Human Daily Activities Using Ensemble Methods Based on Smartphone Inertial Sensors

“…These activities are classified based on feature extraction schemes that are broadly categorized as time and frequency domains. In [19,20,21,22,23,24,25] researchers have implemented the time domain and frequency domain feature extraction as a combined approach. Other researchers in [26,27,28,29,30,31] have used feature extraction in the time domain only, whilst, in [32] researchers applied the frequency domain and time-frequency domain.…”

“…The authors in [33] have chosen the time domain, frequency domain and time-frequency domain for feature extraction. Out of these studies, Saha et al [19] found that the ensemble classifier performs best, with an overall accuracy rate of 94% using accelerometer and gyroscope sensor data. In the research carried out by Mohamed et al [20], a combination of accelerometer data from the arm, belt and pocket analysed using rotation forest with the base learner C4.5, was found to provide the best overall classification accuracy rate of 98.9% [20].…”

Special Issue on Selected Papers from e-Health Pervasive Wireless Applications and Services 2017

“…The human activity recognition which is increasingly used for medical, surveillance, and entertainment applications is addressed by the second paper entitled "An Ensemble of Condition Based Classifiers for Device Independent Detailed Human Activity Recognition Using Smartphones", by Jayita Saha, Chandreyee Chowdhury, Ishan Roy Chowdhury, Suparna Biswas, and Nauman Aslam [3]. For better monitoring, human activity applications require identification of detailed activity like sitting on chair/floor, brisk/slow walking, running, etc.…”