Wavelet based approach for posture transition estimation using a waist worn accelerometer

Abstract

The ability to rise from a chair is considered to be important to achieve functional independence and quality of life. This sit-to-stand task is also a good indicator to assess condition of patients with chronic diseases. We developed a wavelet based algorithm for detecting and calculating the durations of sit-to-stand and stand-to-sit transitions from the signal vector magnitude of the measured acceleration signal. The algorithm was tested on waist worn accelerometer data collected from young subjects as well as geriatric patients. The test demonstrates that both transitions can be detected by using wavelet transformation applied to signal magnitude vector. Wavelet analysis produces an estimate of the transition pattern that can be used to calculate the transition duration that further gives clinically significant information on the patients condition. The method can be applied in a real life ambulatory monitoring system for assessing the condition of a patient living at home.

A Self Organizing Map Based Postural Transition Detection System

“…The signal from these sensors are analyzed by a wavelet-based pattern recognition algorithm in order to detect the postural transitions. This is fairly similar approach to the one previously presented in (Bidargaddi, 2007). Results of an experiment are also given to show a mean classification rate of 70% for this approach.…”

“…The experimental result shows that the detection correctnesses of Si-St and St-Si are 92.2 and 95.6%, respectively. In (Bidargaddi, 2007), a wavelet-based algorithm for detecting and calculating the durations of Si-St and St-Si transitions is developed and reported. The algorithm processes the signal vector magnitude of the measured acceleration signal.…”

Emerging Technology and Analytical Techniques for the Clinical Assessment of Scoliosis

“…Additional research within this area has focused on characterizing in greater detail the nature of the activities, with particular emphasis placed on the extraction of features related to the quality and quantity of walking [96], segmental accelerations [97,98], transitions in positioning and orientation [99], and trunk posture [100][101][102][103][104][105][106]. Within this context, there has been a limited focus of the application of this technology for the measurement of trunk postures in children and youth with Scoliosis.…”

Posture transition identification on PD patients through a SVM-based technique and a single waist-worn accelerometer

“…Several methods have been used in order to study PT, such as electromyography [9][10][11], goniometry [3,12], video [13], photography [14] and pressure platforms [15]. Since these systems rely on cumbersome, heavy or not wearable instruments, they cannot be used in ambulatory monitoring.…”

“…In this sense, numerous works have analyzed PT locating the sensor in different parts of the body. Bidargaggi located an inertial sensor at the waist in order to analyze Sit-to-Stand (SiSt) and Stand-to-Sit (StSi) transitions [14] while Najafi placed the inertial system at the chest [15]. A headband with an inertial system was used by Aloqlah et al for classifying human postures [16] and Bieber et al performed a SiSt and StSi classifier using a mobile phone within a trouser pocket [17].…”