David Duanne Rowlands

Investigating the translational and rotational motion of the swing using accelerometers for athlete skill assessment

In this paper, an accelerometer measurement system comprising three accelerometer nodes was used to identify the correlation between the skill level and the characteristics of the first serve swing in tennis. Three MEMS accelerometers were mounted on the knee, leg, and wrist of the tennis players. The kinematic model for the first serve was observed. Furthermore, this study revealed that side-forward motion of the hand along with the forward motion of the waist of an athlete can be used as indicator to assess the athletes skill level. It is envisaged that this application can provide feedback to tennis players.

Accelerometers: An underutilized resource in sports monitoring

Play based sports monitoring techniques provide coaches and players with the tools to better measure the effects of training or live performance. This paper explores the advantages of using accelerometers units, in an effort to better analyse over ground running in professional athletes. A large portion of studies in player monitoring in the Australian Football League (AFL) utilize GPS to obtain time and distance measurements.

Wavelet coherence of EEG signals for a visual oddball task

Neural co-activation in frontal and central cortex was examined during a visual oddball task using wavelet coherence. EEG was recorded during a visual oddball task, presented to 12 participants with a random mix of 15% oddball targets and 85% frequent non-target letters over 265 trials. Wavelet coherence of individual trials was shown to distinguish frequent and oddball trials. Averaged wavelet coherence showed significant differences: oddball targets showed higher delta-theta activity whereas frequent background stimuli showed higher gamma activity. Increased gamma coherence appeared to be related to expectation of the targets with our analysis showing an R(2) of 0.935 for the relationship between averaged sections of gamma coherence and the number of intervening (frequent) trials since the last oddball.

Automated ECG diagnostic P-wave analysis using wavelets

P-wave characteristics in the human ECG are an important source of information in the diagnosis of atrial conduction pathology. However, diagnosis by visual inspection is a difficult task since the P-wave is relatively small and noise masking is often present. This paper introduces novel wavelet characteristics derived from the continuous wavelet transform (CWT) which are shown to be potentially effective discriminators in an automated diagnostic process. Characteristics of the 12-lead ECG P-wave were derived using CWT and statistical methods. A normal control group and an abnormal (atrial conduction pathology) group were compared. The wavelet characteristics captured frequency, magnitude and variance components of the P-wave. The best individual characteristics (i.e. ones that significantly discriminated the groups) were entered into a linear discriminant analysis (LDA) for four different models: two-lead ECG, three-lead ECG, a derived three-lead ECG and a factor analysis solution consisting of wavelet characteristic loadings on the factors. A comparison was also made between wavelet characteristics derived form individual P-waves verses wavelet characteristics derived from a signal-averaged P-wave for each participant. These wavelet models were also compared to standard cardiological measures of duration, terminal force and duration divided by the PR segment. Results for the individual P-wave approach generally outperformed the standard cardiological measures and the signal-averaged P-wave approach. The best wavelet model on the basis of both classification performance and simplicity was the two-lead model that uses leads II and V1. It was concluded that the wavelet approach of automating classification is worth pursuing with larger samples to validate and extend the present study.

Visualization of wearable sensor data during swimming for performance analysis

Sensor-based biomechanical monitoring of sporting activity requires the interpretation of large data-sets of time series data-sets. Visualization techniques are a powerful method for displaying these data in a meaningful way to assist in understanding the complex interrelationships of the data and biomechanics. In particular, repetitive actions such as seen in many sports, including swimming can benefit from such analysis where overlay and visual comparison of multiple strokes can be advantageous. Many other disciplines, such as medicine visualize repetitive data and are translational opportunities for the investigation of biomechanical data, such as swimming. This paper presents a case study in which inertial sensor time series data from an elite and sub-elite swimmer were compared using visualization techniques to highlight differences in their action and performance. In particular, the metrics of body roll velocity was captured from the gyroscope sensor and was used as the key time series data to be visualized. Visualization techniques investigated were time-series overlay, phase space portraits, ribbon plot overlay, and wavelet scalograms. The phase space portraits, ribbon plots, and wavelet scalograms demonstrated clearly self-consistency of the swimmers action. As a cross-comparison tool, these techniques showed clear difference between the elite swimmer, who had lower variability and thus a more consistent action than the sub-elite swimmer. This paper has demonstrated that there is merit in further examination of these techniques as a tool for feedback. It was found that all the methods presented unique views of stroke biomechanics in a nontechnical yet intuitive way for clearer communication.

FPGA implementation of Morlet continuous wavelet transform for EEG analysis

This article presents the design and implementation of continuous wavelet transform (CWT) of nonstationary Electroencephalogram (EEG) signals using a Spartan 3AN FPGA. The widely applied Morlet wavelet function was used for obtaining the CWT coefficients. The complex convolutions were executed in Fourier space using simple multipliers. Altium designer was used to import Xilinx FFT core configured in Radix 4. Two VHDL controllers were built to control the FFT core operation (which handles both the FFT-IFFT computations) for the first controller; and the multiplication processes in between for the second one. The results showed that the digital architecture of Morlet wavelet function in Fourier space is very time efficient. By an optimized trade-off between speed and silicon area, the design can produce the wavelet coefficients at all scales of 1024 points EEG signal in approximately 1 msec when it runs at maximum clock speed of 125 MHz.

Cloud based activity monitoring system for health and sport

This paper gives the concept, design, and implementation of an activity monitoring system that incorporates a database and a data analysis language as an integral part of the structure. The versatility of the design allows many different analysis techniques to be run on the extracted data. This forms the framework to allow different machine learning techniques to be applied to the data without the construction of separate dedicated systems. As an example application, this paper applies the system to determine some key features of a running based activity.

Design and fabrication of an ECG amplifier on silicon using standard CMOS process

The ECG is a vital part in the armory of the fight against heart disease. In this work we have designed and implemented a 6 lead ECG acquisition system on silicon. The total circuit consisting of instrumentation amplifiers and filters were placed them onto a silicon die 870 /spl mu/m by 500 /spl mu/m. The circuit shows good noise immunity, very high CMRR, and a good frequency response. The output of the circuit is a 0 to 4 volt signal for lead I and lead II. The advantage of this system is the use of the standard CMOS process which will reduce the complexity and cost of manufacture.

Internet based ECG medical information system

Physiological monitoring of humans for medical applications is well established and ready to be adapted to the Internet. This paper describes the implementation of a Medical Information System (MIS-ECG system) incorporating an Internet based ECG acquisition device. Traditionally clinical monitoring of ECG is largely a labour intensive process with data being typically stored on paper. Until recently, ECG monitoring applications have also been constrained somewhat by the size of the equipment required. Today’s technology enables large and fixed hospital monitoring systems to be replaced by small portable devices. With an increasing emphasis on health management a truly integrated information system for the acquisition, analysis, patient particulars and archiving is now a realistic possibility. This paper describes recent Internet and technological advances and presents the design and testing of the MIS-ECG system that utilises these advances.

Derivation of a nonlinear variance equation and its application to SOI technology

An analytic nonlinear equation for variance was derived along with a method based on response surface mapping techniques to calculate the variance using the proposed equation. The technique was applied to the threshold voltage of a 0.1-/spl mu/m silicon-on-insulator MOS device, and the variance value obtained was verified using Monte Carlo simulation. The threshold voltage dependence upon active-layer thickness was found to be highly nonlinear due to the devices going from the fully depleted to the partially depleted regime. Analysis of the variance showed that the effect of the nonlinear terms (18.7%) is more important than the effect of the mixed term (-0.7%) and almost as important as the contribution of the second most dominant input-process parameter (23.6%). This illustrates the importance of the proposed nonlinear equation.