Alexandros Karagiannis

Noise-Assisted Data Processing With Empirical Mode Decomposition in Biomedical Signals

In this paper, a methodology is described in order to investigate the performance of empirical mode decomposition (EMD) in biomedical signals, and especially in the case of electrocardiogram (ECG). Synthetic ECG signals corrupted with white Gaussian noise are employed and time series of various lengths are processed with EMD in order to extract the intrinsic mode functions (IMFs). A statistical significance test is implemented for the identification of IMFs with high-level noise components and their exclusion from denoising procedures. Simulation campaign results reveal that a decrease of processing time is accomplished with the introduction of preprocessing stage, prior to the application of EMD in biomedical time series. Furthermore, the variation in the number of IMFs according to the type of the preprocessing stage is studied as a function of SNR and time-series length. The application of the methodology in MIT-BIH ECG records is also presented in order to verify the findings in real ECG signals.

Noise components identification in biomedical signals based on Empirical Mode Decomposition

Hilbert-Huang Transform (HHT) is composed of the Empirical Mode Decomposition (EMD) as the first step of the procedure and Hilbert Spectral analysis (HSA) as the second step. It is a recent tool in the analysis of signals originating from nonlinear processes as well as nonstationary signals. Empirical Mode Decomposition produces a set of Intrinsic Mode Functions and the core idea is based on the assumption that any data consists of different simple intrinsic modes of oscillations. Statistical significance of the Intrinsic Mode Functions and partial signal reconstruction are investigated in this paper. Application of Hilbert-Huang Transform on biomedical signals such as ECG from MIT-BIH database and experimental respiratory signals acquired by means of accelerometers, reveal the adaptive nature of the method.

Experimental respiratory signal analysis based on Empirical Mode Decomposition

Respiration is a widely used biosignal which is combined with other biosignals in order to extract information about the physiological or pathological conditions that may occur in the development of a treatment. Acquisition of respiration in a clinical environment is usually accomplished by standard hospital equipment and minimum invasive techniques. In this paper a non invasive technique is used for respiration monitoring based on accelerometers. The acquired signal is sampled and transmitted through a wireless sensor network to the gateway point (sink) where it is processed. Empirical Mode Decomposition (EMD) is considered as a method of processing of biosignals such as respiration and the application of the decomposition method in experimental signals acquired by means of a wireless sensor network is evaluated. The processing technique covered in this paper is based on selecting the appropriate signals (IMF) in which respiration is decomposed, by their spectral characteristics that correspond to respiration.

Energy consumption measurement and analysis in wireless sensor networks for biomedical applications

Energy Consumption in Wireless Sensor Networks is a fundamental issue in terms of functionality and network lifetime. Minimization of energy consumption by applying optimization techniques enables pervasive computing especially in the field of biomedical engineering. A framework for energy consumption measurement and analysis is proposed which combines a theoretical approach, a simulation procedure based on a widely used software simulator and the validation by means of a high sensitivity measurement setup. Application driven profiling of energy consumption at the node level is a useful tool for optimal function of energy consumable node components in order to improve total energy efficiency.

A prediction model for the number of intrinsic mode functions in biomedical signals: The case of electrocardiogram

Abstract In this paper, the open issue of prediction of the number of intrinsic mode functions (IMF) extracted from a time series after the application of empirical mode decomposition (EMD) is addressed and a methodology is presented directing towards an a priori prediction model. Parameters related with the time series are measured or calculated and used by the model in order to define a closed set in which the actual total number of IMFs is expected to be included after the application of EMD. The prediction model is verified by a large number of tests on simulated electrocardiogram (ECG) time series and after refinement it is validated using real ECG time series from Physionet MIT-BIH database.

Comparative study of Empirical Mode Decomposition applied in experimental biosignals

Electrocardiogram is a widely used biosignal for diagnosis of pathological situation concerning heart function. Interpretation and analysis of this signal is critical for the selection of the appropriate treatment and this highlights the necessity for clean signals without any kind of artifacts. Several methods have been developed in order to remove artifacts and delineate the characteristics of the ECG that physicians need to evaluate. In this paper, Empirical Mode Decomposition (EMD) is considered and the application of the decomposition method in experimental signals acquired by means of a wireless sensor network is evaluated. The proposed technique is based on the EMD and is studied comparatively to classic techniques of signal processing. Certain metrics are implemented to evaluate the performance of the technique and the results show good results of the EMD based method.

Energy consumption analysis and optimization techniques for Wireless Sensor Networks

WirelessSensorNetworksfunctionalityisclosely relatedtonetworklifetimewhichdependsontheenergy consumptionatthendelevelaswellasthenetworklevel. Analysisoftheenergyconsumptioncontributionofeachnode componenttothetotalenergyconsumptionofanodethrougha meaurementapproachwiththeoreticalformulationofa performancemetricandsimulationresultsfromawidelyused simulatorverifthatoptimizationtechniquesaimedatthe optimaluseoftheradiocommunicationssubsystemcould substantiallyimprovetotalnergyefficiency.

E-SPONDER system: A new communication infrastructure for future emergency networks

This paper describes a novel emergency communication network architecture implemented within the FP7 EU project E-SPONDER [1]. It is characterized by the deployment of heterogeneous wireless systems and also by its holistic approach achieving reliability, high performance, reconfigurability and standalone operation. It is a complete suite of real-time communication technologies built to support the first responders [1] with information services during disaster events. This work investigates the system architecture in an aircraft landing incident and describes a field test carried on at Schiphol airport in Amsterdam. More specifically, a scalable and adaptive telecommunication architecture that ensures voice, video and data between first responders and command centers at all times, even under extreme conditions, is presented. The structure and functionalities of the VoIP subsystem that operates above the proposed heterogeneous E-SPONDER network architecture is described, with a detailed scenario analysis. Finally, the paper presents how the recommended solutions are integrated into an implementable platform.

Pervasive Homecare Monitoring Technologies and Applications

Research studies and population statistics records indicate that elderly population increases in the developed countries whilst forecasts reveal that 65 and over age group will be nearly 20% of the overall population. One of the major challenges related to this observation is the delivery of homecare and the reduction of healthcare cost without compromising the quality of services provided. Pervasive homecare systems provide information and mechanisms to alert when pathological situations are identified. The implementation of technological solutions for homecare applications minimizes the time to provide help in abnormal situations and improve quality of life in elderly and chronically ills. Pervasive homecare networks provide continuous medical monitoring, control of home appliances, medical data storage and processing and emergency situations awareness. Constant monitoring provides early detection of emergency conditions and assists in optimum scheduling of a wide range of healthcare services for people with various degrees of cognitive and physical disabilities. Researchers continuously explore technological solutions in order to provide homecare and healthcare services for the elderly, chronically ill and children. The amount of proposals reveal a rapidly growing scientific area of high impact on sensitive groups that significantly improve quality of life and prevent or deal with life threatening situations. However, the use of this wireless sensor technology in medical practice not only allows a supreme level of complexity in patient monitoring with regards to existing parameters (such as vital signs), but also offers the prospect of identifying new ways of diagnosing and preventing disease. Although wired communication technologies, such us ATM (Asynchronous Transfer Mode) and optical communication, are widely used, the key aspects for pervasive healthcare communication is the transfer of high-speed and ubiquitous health data in every place in earth securely and promptly. Wireless technology came to encompass the e-health monitoring everywhere from any given location, providing the so-called m-health services. Research and development advances in the e-health community include data gathering and transfer of vital information, integration of human machine interface technology into handheld devices, data interoperability and integration with hospital legacy systems and electronic patient records.

Computation time study in biomedical signal processing with Empirical Mode Decomposition: The case of electrocardiogram

In this paper, a study of the Empirical Mode Decomposition (EMD) performance is presented in terms of computation time. Smart resource allocation and management in embedded systems are facilitated by signal processing techniques modeling for time scheduling of tasks. Empirical Mode Decomposition computation time is mainly determined by the number of iterations and the size of Intrinsic Mode Functions (IMF) set which are unknown at the beginning of the process. A metric is introduced to include these factors into a single variable of a linear model developed to a priori estimate methods computation time. In the same framework of Empirical Mode Decomposition computation time study the effects of noisy components and the application of preprocessing techniques are evaluated.