Adaptive filtrate: signal processing in the measurement of haemoglobin using non-invasive method.

Abstract

This project is part of another project which is currently running at the University of Limerick (Ireland) and who is objective is to measure the concentration of haemoglobin (Hb) in blood using a non-invasive optical method. A sensor system, which basically consists of: four optical sources emitting light of different wavelengths; a photodiode which has a response in the appropriate wavelength range; a circuits for signal conditioning and low level processing of the information; and various software algorithms to interpret and calculate the Hb concentration. The main sensor elements are integrated into a finger clip. Despite advances in the development of monitoring systems for haemoglobin using non-invasive techniques, there is still a significant amount of research and development (R&D) required. One of the areas which requires further R&D is to minimise the unwanted effects of motion artefacts on the signal obtained from the photodiode. Through the development of hardware (e.g.: accelerometer, gyroscope, laser and monitor diode, etc.) and software (motion-resistant algorithms), it will gradually achieve the goal of highly accurate and immune to Motion Artefact. This project focuses on the software development. The main objective is to evaluate a number of methods which could potentially remove these motion artefacts through the use of adaptive filtering. One popular adaptive filtering techniques used in such situations is the noise canceller and this option will be explored in the project. In addition different adaptive algorithms will be compared in order to obtain the best solution. Simulations in Matlab and Labview will be used because the latter software is used extensively in the main Hb analysis programme. This project report has the following structure: - Introduction to non-invasive haemoglobin monitoring by LED based optical sensor system. - PPG signal features and motion artefact. - Theory related to adaptive filtering and adaptive algorithms. - Code and simulations in both Matlab and Labview.