Signal Processing in Portable Biopotential Acquisition Systems

Signal acquisition for biomedical applications is a challenging subject with increasing demands in fidelity. Modern biomedical systems require flexibility, portability and reliability of the measurement equipment. Implants and wearable modules are becoming the main focus as the future diagnostic tools. The miniaturization of high-precision equipment involves numerous important trade-offs that have to be addressed in order to meet the imposed requirements. Biopotential signal processing happens in low frequency bands and involves detection of various temporal and spectral features. Relatively small amplitude and inaccessible nature of the signal sources require advanced techniques in offset cancellation and noise filtering. With careful design considerations it is possible to take advantage of the properties of the signal and minimize the processing required to extract the vital information from a patient body in a most convenient way. This work presents a practical implementation of biomedical signal processing chain in embedded cardiogram monitoring system. The tasks of signal acquisition, postprocessing, and out-streaming the signal are thoroughly addressed. The analysis of signal processing techniques that can be applied to encephalogram readings was performed, and the custom wireless streaming protocol was tailored for use with sub-dermal implants. The data handling methods that were studied in both cases give new insights on the related hardware challenges