Market assessment of biosignal quality measurement, enhancement, and noise-robust feature extraction from wearable devices based on modulation sprectrum processing

点击翻译按钮获取中文摘要

Advances in low-cost wearable health monitoring devices have opened doors for numerous new applications, including fitness tracking, stress monitoring, remote health monitoring and disease detection, peak athletic performance assessment, smart-home health monitoring of the elderly or epilepsy detection for children, to name a few. Wearable devices, however, typically rely on dry sensors, which are known to be highly contaminated by e.g., movement, sweat, and/or muscle contraction artifacts, as well as data transmission protocols not optimized for biosignals. Such sensitivity can lead to very noisy signals, which in turn, can lead to false alarms, thus hampering the performance of existing and emerging health applications. The present invention aims to overcome this limitation. More specifically, the invention proposes a new signal processing technique which we have shown to accurately separate signal and noise components, thus allowing us to (1) blindly measure the biosignal noise level, or signal-to-noise-ratio (SNR), (2) filter out only the noise components, thus accurately enhancing the noisy biosignal, and (3) by focusing only on the signal components of the data, to extract relevant parameters (also known as features), thus leading to noise-robust features for e.g., health monitoring, diagnosis, and abnormality detection. The technology has the potential of redefining the performance envelope of applications relying on wearables, ranging from health monitoring, to rehabilitation, human-machine interaction, and smart homes. The market assessment will help us further refine the technology by providing us with crucial information about the most relevant and competitive market(s) to pursue, based on the identified competitive advantages that the technology provides over existing tools.**