Low-sampling-rate wireless sensing of human respiratory movement using UWB radio

Abstract

Rely on their high spatial resolution, robustness to interferences and low power density, UWB signals have recently attracted considerable researches for medical applications, such as contactless sensing of human respiration rate. In the conventional methodology, the UWB signals are acquired by an ultra high sampling rate device (oscilloscope), and there are many signal processing algorithms developed to retrieve human respiration rate from the sampled signals. The whole system brings high hardware cost and complexity. In this Final Year Project report, low-sampling rate algorithms based on Compress Sensing and Finite Rate of Innovation theories are proposed. Throughout series of calculations and simulation results in this report, the proposed strategies are capable to reduce the cost and complexity for the UWB human respiration rate sensing system while maintain its accuracy and robustness for practical applications. This report starts with brief review of UWB theory and conventional methodology of acquiring human respiration rate by UWB techniques. After that, the idea of Compress Sensing theory is discussed, followed by a basic measurement strategy of compress sampling. The concept and simulation results on deploying Pseudorandom Sequence and Waveform Matched Dictionary to improve the effectiveness of the basic measurement algorithm are proposed and evaluated. To realize the algorithm in real situation, explicit hardware implementation scheme is proposed, along with its limitations and modified techniques. The performance of the modified techniques is also discussed. Based on Finite Rate of Innovation theory, this report proposes two measurement methods, followed by comprehensive simulations and hardware implementation scheme discussion. Further research work on applications of the low sampling rate algorithms is recommended at last part of this report.