No abstract

In this paper, a discrete wavelet transform (DWT) based de-noising with its applications into the noise reduction for medical signal preprocessing is introduced. This work focuses on the hardware realization of a real-time wavelet de-noising procedure. The proposed de-noising circuit mainly consists of three modules: a DWT, a thresholding, and an inverse DWT (IDWT) modular circuits. We also proposed a novel adaptive thresholding scheme and incorporated it into our wavelet de-noising procedure. Performance was then evaluated on both the architectural designs of the software and. In addition, the de-noising circuit was also implemented by downloading the Verilog codes to a field programmable gate array (FPGA) based platform so that its ability in noise reduction may be further validated in actual practice. Simulation experiment results produced by applying a set of simulated noise-contaminated electrocardiogram (ECG) signals into the de-noising circuit showed that the circuit could not only desirably meet the requirement of real-time processing, but also achieve satisfactory performance for noise reduction, while the sharp features of the ECG signals can be well preserved. The proposed de-noising circuit was further synthesized using the Synopsys Design Compiler with an Artisan Taiwan Semiconductor Manufacturing Company (TSMC, Hsinchu, Taiwan) 40 nm standard cell library. The integrated circuit (IC) synthesis simulation results showed that the proposed design can achieve a clock frequency of 200 MHz and the power consumption was only 17.4 mW, when operated at 200 MHz.

A real-time muscle fatigue monitoring system was developed to quantitatively detect the muscle fatigue of subjects during cycling movement, where a fatigue progression measure (FPM) was built-in. During the cycling movement, the electromyogram (EMG) signals of the vastus lateralis and gastrocnemius muscles in one leg as well as cycling speed are synchronously measured in a real-time fashion. In addition, the heart rate (HR) and the Borg rating of perceived exertion scale value are recorded per minute. Using the EMG signals, the electrical activity and median frequency (MF) are calculated per cycle. Moreover, the updated FPM, based on the percentage of reduced MF counts during cycling movement, is calculated to measure the onset time and the progressive process of muscle fatigue. To demonstrate the performance of our system, five young healthy subjects were recruited. Each subject was asked to maintain a fixed speed of 60 RPM, as best he/she could, under a constant load during the pedaling. When the speed reached 20 RPM or the HR reached the maximal training HR, the experiment was then terminated immediately. The experimental results show that the proposed system may provide an on-line fatigue monitoring and analysis for the lower extremity muscles during cycling movement.

No abstract