The DSP Implementation Of ECG Signal Processing Algorithms For Wearable Physiological Signal Monitoring System

The monitoring of health status for old people and patients chronic disease has increasingly become an important healthcare issue around globe. Telemonitoring, compared with traditional in-hospital care, offers patients advantages such as low cost, more freedom and less interference with normal life, thus is a promising solution to above mentioned monitoring problem.This project introduces an innovative wearable remote physiological monitoring system. It is composed of two parts: a healthcare center and wearable terminals. Those two parts communicate with each other via GPRS network. In the terminal, an ARM controller is in charge of multi-channel physiological signal sampling (ECG, blood pressure, body temperature, etc), and a DSP processor is employed to conduct real-time data processing. The notable feature of this design include: firstly, GPRS wireless communication network is utilized as data exchange media, which overcomes the shortcomings of traditional wired and short distance Radio Frequency methods, providing wide signal coverage and stable communication channels; Secondly, the design is base on the concept of intelligent handle-hold device, which utilizes ARM and DSP dual processors to conduct local physiological signal processing. This strategy extends the function of the terminal from the traditional data sampling and transmission, to data processing and analysis, which is able to provide the user with real-time feedbacks and warnings.In this paper, the design and implementation of the ECG signal processing algorithm is discussed in detail. In this design, the DSP chip is the TMS320VC5509, manufactured by Texas Instrument Cooperation. This processor is featured as low power consumption and high computation capacity, which is very suitable in embedded application. The signal processing algorithm has the 3 main function modules: an R-wave detection algorithm based on wavelet transformation and threshold detection, a lossless data compression algorithm derived from Lz77 classical algorithm, and a data encryption algorithm based on AES method. Besides, discussion also includes data exchange strategy between ARM and DSP processors, the use of DSP/BIOS operation system to manage multi-task scheduling.The experimental results demonstrate that the designed algorithm is very efficient on the DSP. Processing results can be obtained within desired time and accuracy. This paper offers a successful example for customizing general DSP device to medical signal processing applications.