A versatile design method of fast, linear-phase FIR filtering systems for electrocardiogram acquisition and analysis systems

Electrocardiogram signals, particularly those taken in ambulatory settings, are often corrupted by biological and electrical background noise. A fast finite impulse response (FIR) filtering system that can be easily implemented without a digital signal processing (DSP) processor based data-acquisition system can provide a solution to this problem. The filters developed were formed by cascading several moving average filters, and a short corrector filter. An algorithm is described that determines the best linear-phase corrector filter for any given cutoff frequency, sampling frequency, set of tap weights, and lengths of the moving average filters. At observed sampling rates of up to 500 Hz, the filtering systems consisting of the corrector and moving average filters were shorter in length and has less passband ripple than filters designed with the Parks-McClellan, Kaiser, and Hamming methods