Biomedical Engineering Reference
In-Depth Information
SIGNAL PROCESSING
2.0
INTRODUCTION
All of the biomechanical variables are time-varying, and it doesn't matter
whether the measure is kinematic, kinetic, or EMG; it must be processed like
any other signal. Some of these variables are directly measured: acceleration
and force signals from transducers or EMG from bioamplifiers. Others are a
product of our analyses: moments-of-force, joint reaction forces, mechanical
energy and power. All can benefit from further signal processing to extract
cleaner or averaged waveforms, correlated to find similarities or differences
or even transformed into the frequency domain.
This chapter will summarize the analysis techniques associated with auto-
and cross-correlations, frequency (Fourier) analysis and its applications cor-
rect data record length and sampling frequency. The theory of digital filtering
is presented here; however, the specific applications of digital filtering of kine-
matics appears in Chapter 3 and analog filtering of EMG in Chapter 10. The
applications of ensemble averaging of variables associated with repetitive
movements are also presented.
2.1
AUTO- AND CROSS-CORRELATION ANALYSES
Autocorrelation analyzes how well a signal is correlated with itself, between
the present point in time and past and future points in time. Cross-correlation
Search WWH ::
Custom Search