Time Series Analysis in Biomechanics
The quantification of human motion is based on the conversion of a continuous biomechanical signal to a discrete sampling of data points that vary with time known as a time series. This process enables the biomechanical signal to be stored digitally, subsequently allowing the various time series to be further processed. The goal of a biomechanical analysis is straightforward but the theory and techniques for obtaining accurate and meaningful data are often not. For example, if the biomechanical signal varies with time, how often do we need to sample it to prevent missing information? How can we insure that the sampled or digital signal accurately reflects the original biomechanical signal? Can we improve this accuracy after the signal has been digitized? Can we decompose the signal into constituent parts and examine these parts independently? Can we remove unwanted and distracting parts while retaining the vital information contained in the data? This chapter will cover the basic requirements to collect valid data free from sources of error and noise and some of the common processing and analysis techniques used for decomposing, rejecting, and recomposing the biomechanical signal.
KeywordsSignal Cross-correlation Autocorrelation Sampling theory Frequency Fourier transform Nyquist rate Nyquist Frequency Numerical differentiation Numerical integration First central difference method Trapezoid rule Noise Smoothing Filter Joint time frequency analysis Short term Fourier transform Wavelet transform Continuous wavelet transform Discrete wavelet transform
- Winter DA (1990) Biomechanics and motor control of human movement, 2nd edn. John Wiley & Sons, TorontoGoogle Scholar