Signals

Question 1

limitations of signals are based on what? (3)

Answer 1

• sensor error
• sampling
• environment

Question 2

4 ways to deal with error

Answer 2

• filtering data
• sampling properly
• smoothing data
• understanding signals and transformations that we intend to perform on them

Question 3

explain the nyquist shannon sampling theorem

Answer 3

• taking frequent samples of an analog signal to digitize it
• for the digital signal to represent the analog signal perfectly you must increase the sampling frequency to at least TWICE the analog signal highest frequency

Question 4

with the nyquist shannon sampling theorem, to ensure sharp directional changes you must have a frequency that is?

Answer 4

5-10 times faster then the analogs signals frequency

Question 5

raises the level of the signal relative to the noise

Answer 5

amplification

Question 6

name and explain 2 smoothing techniques

Answer 6

1. curve fitting: fits a best fitting function to the data
   - -> polynomials
   - -> splines
2. moving averages: over a window of samples, simple averaging gets rid of noisy spikes

Question 7

4 essential components of a signal

Answer 7

1. frequency: # of repeated cycles per unit of time
2. amplitude: magnitude of change
3. offset: distance of mean amp from zero (above or below)
4. phase shift: a horizontal offset

Question 8

when a signal is decomposed it can then be represented in a ?
–> then if we multiplied each individual sine wave by each other what would we get?

Answer 8

frequency graph of sine waves with only the stuff we want to keep (all the noise cut out)
- the original signal time domain graph

Question 9

the time domain graph shows what?

and the frequency domain graph shows what?

Answer 9

• amplitude for each frequency throughout time

- average amplitude at each frequency in a bar like fashion

Question 10

a time domain signal is transformed into a frequency domain signal using a mathematical technique called?

Answer 10

fourier transform (decomposing the signal into components) 
--> find the frequency component of a time- domain signal that is mixed with noise

Question 11

4 types of digital filtering

Answer 11

1. low pass: let everything below pass through (record)
2. high pass: let everything above pass through (record)
3. band pass: above and below cannot pass through
4. notch: filters out a specific frequency

Question 12

a problem with digital filtering

Answer 12

• phase distortion (usually phase lags occur)

Question 13

what is the most common source of noise in human motion capture?

• -> this noise is typically what?
• -> so we would apply what kind of filter?

Answer 13

small errors in position of biomechanical markers during the digital processing

• low amplitude, high frequency
• low pass filter

Question 14

when is noise amplified?

Answer 14

at every step of integration and differentiation

Question 15

in human locomotion, the highest voluntary frequency is what? this helps what?

Answer 15

• 30 hz

- helps define the minimal sampling rate