1 MATLAB assignment
What is the purpose of line 24 of the “Filter_Data.m” script? thigh_accel =
th_R*[th_acc_cs(:,1),th_acc_cs(:,2),th_acc_cs(:,3)]';
The reason for the line 24 is in the filter data script, it’s to reotate the
accelerometer data so it has its reference frames. The rotation was done by
multiplying the raw accelerometer data with the rotated one
2. Why smoothing the z-component of acceleration using “smoothdata” with
no option, i.e. simple moving average, is not sufficient for correctly
identifying the peak points (Figure 3)?
Smoothing data doesn’t always give the best noise free data. To go
over the smoothing data, the average of the numbers and data that’s taken
is used to fix the peaks of the graph. But when the peaks are smoothed,
there might be a bit of attenuation that makes the graph challenging to read
or to process.
3. How does the code find the minimum point of the acceleration using the
same “findpeaks” command/function?
The code finds the minimum points by determining the same find
peask command. This is then run through a filter command along with the
inversion command. The peaks will then be applied to these commands and
the corresponding minimums and peaks will be separated. The graph is then
plotten with time vs the data.
4. Which of the smoothing approaches is more suitable for identifying the
peaks? Why?
The gaussian smoothing methods was seen to be the most liked
method to smoothing because it applies the gaussian distribution to the data
effectively and it also reduces noise while preserving important information
like the peaks. It also balances the noise and retention of the peaks.
5. How did the code find the peaks in the y-component of the acceleration
signal without significant smoothing? Elaborate on how using the
'MinPeakProminence' enables finding the peak.
The code uses a technique called thresholding to identify peaks and
from there, it takes the MinPeakProminence command to find peaks in the
findpeaks function. This finds the minimum of the peaks. The code identifies
the peaks in the x and y scale of a graph and from there it balances the noise
to retention ratio.
