Data before filtering:
Data after outlier removal:
Data after filtering:
Another part of the preprocessing phase includes removing outliers, data points that change too drastically. We observed that this may be caused by the instability of the hardware. We removed data points that lie outside one standard deviation of the average value.
Having these graphs as visualization tools really helped us a lot in the development phase. It is easy to identify which part of the system goes down. We can see whether the chips are connecting and sending data, and whether the proximity detection works. We are also able to determine what kind of results the filters are providing.