A Framework for Human Activity Recognition Based on WiFi CSI Signal Enhancement

Algorithm 4

Activity segmentation.

	Input: Se—the enhanced signal of CSI
	W—the size of sliding window
	N—the length of the sequential data of CSI
	step—the step size of window movement
	Output: Ts—the start time point of activity
	Te—the end time point of activity
	Step 1: for each subcarrier Sj
	Step 2: V = Ø
	Step 3: for (k = 0; k + W ≤ N; k = k + step)
	Step 4: calculate the mean(mk) of sequential data in sliding window from Sj in Se
	Step 5: append mk to V
	Step 6: end
	Step 7:
	Step 8: sort in ascending order
	Step 9: t = the numerical value of third quartile (75%) in sorted VS
	Step 10: filter out the value that is less than t in
	Step 11: the range of the remaining continuous data in is the start time() to the end time() in sequential data for the activity
	Step 12: end
	Step 13: Ts = min()
	Step 14: Te = max()
	Step 15: return Ts, Te