An Novel Indoor Crowd Counting Approach Based on WiFi Signals

Abstract

Crowd counting approaches are widely used in personnel arrangements in crowded places and early warning of critical situations for their ability to accurately and quickly estimate the number of crowds. At present, researchers have proposed some methods for crowd counting using cameras and wearable devices. But these methods have some defects, for instance, cameras can only provide monitoring within the visible range, and wearable devices need to be consciously worn by the monitored. There are also some methods used radar-related technology to realize through-wall perception and recognition, but this kind of system has a complex design and high application cost and is mostly used in the military field. Considering the above problems, this paper proposes a crowd counting method based on WiFi signal, which uses the amplitude fluctuation of Channel State Information (CSI) to describe the change of indoor population and realizes crowd counting based on the Resnet network algorithm. At the same time, we verify the performance of the approach, and the experimental results show that the performance of this method is better than other algorithms.