Dynamic Time Warping Based Passive Crowd Counting Using WiFi Received Signal Strength

In the current era, the Internet of Things is developing rapidly. The increasing number of persons pay attention to personnel information in an area. Crowd counting is favored by many researchers. It can be applied in many people-centric scenarios, such as the smart home and supermarket energy management. In this paper, we only use a pair of transceivers, relying on the Received Signal Strength Indicator (RSSI) information of the commercial WiFi signal to count the crowd without requiring the people carry any device. We first model the received signal into three parts, which are the Line-of-Sight (LoS) path blockage effect, Multipath (MP) effect on the received signal, and the multipath effect resulting from signal reflection by the fixed objects. Then, we analyze the Probability Density Function (PDF) of the received signal based on the characteristic function and then combine two different distributions to characterize the relationship between the number of persons and the PDF of the received signal amplitude. Finally, we use the Dynamic Time Warping (DTW) algorithm for crowd counting. We validate the performance of the approach in an outdoor environment, and the experimental results show that our approach can count four persons with an average accuracy of 96.25%.