Pedestrian characterisation in urban environments combining WiFi and AI
Antonio Guillén, Maria‐Dolores Cano · International Journal of Sensor Networks · 2021
Knowing how many people there are in a given scenario offers new possibilities for the development of intelligent services. With this goal in mind, the use of sensors and Radio Frequency (RF) signals is becoming an interesting alternative to other classic methods such as image processing for counting people. In this paper we present a novel method for counting, characterizing, and localizing pedestrians in outdoor environments, called iPCW (intelligent Pedestrian Characterization using WiFi). iPCW is a passive, device-based sensor system that incorporates artificial intelligence techniques, more specifically, machine learning techniques. Performance evaluation using intensive computer simulations shows that iPCW achieves excellent results, with moving and static pedestrian detection accuracy above 98% and positioning accuracy above 92%.