Travel Mode Detection based on Wi-Fi Signal Probe Using ANFIS

Recognizing the pattern of movement and the traveling mode is one of the most important issues in the field of urban analysis and transportation management. Disadvantages of the traditional method such as the questionnaire is the lack of availability and cost. So, after the emergence of new technologies, communication tools such as mobile phones are used to collect and analyze traffic data. The Wi-Fi network has been considered by intelligent transportation systems due to high availability and low cost. In this research, users in the specific area are classified into three categories using the Wi-Fi scanner. Pedestrians, passing cars and users who stop in the area for long time. This classification has been made using the extraction three features on the users' Wi-Fi signals and using an adaptive network-based fuzzy inference system (ANFIS) model. The results show that the proposed model for using the subtractive clustering method in the first layer of ANFIS to determine the initial membership function of the features has been able to classify the three categories with a accuracy of 84%, and the precision and recall of the user's detection Which used cars in this area, was 75% and 90%, respectively.