Evaluation of a transport mode detection using fuzzy rules

Abstract GPS and other sensor data available from smartphones are more and more used to derive the mobility behavior of their users. This contribution evaluates an own developed three step algorithm to derive transportation modes from such data by taking the trajectory data of some dozen trips of users with a focus on obtaining road traffic information. The algorithm itself consists of first a segmentation phase, second fuzzy rule transport mode detection and third consistency corrections. Results obtained so far are promising with correct detections of about 75 %, but further analysis will be needed to substantially enhance the algorithm.