Automated human mobility mode detection based on GPS tracking data

GPS-enabled devices have greatly facilitated the continuous collection of highly accurate locational data for moving objects including humans. The effective and efficient detection of mobility mode from raw GPS data is critically important for social behavior and public health research. In this paper, we propose a new method that can be used to automate the detection of human mobility mode by synthetically analyzing geographic location, duration, speed as well as spatial context information in the data. Ancillary GIS layers such as building footprints are also included to help identify the GPS points taken indoor. Five mobility modes: transporting, parking, walking, roaming and indoor are detected in our method. Preliminary testing of the method applied on three datasets showed an overall accuracy of above 90% when compared with the human interpretation results. A software application has also been developed to automate the detection and report generation.