Abstract
Compared to vehicle trajectories that are solely generated from outdoor environments, most pedestrian GPS trajectories are recorded in mixed indoor and outdoor environments. Due to the problems of poor indoor accuracy and sparseness of signal points, processing of indoor GPS trajectories is significantly different from that of outdoor GPS data. Existing research often assumes that GPS signal is completely missing in indoor environments. However, with the sensitive GPS receivers and some big windows, satellite signals can also be picked up in indoor environments. To address the above problem, this chapter presents a signal-loss-based method to segment and analyze mixed indoor/outdoor pedestrian GPS trajectories. Firstly, by considering the signal-loss periods in indoor environments, a clustering method is proposed to segment indoor/outdoor sub-trajectories from each trajectory. Based on that, the approach for understanding trajectory patterns is developed, which uses features such as speed, distance and time to recognize “passing” pattern and “indoor activity” pattern in indoor environments, as well as “move-stop” pattern, “more-move” pattern and “more-stop” pattern in outdoor environments. Finally, we evaluate the proposed method with some real trajectories to study its feasibility in segmenting and analyzing mixed indoor/outdoor pedestrian GPS trajectories.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
