Indoor Trajectory Mapping Research Based On Opportunistic Encounter Detection

With the rapid development of communication technology and intelligent hardware, smartphones has gained popularity rapidly, and the location based service on smartphones has been widely used. Because of the particularity of the indoor environment, the research method of indoor location and track mapping has encountered new challenges, but also new opportunities. Methods that locate and track indoor users have been mature, but can not be used widely in practice, one of the main reasons is most locating and tracking methods rely on constraints of indoor floor plans, which is not available in reality. In view of this, we study the behavior and characteristics of indoor pedestrian path constraint, and propose an indoor pathway mapping system and give the experimental simulation of this system. Then main contents of this paper are as follows.First of all, we include the method and main principle of states and events recognition, indoor localization and path mapping of indoor users, based-on smart phones, and analyze the advantages and disadvantages of current approaches. Then we summarize the main constraint factors to apply these methods.Secondly, based on the characteristics of wireless signal of indoor smart phones users, a new encounter events detection model using Wi-Fi direct mode was raised. This method log the change trend of the Wi-Fi signals scanned by each other with different of the distance and obstacles feature between them to recognize encounters. The encounter events are classified to meet, concomitance and pseudo encounter events.Thirdly, using the information of encounter events and other behavior identification methods of indoor users, we connect the path segments of indoor pedestrians generated by dead reckoning, iteratively train a unit correction vector, and apply it on corresponding users to improve the precision of trajectories.Lastly, we design complete experiment to evaluate our system and prove the feasibility of the two methods proposed above. Through error analysis, we validate the accuracy of encounter detection approach, and improve the accuracy of path mapping, to reduce the influence of drift error caused by hardware.