Abstract

There is still limited understanding of people’s behaviours in transit situations, which results in a gap in smoothly"br" guiding users from a start to a destination in multi-modal transportation networks. Thereby an increasing number"br" of personal mobility scenarios require the localization of individuals or a group of end users within a group or"br" neighbourhood. Not rarely, this is the case in adverse operating conditions where only limited positioning"br" persons is not sufficient for real-time passenger guidance. A solution to this problem is referred to as Cooperative"br" Positioning (CP). In CP, end user localization is performed using its own measurement data plus any additional"br" information coming from neighbouring users in the form of inter-nodal ranges or other means. In this study,"br" absolute indoor localization of a user group is enabled with a novel differential Wi-Fi (Wireless Fidelity, a.k.a."br" WLAN) positioning approach. Similar to Differential GPS (DGPS), the received signal strength (RSS)"br" measurements of the Wi-Fi access points (APs) are corrected to reduce short- and long-time variations of the RSS"br" at the user side by the use of reference stations. In addition, continuous positioning is achieved by integration of"br" other smartphone sensors, such as accelerometer, magnetometer, gyroscope and barometric pressure sensor. From"br" the calculated positions, trajectories of the users are derived which are processed with machine learning methods"br" and analysed in view of the movement behaviour of the group. Using these trajectories guidance information is"br" derived with regard to a possible group subdivision and management on different routes, for instance, in the case"br" of large crowds during an emergency situation or service interruption. Thus, the traffic flow and security aspect is"br" fulfilled. A further main attention in this study is led on the support of people with limited mobility and visually"br" impaired. These challenging tasks can only be achieved if a positioning and tracking concept on the micro-level"br" in the stations is provided for all end-users. In the practical evaluation, static and kinematic tests using a number"br" of users carrying different smartphones were conducted in an indoor lab setting. Raspberry Pi’s served as reference"br" stations and APs together with inclusions of available and visible Wi-Fi APs. One user enabled the hotspot function"br" on his mobile device so that the other users can measure the RSS to derive inter-nodal ranges. It could be proven"br" that the new CP DWi-Fi approach outperforms conventional localization algorithms due to a significant"br" improvement of the indoor positioning and tracking accuracy.