Fischer, Carl and Gellersen, Hans (2012) Localisation, tracking, and navigation support for pedestrians in uninstrumented and unknown environments. PhD thesis, Lancaster University.
thesis_revised.pdf - Accepted Version
Available under License Creative Commons Attribution.
Download (11MB)
Abstract
Research into localisation and tracking of pedestrians is a growing area, but developed techniques have tended to be infrastructure-based or to rely on accessto prior information such as floorplans and maps. There are situations, such assearch and rescue missions, where more robust and versatile self-contained localisation systems are desirable. To our knowledge, this a relatively unexplored area of research. Relative localisation, involving only objects in the vicinity of the target being tracked, offers a partial solution by not requiring any infrastructure. We demonstrate and discuss some techniques based on this approach, and develop an algorithm suitable for tracking highly mobile sensor networks. We also highlight its limitations and look for complementary oolutions. Pedestrian dead reckoning (PDR) based on foot-mounted inertial sensors is a promising method which we describe in detail, including its inherent flaws. We combine the dead reckoning and sensor node techniques to perform simultaneous localisation and mapping (SLAM) for pedestrians in indoor environments. Finally, we include our SLAM algorithm in a complete navigation solution which we evaluate in a virtual environment.This study allows us to offer insight into problems and opportunities offered by these technologies and their application in the field of pedestrian navigation in uninstrumented and unknown environments. We describe the potential offered by tracking and navigation once they are no longer dependent on infrastructure, predeployment, or prior knowledge of an area.