Pedestrian GraphSLAM using Smartphone-based PDR in Indoor Environments
Dr. Mahi Abdelbar, Wireless@VT speaks on Simultaneous Localization and Mapping (SLAM) for pedestrians is a relatively new approach for the indoor localization problem.
Abstract:Adopted from robotics, SLAM for indoor pedestrians presents a new and efficient framework for tracking users’ movement trajectories within buildings. With the advancements in smartphone technology, pedestrian SLAM has transitioned towards utilizing smartphones’ integrated sensors through Pedestrian Dead-Reckoning (PDR) techniques. GraphSLAM models the spatial structure of a sequence of user’s positions inside a building as a graph optimization problem, based on estimated positions through PDR. In addition, GraphSLAM depends on loop-closures and/or landmarks as constraints in the trajectory estimation problem, which in pedestrian SLAM is still very challenging.
This talk present an overview of the pedestrian SLAM problem and its challenges, as a solution to the indoor localization problem. This work comes at the intersection of indoor localization using movement trajectories, pedestrian dead-reckoning (PDR) and Graph Simultaneous Localization and Mapping (GraphSLAM).