Indoor Localization for the Visually Impaired

open_access
Open Access
Author:
Kawchak, Thomas Michael
Area of Honors:
Computer Engineering
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
Dr. Vijaykrishnan Narayanan, Thesis Supervisor
John Morgan Sampson, Thesis Honors Advisor
Keywords:
SLAM
Indoor Localization
GIST
SURF
Computer Vision
Position Tracking
Drift
Navigation
Abstract:
Visually impaired people rely on sighted people to perform many navigation tasks. With advances in technologies to perform localization and navigation, visually impaired people will gain autonomy for actions like grocery shopping or moving around within buildings. This Thesis reviews recent techniques for indoor localization and proposes a system that can be used to help visually impaired people navigate. Currently, navigation systems focus on robotics, autonomous vehiclesandaugmentedreality(AR)experiences. Further,stateofthearttechniquesforsimultaneous localization and mapping (SLAM) do not scale well to large areas due to associated problems like drift. This work will propose and test a system that can be used to help localize and navigate visually impaired users in larger spaces using a portable device with very little prior exposure to andinformationofthearea. TheproposedsystemwillusetheGISTimagefeaturetoformarough location estimate and the ZED camera spatial mapping features to perform finer localization. At certain levels of localization granularity, the system is able to achieve better performance than the ZED camera. The impact of this work will have implications for SLAM that is being performed at large scales with the aim ofhelping to provide a system thatwill helpvisually impaired users in indoor environments.