Real-Time Indoor Localization using Visual and Inertial Odometry

This project encompassed the design of a mobile, real-time localization device for use in an indoor environment. A system was designed and constructed using visual and inertial odometry methods to meet the project requirements. Stereoscopic image features were detected through a C++ Sobel filter implementation and matched. An inertial measurement unit (IMU) provided raw acceleration and rotation coordinates which were transformed into a global frame of reference. A Kalman filter produced motion approximations from the input data and transmitted the Kalman position state coordinates via a radio transceiver to a remote base station. This station used a graphical user interface to map the incoming coordinates