Undergrad Research Project - Data fusion for Improved Localization

Fall 2015

Joel Loo
Martial Hebert
Project description

The goal of this project is to improve the Global Positioning System (GPS) localization of a vehicle by fusing it with other data. We have a large dataset of Global Positioning System (GPS), image, accelerometer, gyroscope, and orientation data currently available, and we will first use a Kalman filter (in particular an open source implementation available on the Robot Operating System (ROS) as a package) to attempt to fuse the GPS data with the accelerometer and gyroscope data, to improve the localization capabilities of the vehicle. To achieve this we will also need to investigate and understand the uncertainty characteristics of the sensors on the vehicle being used for localization. Once this is achieved, we will additionally try to use other information to further improve the vehicle's localization capabilities - in particular, we will try to use map information and landmarks like Stop signs in localization. In all, we hope to improve the vehicle's localization capabilities in practice.

Return to project list