Vehicle Navigation in Dense Urban Environments Using GPS, Image, and IMU Measurements from Commercial Off-The-Shelf Sensors

Johnathan Rohde, Jason Png, John Raquet, and David Jacques

Abstract:	This paper describes an algorithm and test results for integrating GPS, monocular vision, and IMU measurements from COTS sensors in order to navigate a vehicle in a heavy urban environment, where there are not sufficient GPS measurements available to achieve an accurate solution. A 2-D zero velocity update approach is used for the monocular vision update, since scale cannot be measured. The 2-D update leverages IMU data and epipolar constraints from tracked features between subsequent images. The algorithm is tested using both simulated and real data, and performance is evaluated as a function of camera direction and number of GPS satellites.
Published in:	Proceedings of the ION 2015 Pacific PNT Meeting April 20 - 23, 2015 Marriott Waikiki Beach Resort & Spa Honolulu, Hawaii
Pages:	591 - 598
Cite this article:	Rohde, Johnathan, Png, Jason, Raquet, John, Jacques, David, "Vehicle Navigation in Dense Urban Environments Using GPS, Image, and IMU Measurements from Commercial Off-The-Shelf Sensors," Proceedings of the ION 2015 Pacific PNT Meeting, Honolulu, Hawaii, April 2015, pp. 591-598.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In