Development of a low-cost, self-contained, combined vision and inertial navigation system

In the field of autonomous vehicle research, maintaining accurate estimates of a vehicle's velocities, position, and orientation is crucial to successful navigation. Normally, this information can be obtained through the use of a GPS system. However, GPS signals are not always available in some of the more exotic or dangerous locations that a vehicle might be dispatched to. This paper proposes a hybrid navigation system for autonomous vehicles that uses inertial sensors and a stereo camera system to accurately determine the velocities, position, and orientation of a vehicle. Such a system would be ideal for vehicles that have to operate in underground, indoor, extraterrestrial, or other GPS-denied environments.