Real-time bounded-error pose estimation for road vehicles using vision

This paper is about online, constant-time pose estimation for road vehicles. We exploit both the state of the art in vision based SLAM and the wide availability of overhead imagery of road networks. We show that by formulating the pose estimation problem in a relative sense, we can estimate the vehicle pose in real-time and bound its absolute error by using overhead image priors. We demonstrate our technique on data gathered from a stereo pair on a vehicle traveling at 40 kph through urban streets. Crucially our method has no dependence on infrastructure, needs no workspace modification, is not dependent on GPS reception, requires only a single stereo pair and runs on an every day laptop.

I. Motivation and Background

It is hard to understate the importance of the transport of goods and people in daily life. We are totally dependant on it, thus, any increase in efficiency, access, safety or reliability will have a major economic and societal impact. This paper describes work towards this goal, motivated by the belief that information engineering, computing and robotics can provide a low cost solution for smart vehicles in civil, defence and industrial domains. Such vehicles offer the possibility of end-to-end goods transportation, improved efficiency and safety on our roads, and give our aged, infirm and sensorially impaired citizens the hope of independent personal transport.