Multisensor data fusion for an autonomous ground vehicle

Abstract

Multisensor data fusion is an emerging technology applied to military areas such as automated target recognition, battlefield surveillance, guidance and control of autonomous vehicles and also to non-military applications such as monitoring of complex machinery, medical diagnosis, and smart buildings. Techniques for multisensor data fusion include artificial intelligence, pattern recognition, statistical estimation, and other areas. In this work, we have used multisensor data fusion for navigation and to control an autonomous vehicle. Various sensors such as GPS, digital camera, digital compass, laser scanner and ultrasound sensor have been used to gather data from structured surroundings. This data is then preprocessed and selectively fused to make reliable decisions. The sensors are interfaced with an onboard vehicle controller and the vehicle navigates a certain path on the basis of a robust multisensory data fusion algorithm.