A New Observation Model to Solve Vehicle Localization Problem

To localize itself, an autonomous vehicle should be able to observe its surrounding environment and extract the significant information to calculate its pose. We propose a new observation model to solve the localization problem of an autonomous vehicle equipped with laser rangefinder. In this method, a feature named "shortest-distance point" of natural landmarks is employed to build the observation equations for laser rangefinder to correct the vehicle's pose. An estimator using this method was designed and tested, showing a robust and effective localization