Environment mapping for autonomous driving into parking lots

We are presenting a proceeding of registering and evaluating data of the environment using a laser scanner and ultrasonic sensor cells programmable in Embedded Matlab/Simulink Software, having the aim of real time navigation and path determination of a car like vehicle. We assume that laser scanners have the advantage of producing reliable data with well understood characteristics for map generation. Also we are using ultrasonic sensor cells for object identification in short ranges. A vision system completes the floor marking identification available in some parking lots. The final goal of our project defines the capability of a fully autonomous vehicle to safely drive through the environment until reaching the standard parking lots and complete autonomous a parking procedure. By implementing evolutive algorithms we process data into lines representing edges of the surrounding objects and create a simplified representation of the environment. Because of the dynamic evolution of the map, during the movement of the vehicle we are considering of merging and fitting the data by applying a shape correlation.