MAP-BASED FREE NAVIGATION FOR AUTONOMOUS VEHICLES

A general method for the construction of a map-based navigation system is proposed, which is based on a Kalman filter and can be equipped wi th many commercially available navigation sensors. A prototype has bee n implemented using a laser rangefinder and a 2D model of the natural environment. No artificial landmarks are necessary. The computing rime required for matching the measurement data with the model elements is low and independent of the model size. This is achieved by precompili ng the navigation map to form a lookup table. The classical Kalman fil ter algorithm is modified to take into account the different speeds of the absolute and relative sensors as well as the computing time of th e algorithm itself. A systematic way is given to decouple the time whe n absolute measurement data ave taken from the time when these data ar e processed. Loss of track resulting from errors in the navigation map or incorrect dead reckoning is detected by comparing the real measure ment data with on-line simulated measurement data. The on-fine simulat ion is performed in real time using the same lookup table which is uti lized for the matching process. For industrial applications this self- supervision is an essential feature.