This paper presents the sensor data fusion approach of the European Union(EU) project FRICTION. The project objective was to create an on-board system for estimating friction and road slipperiness to enhance the performance of integrated and cooperative vehicle safety systems. Using a single sensor approach it is difficult to provide high quality friction information for multiple safety systems. Consequently, three types of sensors were used in the study: (I) existing in-vehicle sensors measuring vehicle dynamics, (II) environmental sensors, and (III) a tire-based sensor prototype. Adata fusion architecture is presented consisting of three feature fusion modules and a decision fusion module. The algorithms have been validated with a development vehicle and test results are presented achieving nearly continuous friction estimation.
Abstract