A Fuzzy Adaptive Data Fusion Strategy for Intelligent Vehicle Systems

Progress in science and technology has accelerated the pace of development in the intelligent vehicle system. Multiple sensors distributed in the system are used to gather and determine vehicle environmental information. Regardless of the accuracy of sensors, the data fusion algorithm employed in multi sensors system affects the decision-making directly and deeply. According, it is necessary to design appropriate algorithms to achieve sensors data processing. This paper presents a novel fuzzy adaptive data fusion strategy based on the federal Kalman filter and fuzzy inference so as to achieve more accurate results. In the strategy, fuzzy inference is used to adjust the level of measurement noise, while the purpose of federal Kalman filter is to obtain the smooth data. Besides, calculating the mean value of residuals, we illustrate a method to reject measurement defects and substitute the previous effective value for them. Experimental results show that the proposed strategy can effectively detect and correct defects, and obtain more accurate data processing results compared with the federal Kalman filter.