Data fusion using fuzzy-valued logic

Data fusion is an important function of all intelligent vehicle highway systems (IVHS) components. Raw data on traffic conditions are received from various sources, in several formats, and at different time intervals. The goal of data fusion is to combine such data into meaningful inferences about traffic conditions. But it is quite common for these input data to have inconsistencies, uncertainties, and a lack of completeness. Applying binary logic and Bayes decision theory is inappropriate because some contradictions and only partial information are typically present in the input data. This paper presents an alternative approach to data fusion using a fuzzy-valued logic generalized from Belnap's four valued logic.