Uncertainty evaluation of 2D vehicle position measurement by probability and theory of evidence approaches

A two dimensional example of uncertainty expression and propagation with different approaches is described. A first approach is the already known method based on the theory of probability which uses probability density functions to express uncertainty and relies on the Monte Carlo method for uncertainty propagation. A second approach, which is innovative and not yet well known in measurement, is based on the theory of evidence, employing random-fuzzy variables to express and propagate uncertainty. This work describes the application of those two approaches to the 2D case of vehicle position measurement on a plane. The definition of a procedure for uncertainty evaluation in multi dimensional measurements is currently under development for the probabilistic approach, and is completely new for the theory of evidence approach. The results show that the confidence regions obtained by the two different methods are compatible and that, the RFV approach yields larger confidence regions than the probability one, reducing the risk of uncertainty underestimation in presence of complete ignorance situations and systematic effects.