The Recognition of Roadbed Based on Fuzzy Etalons.
In the paper we solve the problem of recognizing the roadbed using information about its possible typical designs. As shown on the paper typical designs can be adequately described by the set of fuzzy features, derived from the polygonal representation of the roadbed cross profile. The recognition is based on the minimization of a functional that takes into account how thepolygon representation fits to the geometrical form of a measured profile and how its features obey the fuzzy restrictions. We propose to minimize this functional using probabilistic genetic algorithm. In the paper we also discuss the obtained results of the conducted experiments.