Estratégia para função de detecção e reconhecimento de faixas em rodovias com aplicação a sistemas de assistência ao condutor

Abstract

The automotive functions of ADAS (Advanced Driver Assistance Systems) in the current scope are criterion of technological and commercial differentiation, in the future such functions will be essential in the terrestrial vehicles. The development of these functions with the model-based design methodology ensures compatibility and safety for current and future vehicles, with the possibility of several verifications and validations aligned to the development. This work is focused on the development of the strategy of an ADAS function for the lanes recognition, using RANSAC (Random Sample Consensus) statistical method, for the application of LDW (Lane Departure Warning) and resulted in a robust algorithm that guaranteed security and compatibility, but still with possibilities for improvements. The quantitative and qualitative tests performed with the algorithm demonstrate the assertiveness in the lanes recognition in several situations, and consequently the assertiveness of the LDW.