Study and implementation of an autonomous navigation algorithm for a scale electric car

Abstract

This research describes the study of the autonomous navigation algorithm named Pure Pursuit. First, simulations were carried out in Matlab to evaluate its performance and then it was implemented in the embedded system myRIO of a scale electric vehicle called Aguara'i. It was possible to establish the values of lookahead distance so the navigation errors do not exceed a pre-defined threshold. For example, in a path for bicyclists at the city of Hernandarias for a velocity of 2.5 m/s the anticipated distance was about 1.8 m. One of the implementations developed was carried out in the Robocar Race 2018 competition (Santo André, Brazil), where the algorithm had a good performance, leaving Aguara'i with the 4th best time.