Template-based autonomous navigation in urban environments

Abstract

Autonomous navigation is a fundamental task in mobile robotics. In the last years, several approaches have been addressing the autonomous navigation in outdoor environments. Lately it has also been extended to robotic vehicles in urban environments. This paper focus in the road identification problem, which is an important capability to autonomous vehicle drive. Our approach is based on image processing, template matching classification, and finite state machines processing. The proposed system allows to train an image segmentation algorithm in order to identify navigable and non-navigable regions (inside/outside roads), generating as output the steering control for an Electric Autonomous Vehicle, that should stay following the road. Several experimental tests have been carried out under different environmental conditions to evaluate the proposed techniques.

Vision-Based Autonomous Navigation Using Supervised Learning Techniques

“…3 (d)). Each template is composed by a mask of 1s and 0s [15]. The value of each mask is multiplied by the correspondent value into the navigability matrix (values obtained from the ANN classification of the correspondent blocks of the image).…”

“…In this work, we use the FSM with only 2 intermediate transitions between the states and have produced reasonable results. Detailed information can be seen in [15].…”

Investigation on the evolution of a robotic controller for autonomous vehicle navigation

“…The test platform CaRINA I -Intelligent Robotic Car for Autonomous Navigation. The development of the CaRINA is an ongoing project, as can be seen in [5,23,24]. non-safe/non-navigable areas from images acquired from a video camera.…”

“…It also addresses a more embracing set of parameters that impact on the GA and ANN behavior. In previous paper [23,24] we have evaluated the use of image pro-cessing techniques and neural networks in conducting an autonomous vehicle. This paper extends previous work aiming at present how the ANN could be evolved by using Genetic Algorithms.…”

Navegação autônoma para robôs móveis usando aprendizado supervisionado.

“…Um deles identifica uma rua em linha reta, dois identificam curvas suaves esquerda e direita e dois curvas rígidas esquerda e direita. Cada modelo (Template) é composto por uma máscara de uns e zeros (Figura 4.6), como proposto em Souza et al (2011a). O valor de cada máscara é multiplicado pelo valor correspondente da matriz do mapa de navegabilidade (Figura 4.3 (c)).…”