Object detection and classification in outdoor environments for autonomous passenger vehicle navigation based on data-fusion of an artificial vision system and laser sensor

• Grant number: 11/03986-9
• Support Opportunities: Scholarships in Brazil - Master
• Effective date (Start): August 01, 2011
• Effective date (End): January 31, 2014
• Field of knowledge: Interdisciplinary Subjects
• Principal Investigator: Marcelo Becker
• Grantee: Henry Antonio Roncancio Velandia
• Host Institution: Escola de Engenharia de São Carlos (EESC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Associated research grant: 08/57858-9 - National Institute for Optics and Photonics, AP.TEM
• Associated scholarship(s): 12/21079-1 - Path detection for autonomous vehicles using computer vision and on-line learning, BE.EP.MS

Abstract

This research project takes part in the SENA project (Autonomous Embedded Navigation System,) which will be developed at the Laboratório de Robótica Móvel of the Mechatronics Group at Engineering School of São Carlos - University of São Paulo (EESC - USP) in collaboration with the Instituto de Ciências Matemáticas e de Computação (ICMC), and the Instituto de Física de São Carlos (IFSC). The main goal for SENA project is to develop assistive and autonomous technologies that can deal with the real needs of different types of drivers (unskilled, old aged, disabled, etc.) For this purpose, this project is currently receiving funding from 2 INCTs, the INCT- INOF - Instituto Nacional de Óptica e Fotônica and the INCT - SEC - Sistemas Embarcados Críticos. It is expected that the application on a large scale of this kind of technology may reduce drastically the quantity of injured and dead people in automobile accidents happening in urban-like environments and streets. In this way, this research project aims at providing information of the environment (around the vehicle) for the decision making and control system embedded in an autonomous vehicle. The most basic information to be provided is the existence of objects (obstacles) around the vehicle; moreover, the sort of object, i.e., its classification in cars, pedestrians, road signs, sidewalks and street marks as well as its position and scale will be also provided. The data of the environment will be acquired by using a camera and a laser measurement system (LMS). One of the challenges of this research is to accomplish a robust classification invariant to image scale, rotation, changes in viewpoints of the objects, and changes of illumination in the environment. (AU)