Curb Tracking and Estimation (CuTE)
Speaker: Sarath Kodagoda, (speaker organisation unavailable)
When: 2003-10-22 10:00:00
Venue: 78-420
Host: A/Prof Janet Wiles
Abstract:Autonomous navigation in an urban environment is a challenging
task. Among other functions, an important task faced by designers of
such navigation systems is the sensing, detection and tracking of
roads. An urban/ semi-urban road environment is characterized by
curbs and painted lane markings in almost all sections of the road
network at least in my country of residence, Singapore, thus the
main inspiration and motivation of the research was to exploit such
features.
Since curbs and lane markings are highly correlated with road
boundaries, those can effectively be fused to enhance
robustness. Curbs are extracted using laser based range/bearing
measurements and lane markings are extracted using camera based
image. The tracking of curbs (equivalent to road boundaries) is
formulated as detecting and tracking maneuvering targets in clutter
from a moving platform using on board sensors. The maneuvering
target, i.e. the curb, is modeled as a nonlinear Markov switching
process. The target (curb) observations are of three kinds: laser
road boundary measurements modeled as traditional point-mass
observations, image road boundary measurements and image mode
measurements modeled as a discrete-time point process. These
observations are fused in a unified manner using an Image Enhanced
Interacting Multiple Model (IE-IMM) approach to realize the CuTE
(Curb Tracking and Estimation) algorithm. Extensive simulation and
experimental results are presented from the application of the CuTE
algorithm to a campus site environment to demonstrate its viability,
effectiveness and robustness. It can be concluded that the LMS and
camera based multi-sensor fusion using an image-enhanced interacting
multiple model framework is viable, effective and efficient in road
boundary extraction and tracking.
Biography:Sarath Kodagoda received the BSc. Eng. Hons. degree in electrical
engineering from the University of Moratuwa, Sri Lanka, in 1995, and
his M.Eng. degree in the area of mobile robots from Nanyang
Technological University, Singapore in 2000. He has finished his
research work on Ph.D. in school of Electrical and Electronic
Engineering, Nanyang Technological University, Singapore and
submitted the thesis for reviewing.
He worked as a Design and Sales Engineer in a multi-national
transformer manufacturing company: Toroid International Pvt Ltd from
1996-1998. From 2000-2002, he worked as a Research Associate in the
project: Development of an autonomous navigation system for an
outdoor autonomously guided vehicle - (ARC 3/95), in School of EEE,
Nanyang Technological University, Singapore contributing to the
design and development of the AGV. He is an author of more than 16
international journal and conference papers. His current research
interests are, mobile robotics, sensor fusion and target tracking.
Mr. Kodagoda is a Member of the IEE and Associate Member of the
Institution of Engineers, Sri Lanka. His biography is published in
the 07th edition of Who’s Who in Science and Engineering, USA.
Type: ITEE Seminar
Contact:A/Prof Janet Wiles, seminar host (janet@itee.uq.edu.au)
or Guido Governatori (ITEE seminar co-ordinator)
(guido@itee.uq.edu.au)