Page 19 - Autonomous Mobile Robots

2 Autonomous Mobile Robots

further improve the perceptual capabilities of systems. This issue is discussed,
with examples on the fusion of image data with LADAR information. The
chapter concludes with a discussion on the open problems and challenges in
the area of visual perception.
Where visual sensing is insufficient, other sensors serve as additional
sources of information, and are equally important in improving the naviga-
tional and perceptual capabilities of autonomous robots. The use of millimeter
wave RADAR for performing feature detection and navigation is treated in
detail in the second chapter of this part. Millimeter wave RADAR is capable of
providing high-fidelity range information when vision sensors fail under poor
visibility conditions, and is therefore, a useful tool for robots to use in perceiving
their environment. The chapter first deals with the analysis and characterization
of noise affecting the measurements of millimeter wave RADAR. A method is
then proposed for the accurate prediction of range spectra. This is followed by
the description of a robust algorithm, based on target presence probability, to
improve feature detection in highly cluttered environments.
Aside from providing robots with a view of the environment it is immersed
in, certain sensors also give robots the ability to analyze and evaluate its
own state, namely, its position. Augmentation of such information with those
garnered from environmental perception further provides robots with a clearer
picture of the condition of its environment and the robot’s own role within
it. While visual perception may be used for localization, the use of internal
and external sensors, like the Inertial Navigation System (INS) and the Global
Positioning System (GPS), allows refinement of estimated values. The third
chapter of this part treats, in detail, the use of both INS and GPS for position
estimation. This chapter first provides a comprehensive review of the Extended
Kalman Filter (EKF), as well as the basics of GPS and INS. Detailed treat-
ment of the use of the EKF in fusing measurements from GPS and INS is
then provided, followed by a discussion of various approaches that have been
proposed for the fusion of GPS and INS.
In addition to internal and external explicit measurements, landmarks in the
environment may also be utilized by the robots to get a sense of where they
are. This may be done through triangulation techniques, which are described
in the final chapter of this part. Recognition of landmarks may be performed
by the visual sensors, and localization is achieved through the association of
landmarks with those in internal maps, thereby providing position estimates.
The chapter provides descriptions and experimental results of several different
techniques for landmark-based position estimation. Different landmarks are
used, ranging from laser beacons to visually distinct landmarks, to moveable
landmarks mounted on robots for multi-robot localization.
This part of the book aims to provide readers with an understanding of the
theoretical and practical issues involved in the use of sensors, and the important
role sensors play in determining (and limiting) the degree of autonomy mobile

© 2006 by Taylor & Francis Group, LLC

FRANKL: “dk6033_c001” — 2006/3/31 — 16:42 — page2—#2