Page 170 - Autonomous Mobile Robots

154 Autonomous Mobile Robots

the environment is two-dimensional. The position of the robot with respect to
this map is unknown and needs to be determined.
Therearethreekindsoflandmarksthatareconsideredinourdesign(i)single
strip retro-reflective landmarks, (ii) digital landmarks, and (iii) geometric land-
marks. The positions of the first two kinds of landmarks are pre-input into a
global map of the robot’s environment. The positions of the geometric land-
marks are abstracted and then registered into the global map dynamically. The
next three sections describe their application in robot navigation respectively.
Note that our navigation system is not restricted to these three kinds of land-
marks, and can be easily extended into other kind of landmarks and their
combination.

4.3 LASER SCANNER AND RETRO-REFLECTIVE LANDMARKS

4.3.1 Laser Scanner and Angle Observation
The localization system based on the laser scanner and retro-reflective land-
marks is a promising absolute positioning technique in terms of performance
and cost [8]. Using this technology, the coordinates of retro-reflective land-
marks are prestored into an environment map. During its operation, the robot
uses its onboard laser sensor to scan these landmarks in its surroundings and
measure the bearing relative to each of them [1]. Then the position estimation
of the mobile robot is normally calculated by using two distinctive methods:
triangulation [12] and Kalman filtering [14].
Research here is based on a rotating laser scanner that is able to measure the
angle between the robot base line and the beam line from either the leading or
the falling edge of landmarks in the horizontal plane. As shown in Figure 4.2,
the laser scanner is situated on top of the physical center of the robot, scanned
360 in azimuth up to 50 m range at a constant speed of 2 Hz. Note that an
◦
IR laser beam (870 nm) from a HeNe laser diode emits energy of 0.5 mW,
which is eye-safe. As can be seen, there are six landmarks in this environment,
namely B 1 , B 2 , ... , B 6 . The landmarks are in the form of a single strip for easy
detection from a large distance, instead of traditional bar-codes. All landmarks
have an identical size of 50 cm in length and 10 cm in width. The positions of
the landmarks are surveyed in advance and prestored into the robot memory as
a look-up table, represented by the coordinates in the world frame:

=[B 1 , ... , B i , ... , B N ]=[(b x1 , b y1 ), ... , (b xi , b yi ), ...] (4.1)

where (b xi , b yi ) are the coordinates of the ith landmark and N is the total number
of landmarks in use.
These landmarks can return strong reflective signals to the scanner, that
is, the area inside dotted lines in Figure 4.2. The reflected light from these

© 2006 by Taylor & Francis Group, LLC

FRANKL: “dk6033_c004” — 2006/3/31 — 16:42 — page 154 — #6