Page 184 - Autonomous Mobile Robots

168 Autonomous Mobile Robots

(a) Landmark
y
x

……
a 0 a n
a i

Robot trajectory
z i z 0 z n
Position estimation based on a single landmark
(b) Landmarks
y

x
……

a 2i
a 20
a 1i
a 10
Robot trajectory
z i z 0
Position estimation based on two landmarks

FIGURE 4.9 Landmark-based localization of the robot.

Equation (4.28) shows the relationship between localization errors and char-
acter extraction errors. It can be seen that the localization error is relative to l 1
and l 2 , as well as the distances between the robot and the features, which will
be large when the observing angle is large. In the two-landmark case, the two
features p 1 and p 2 can be selected from different landmarks, which can provide
more accurate position results.

4.4.3 Least Square Estimator (LSE)

In a real application, the robot continuously samples data using its onboard
camera. Errors may be reduced by fusing the data of individual samples. In this
section, LSEs are used in terms of two different cases: single landmark case

© 2006 by Taylor & Francis Group, LLC

FRANKL: “dk6033_c004” — 2006/3/31 — 16:42 — page 168 — #20