Page 450 - Introduction to AI Robotics

11.10 Exercises
Exercise 11.11 433
Consider how the robot moves from one frontier to another in frontier-based explo-
ration. What are the advantages and disadvantages of explicitly planning a path
between frontier centroids versus using a purely reactive move-to-goal and avoid set
of behaviors?

Exercise 11.12
Compare frontier-based exploration with GVG exploration. Is one better suited for
certain environments than the other?

Exercise 11.13
Consider how both exploration strategies (frontier and GVG) often have to choose
between possible areas to explore and then save the choices not taken for further
exploration. Which data structure do you think would be better to store the choices,
astack or apriority queue? Why?

Exercise 11.14 [Programming]
Write a program that displays a sonar model on a 26 by 17 grid. The origin of the
sonar should be at (26,8), with a = . Assume a sensor reading of s = wi 1th a 5 1 0
tolerance of 2 units.

a. Create a simple ASCII display. Each element should be labeled with the number
of its corresponding region: 1, 2, or 3. For elements outside of the field of view,
the label is 4.
b. Create a simple Bayesian display. Each element should be labeled with the correct
P (sjO ) value .
ccupied
c. Create a simple Dempster-Shafer display. Each element should be labeled with the
correct m(Occupied ) value (Dempster-Shafer).
d. Save the data from parts b. and c. and put in a spreadsheet or plotting program.
Display as a 3D plot.

Exercise 11.15 [Programming]
Write a program for a robot which moves to the centroid of an unknown area. Test it
for a variety of room configurations, and describe where it stops in each case.

Exercise 11.16 [Advanced Reading]
Read “Dervish: An Office-Navigating Robot,” by Illah Nourbakhsh, and “Xavier: A
Robot Navigation Architecture Based on Partially Observable Markov Decision Pro-
cess Models,” by Sven Koenig and Reid Simmons in Artificial Intelligence and Mobile
Robots: Case Studies of Succesful Robot System,ed. Kortenkamp,Bonasso, Murphy,
1998. Describe how these approaches would work for the example in Fig. 11.21. Com-
ment on the similarities and differences.