(figure 18). 15 How about developing a
“deep action” system, with cross connections to the perceptual hierarchy
and using (perhaps) hierarchical reinforcement learning to learn the
actions? 16 One could then try to use
both hierarchies to control a robot.
3. LIFE, November 20, 1970.
4. For more information, see the T-R website
5. See wiki.ros.org/smach.
6. The figure is from Nils J. Nilsson, “A
Mobile Automaton: An Application of Artificial Intelligence Techniques,” Proceedings
of the International Joint Conference on Artificial Intelligence, 7–9 May, 1969. Washington, DC. Los Altos, CA: William Kaufmann
7. See en.wikipedia.org/wiki/hierarchical_
task_network for more information.
8. See www.ai.sri.com/~sipe.
9. See www.cs.umd.edu/projects/shop.
10. See aigamedev.com/open/review/plan-
Figure 17. Some of the People Who Worked on Shakey
Front Row, left to right: Richard Fikes, Helen Chan Wolf. Rear row, left to right: Charles A. Rosen,
Bertram Raphael, Richard O. Duda, Milt Adams, Jerry Gleason, Alfred E. (Ted) Brain, Peter E. Hart, and Jim Baer.
11. Email from Steven Woodcock sent to
Nilsson on 6/14/2003.
12. See, for example, en.wikipedia.org/wiki/
13. For a video of the Hough Transform in
action, see www.youtube.com/watch?v=
14. See, for example, www.cs.toronto.edu/
15. Diagram from willcov.com/bio-con-
16. The following paper seems relevant to
this problem: Nicholas K. Jong and Peter