version of the Hough transform to include circles and
analytic curves and to use a rho-theta parameterization (Duda and Hart 1972). Their paper gets 4500 hits
on Google Scholar.)
The modern form of the Hough transform is used
in automobiles to detect lane markings to warn the
driver about drifting out of his or her lane. 13
One reason for the success of the Shakey project and
for its extensive legacy is that we were the first group
to think that developing a robot that could perceive its
environment and make and execute plans was a feasi-
ble idea. At the time, there was little existing software
for us to use, so we had to invent what we needed. It
turned out that the new inventions were ones that had
broad applicability once people heard about them.
Another reason for our success is that we had a
very talented team of AI researchers and software
developers, along with people who could make the
connections between software and hardware (figure
17). Some team members had a reunion at SRI International in November 2014.
There are still many problems in AI where talented
researchers could be first. An idea mentioned by Nilsson during the panel is to develop an “action hierarchy” analogous to the deep learning hierarchies that
are being used for vision and speech recognition. 14
Some of these are said to be rough models of the perceptual part of the neocortex. But the cortex also coordinates and plans actions, as illustrated in the diagram
Figure 15. A Hierarchical Task Network.
Figure 16. Region Finding as Used by Shakey’s Vision System.
Reprinted with permission from Claude Brice and Claude Fennema, Scene Analysis Using Regions. Artificial Intelligence 1970 ( 3-4).