Controller for a four legged walking machine

When I was in high school and working at the Sunnyvale Public Library, I remember sneaking a read at an article in one of the Smithsonian Magazine issues that I was supposed to be shelving. It described analog, insect-like robots from work in Los Alamos. It has a big effect on me. This was before I knew I had any interest in distributed systems or any distaste for old fashioned AI, but I was really impressed by how simple simple robots were exhibiting more complicated and adaptive behavior than the most centralized, state of the art robots had managed. I'm pretty sure that this paper describes some of the work that I read about. It is a '98 paper, and I worked at the library through '99.

It describes itself as a hybrid based on Ferrell's definition of the term as a compromise between 'patterned' and 'reflexive' approaches to locomotion controllers. Though it is thought o fas a hybrid, it has no sensors. My understanding of their circuits and signals is shoddy, but it seems that the sensing/feedback is an implicit result of the behvior of the way the motors are wired.

CPG is a Central Pattern Generator.

I really like how this article is written. It is more of a detailed 'note to self', or lab report, than it is an argument for this or that pedagogy. It seems to have been written solely for the general utility of recording what happens in one's experiments.