The Roomba’s Ancestor: The Cybernetic Tortoise
spectrum.ieee.org
spectrum.ieee.org
The Johns Hopkins Beast (https://en.wikipedia.org/wiki/Johns_Hopkins_Beast) held an almost mythological sway over my imagination when I learned of it in the early 1960's. Lee Hoevel, then a graduate student of Mike Flynn, spoke affectionately of the beast and the noises it made in its nocturnal foraging for electric power sockets on the 2nd floor of the lab where he began his graduate work.
Digging into the pre-history of another experience turned up the following story. Thanks to Bruce Koball who was there.
Briefly, in 1987, Key Dewdney attended the Asilomar Microcomputer Workshop. In conversations with John Wharton and Bruce Koball, he challenged them to come up with a realization of Valentino Braitenberg's thought experiment as described in his monograph "Vehicles: Experiments in Synthetic Psychology" https://mitpress.mit.edu/books/vehicles. The prize was entry to the exclusive First Conference on Artificial Life to be held that fall Los Alamos National Labs. http://alife.org/conference/alife-1 John and Bruce took up his challenge; Bruce built the h/w (using a gutted radio-controlled model car) and John wrote the firmware. As a "front-end" we used a simple neural net simulation program written for Macintosh called "MacBrain" that Bruce had been experimenting with. It had a nifty GUI which allowed him to define neural net topologies on-screen and then download the net into the vehicle via an async serial port where it was executed in John's code. Their vehicle debuted at A-Life 1 in 1987, and then was shown again at the annual AMW (the Asilomar Microcomputer Workshop) and in Stanford's iconic Colloquium, EE380. Then, at Fred Coury's urging, Bruce wrote a brief paper, "A Test Vehicle for Braitenberg Control Structures", on our efforts for COMPCON Spring 1988. https://www.researchgate.net/publication/3496636_A_test_vehi... [Bruce R. Koball, Motion West, Berkeley; bkoball@well.com, bkoball@motionwest.com]
I, Dennis Allison, was an observer and not a participant in both these anecdotes. It was inspiring to see Bratenburg vehicles (https://en.wikipedia.org/wiki/Braitenberg_vehicle), designed as a thought experiment, actually demonstrate very complex behaviors even though the control program used a very simple rule set. John released several vehicles during the RATS session at the annual Asilomar Microcomputer Workshop where they scurried about under tables and chairs. Some hid, others vamped. They were an immediate hit, a showstopper with an audience of jaded techies. It took a while to find them all as some went into deep hiding mode. Later, John repeated the release as part of an EE380 Colloquium talk at Stanford, but Terman Auditorium lacked the hiding spaces Asilomar provided.