Logic and Closed Loops for a Computer Junket to Mars

Abstract

At present, three of us, Jose da Fonseca, Roberto Moreno Diaz and I, are assisting Louis Sutro of the Instrumentation Laboratory of M.I.T.; in his planning of a device to land on Mars to look for signs of life, say moving things or those bigger on top than on the bottom, like a tree or man. Because it must retrorocket in a thin atmosphere, it must crawl or roll out of the region it has blasted. Our primary work is with its vision, but it must have several other types of input for safe locomotion. It will have pressure and touch receptors on a hand to poke, and strain gauges on its wheels or springs and, of course, accelerometers to know which way is up and how it is moving. Without these it could not form, in its computer, a model of its surroundings to guide its motions. There is no possibility for it to send back to earth a full series of pictures detailed and fast enough for us to guide it. The channel back to earth is noisy and, with the little power promised, the rate of transmission is about what an amateur Radio Operator sends by hand. Worse yet! The time required for a single signal back to earth is of the order of 10 min. Hence we must build into it enough sense to observe a moving bug or worm and tell us what it saw or to study the landscape and describe it to us in a few well-chosen words.