Instant Insanity – Computer Vision & Robotics

[Recorded: 1971]
Over the last decades, computer vision systems have become increasingly capable of controlling robotic movement. One example of early research and development of computer vision and robotic systems was recorded at the Stanford University Artificial Intelligence Laboratory in 1971. This film shows a computer vision system and robotic arm solve the Instant Insanity puzzle, which has been around for more than a century under various aliases. It consists of a set of four cubes with one of four colors on each of their six faces. The goal is to arrange the four cubes in a row so that all four colors appear on each of the row’s four long sides. The order of the cubes doesn’t matter, but that simplicity is deceptive. There are 41,472 different ways of arranging the four cubes in a row, so this is not a trivial task. The computer vision system first finds each of the four cubes by matching the visual edges to a prototype cube. In the case of a cube with only two faces visible, the arm turns the cube 45º so that three faces will be visible. The colors of the faces are then determined by reading in the scene again under three different color filters. The cubes are then turned over so that the three hidden back faces are visible to the camera and the process repeated. Once a solution is found the computer directs the arm to stack the blocks in the required order.
Originally, this film was recorded without sound. A narration by Les Earnest and Lou Paul, of the Stanford AI Lab was added in 2009.
Original Film Credits: Richard Paul, Karl Pingle, Jerome Feldman, & Alan Kay
This film made available to the Computer History Museum courtesy of Stanford University.

Duration : 0:6:27

[youtube O1oJzUSlTeY]