In the early 1960s, a robot that could think for itself was the stuff of science fiction. When Charles Rosen, who had founded the Machine Learning Group at Stanford Research Center (now SRI International), proposed building one in 1964, his idea was met with skepticism by many in the nascent AI field.
Rosen applied for funding to the Advanced Research Projects Agency (the government entity now known as the Defense Advanced Research Projects Agency, or DARPA), which grants funds for the development of emerging technologies. It took Rosen two years to get the funding. And it took six more years, until 1972, before engineers at Stanford’s AI Center finished building Shakey [above]. It was the first mobile robot programmed to “think” through its actions, such as navigating its way from room to room. Unlike other robots at the time, which had to be programmed on how to complete each step of a complex task, Shakey could analyze complex commands, break them down, and act accordingly.
The robot was recognized this month with an IEEE Milestone. Administered by the IEEE History Center and supported by donors, the Milestone program recognizes outstanding technical developments around the world.
From the start, Rosen and his team faced many challenges. Questions abounded. Building an intelligent mobile robot had never been attempted, so what should it look like? Should it mimic the bots found in science fiction movies and novels? How much effort should be allocated to engineering the physical robot relative to developing the artificial intelligence methods and writing the software?
In the end, Shakey looked nothing like the humanoid robots we see today. Standing a little less than 2 meters tall, it had three sections. At the bottom was a wheeled platform that gave Shakey its mobility. Atop that were what looked like three slide-in units in a rack. Those held Shakey’s onboard logic and a camera-control unit. Stacked on the uppermost unit was a TV camera and a range finder, with a radio antenna protruding from the top.That platform at the bottom held two stepping motors, one connected to each of the side-mounted drive wheels. Shakey could therefore pivot around its center, which made steering much easier to control than if it steered like an automobile. (The robot is now on display at the Computer History Museum, in Mountain View, Calif.)
As for its name, according to the history museum’s website, Rosen explained it this way: “We worked for a month trying to find a good name for it, ranging from Greek names to whatnot, and then one of us said, ‘Hey, it shakes like hell and moves around, let’s just call it Shakey.’”
From today’s perspective, Shakey looks bulky enough to contain a lot of onboard computing. In reality, there was virtually none. All the onboard electronics were for local control of sensors and motors, and for communicating with a time-shared mainframe that did the real computing.
Initially, an SDS 940 mainframe computer was used. Around 1970, a more powerful DEC PDP-10 replaced the SDS 940. The PDP-10 used a large magnetic drum memory for swapping time-shared jobs in and out of working core memory. The drum was the size of a refrigerator, held a little over 1 megabyte, and cost around US $1.5 million in today’s money.
The robot’s control software used a multilevel hierarchy. Its physical actions were at the first level, autonomous planning in the middle, and plan execution (with the ability to recover from errors) at the top. That design was adopted by subsequent robots including Stanley, a self-driving vehicle developed by researchers at Stanford University that won the 2005 DARPA Grand Challenge for driving itself 281 kilometers across California’s Mojave Desert.
Shakey’s automated planning system, STRIPS (Stanford Research Institute Problem Solver), was developed by Nils Nilsson and Richard Fikes. STRIPS composed a sequence of actions into a plan that enabled Shakey to achieve goals that were beyond the capability of its pre-programmed actions.
Researchers also developed the A* algorithm, which computes the shortest path possible through a space using the least amount of computation. Today, descendants of A* compute your driving directions, compute the paths of characters in video games, and compute the paths of Mars rovers.
The AI breakthrough was recognized on 16 February during a ceremony at the Computer History Museum in Mountain View, Calif. The bronze plaque dedicated at that event will soon be mounted in the front lobby of SRI International, in Menlo Park, Calif. The plaque reads:
Stanford Research Institute's Artificial Intelligence Center developed the world’s first mobile intelligent robot, SHAKEY. It could perceive its surroundings, infer implicit facts from explicit ones, create plans, recover from errors in plan execution, and communicate using ordinary English. SHAKEY’s software architecture, computer vision, and methods for navigation and planning proved seminal in robotics and in the design of web servers, automobiles, factories, video games, and Mars rovers.