sourced story
1966-1972Peer-reviewed · 2 sourcesWell documented

Shakey becomes the first robot that reasons about its actions

SRI's four-layer robot plans routes, moves blocks, and recovers from its own mistakes

On the timeline · around 1966-1972 · Symbolic AI, Booms and WintersFoundationsSymbolic AI, Booms and WintersShakey becomes the first robot that reasons about its actions19601965197019751980

Quick facts

Institution
Stanford Research Institute (SRI)
Key researchers
Charles Rosen, Nils Nilsson, Peter Hart
Planning system
STRIPS

What happened

Researchers at the Stanford Research Institute, led by Charles Rosen, Nils Nilsson, and Peter Hart, built Shakey, a wheeled robot that operated in a set of rooms containing blocks and ramps. Shakey used a camera and bump sensors to perceive its surroundings, and a planning program called STRIPS (Stanford Research Institute Problem Solver) to break a goal such as 'push block 9 to doorway 4' into a sequence of achievable actions. If an action failed partway through, Shakey could re-plan around the failure rather than simply stopping. Its control software was organized into layers, from low-level motor control up to high-level planning, an architecture that later became a template for other robots. It was not fast or practical: a single task could take Shakey hours to complete.

Why it matters

Shakey was the first machine to combine perception, planning, and physical action in one system, and the STRIPS planning language it used influenced automated planning research for decades afterward.

How we know

SRI International and the Computer History Museum, which preserves the physical robot, both document Shakey's capabilities, the STRIPS planner, and its layered control architecture.

Sources

See something wrong? . Corrections with a source get fixed fastest.

Part of a timelineArtificial Intelligence30 events · From a wartime theory of neurons to machines that write, paint, and fold proteinsView all →