Artificial Intelligence and Sensor Technologies

SRI's vast talent and expertise in artificial intelligence, algorithm development, and sensors research have yielded a prolific environment for the development of disruptive technologies that are being incorporated in ever-evolving robot platforms.

One of many SRI Centibots

LAGR (2008)

SRI's Artificial Intelligence Center includes some of the world's top experts in this field. Read about the latest achievements, recognition and projects of our AI team, and view a complete timeline of AI Center projects.

Our Robots: History and Future

SRI's activities in intelligent systems began more than 50 years ago, with a focus on learning machines and self-organizing systems. Following is an overview of SRI's robots in the artificial intelligence arena, from Shakey in the 1960s to today's LAGR, which can orient itself and navigate autonomously off-road.

Shakey (1966-1972)

Flakey (1982-1995)

LURCH (1993-2001)

ERRATIC and Pioneer (1994-present)  

Small Vision Module (1995-present)

Centibots (2002-2004) 

LAGR (2006-present)

Shakey

(1966-1972)

Shakey, the first mobile robot with the ability to reason and react to its environment, has had a substantial influence on present-day artificial intelligence and robotics.

Using a TV camera, a triangulating range finder, and bump sensors, Shakey was connected to DEC PDP-10 and PDP-15 computers via radio and video links. Interoperating programs with varying levels of sophistication provided Shakey the ability to combine simple movements and environmental perception into robust, complex tasks, enabling it to achieve goals given by a user. The system also generalized and saved these plans for possible future use. Inducted into the Robot Hall of Fame in 2004, Shakey is today on display at the Computer History Museum in Mountain View, California.

Learn more about Shakey here.

Flakey

(1982-1995)

Flakey was a custom-built platform approximately three-feet high and two feet in diameter operating within SRI’s own office environment. Two independently driven wheels provided a maximum velocity of about two feet per second. Flakey’s sensors included a ring of 12 sonar range finders, wheel encoders, and a video camera used in combination with a laser to provide dense depth information over a small area in front of the vehicle. Flakey’s onboard computers included a workstation and other processors dedicated to sensor interpretation, motor control, and radio communications.

Learn more about Flakey here.

LURCH

(1993-2001)

LURCH (Large, Useful Robot Controlling Hazards) is designed for control of robot operations in realistic outdoor terrain and is controlled using high-level directives from a remote operator control station connected via a packet-switched radio network. LURCH was created by modifying an Andros Mark V-A robot from Remotec, Inc. to incorporate SRI’s planning and control system. Enhancements include onboard control of the mobile base and a manipulator arm based on a ruggedized PC system. Onboard sensors include stereoscopic vision, 16 ultrasonic sensors, and encoders for all robot, manipulator, and camera motions.

Learn about applications of LURCH in field robotics here.

ERRATIC and Pioneer

(1994-present)

Addressing the need for an easy-to-construct, low-cost robot development platform, SRI designed ERRATIC to run as a robot server from a host computer over a remote serial connection. It provides basic functions of forward/back velocity and angular position integration, stall sensing, and sonar ranging. Pioneer I is a production version of the ERRATIC platform. The real-time controller for ERRATIC (Saphira) is based on software developed at SRI International on the Flakey project. The software runs a reactive planning system with a fuzzy controller, behavior sequencer, and deliberative planner with integrated routines for sonar sensor interpretation, map building, and navigation.

Small Vision Module

(1995-present)

In a project cosponsored by DARPA, SRI developed an inexpensive, compact, self-contained, low-powered range sensor for mobile robots and surveillance applications. The small device measures only six cubic inches and combines commercially available components to perform real-time analysis of stereo images in a very small footprint, using very little power. SRI is developing the next generation of this device, which will feature nearly a 600-percent performance improvement.

Centibots

(2002-2004)

The Centibots are mobile, coordinated robots that can autonomously and effectively explore, map, and survey the interior of unknown building structures. The Centibots marked a milestone in robotics, representing the largest collection (more than 100) to date of coordinated autonomous mobile robots. These autonomous team robots were designed to augment the situational awareness of human teams—such as crisis response teams—in situations that could pose a threat to people. Centibots improve upon current robot architectures, which rely on large, power-hungry subsystems for mobility, communication and control, and are limited to only individual or small teams of robots.

Learn more about the history of Centibots here.

LAGR
(2005-present)

Real-time vision and learning technologies are at the core of the DARPA Learning Applied to Ground Robotics (LAGR) program to develop autonomous off-road navigation. The goal is to develop sensing- and-camera-based techniques for learning the mobility properties of objects in a new environment and planning and control techniques for using this information to avoid such difficulties as loose sand, bushes, and cul-de-sacs. SRI developed color-and-texture-based techniques for learning and recognizing paths and obstacles; a real-time, stereo-based visual odometry technique for precisely locating the robot as it moved through complex outdoor environments; mapping of features for later runs; and very efficient, low-level control techniques so the robot could rapidly traverse planned paths and quickly free itself.

Learn more about LAGR here.