Robo sapiens Robo sapiens The Forefront of AI? CPSC 433 Christian Jacob Dept. of Computer Science Dept. of Biochemistry & Molecular Biology University of Calgary Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction Challenges Challenges of the Physical W orld of the Physical W orld � 2 � • Non � episodicity The physical world we live in is very demanding • E � ects of an action change over time. � especially for robots. • Dynamics • Inaccessibility • Robots need to decide when to � think � and when to • Arti � cial sensors are still imperfect � or too perfect � . act immediately. • Nondeterminism • Continuity • A robot needs to deal with uncertainty. • States and actions are drawn from a continuum of physical con � gurations and motions. Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction Motion Planning • Fine � motion � bounded � error � planning • Assumes bounds on sensor and actuator uncertainty. What are Robots Good For? � Partial knowledge of the environment is known to the system. � Most of the planning is done o � ine. � Used for planning small, precise motions of assembly robots. Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction
Manufacturing and Materials Fine � Motion Planning Handling Assembly Robot Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction Navigation & Navigation Motion Planning • Cell / map decomposition • Landmark � based navigation • Break continuous space into a � nite number of • This method assumes that some regions exist in discrete search problems. which the robot location can be pinpointed using landmarks. • Skeletonization • Outside those regions there may only be orientation • Compute a uni � directional � skeleton � � subset � of the information. con � guration space � ==> graph search problem � Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction Landmark Navigation Gofer Service Robots Bell & Howell Mail Mobile Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction
Navigation & Exploration Motion Planning • Online Algorithms • The robot makes decisions at run time � no need for o � ine planning � . • This method assumes that the environment is completely unknown, and has to be � explored � . • The robot is usually equipped with a position sensor and knows the location of its goal. Nomad � Carnegie Mellon University � Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction Exploration Online Algorithm Text Mars Path � nder Sojourner http://www � 2.cs.cmu.edu/ � lri/nav97.html Rover Mars Path � nder Mars Mission 2004 Sojourner Nomad � CMU, 1997 � Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction More Robot Examples Hazardous Environments Useful and curious ... Lunokhod Moon Robot Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction
Hazardous Environments Telepresence Dante II Frame W alking Robot Wheelbarrow : Bomb Disposal Robot Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction Telepresence Telepresence ARTEMIS : Advanced Robot & Telemanipulator Advanced Tethered V ehicl � System for Minimal Invasive Surgery Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction Augmentation of Human Augmentation of Human Abilities Abilities Sigourney W eaver in the movie Aliens Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction
What Makes Robots W ork? What Makes Robots W ork? • Sensors: tools for perception • E � ectors: tools for action • Proprioception • Locomotion • Force sensing • Manipulation • Tactile sensing • ... Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction Locomotion Proprioception robots/mit_robots.html robots/mit_robots.html http://www.computerthai.or.th/Technology/MIT http://www.computerthai.or.th/Technology/MIT MIT � s 3D Hopper � 1984 � Spring Flamingo � MIT � Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction Proprioception Proprioception robots/mit_robots.html robots/mit_robots.html http://www.computerthai.or.th/Technology/MIT http://www.computerthai.or.th/Technology/MIT Spring Flamingo � MIT, 1996 � 2000 � M2, Bipedal W alker � MIT, 2001 � present � Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction
Haptics What Makes Robots W ork? • Sensors: tools for perception • Proprioception • Force sensing • Tactile sensing • Sonar • Camera, visual perception MIT T ouch Lab Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction Sonar Sensors Camera Sensors Peoplebots � ActivMedia � Fast Eye Gimbals � MIT � Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction Kisme � Robots in our � Daily Lives � MIT Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction
Robot Teams LOGO Robot Seymour Papert AIBO � Sony � and RoboCup Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction W ABOT � 2 : Organ Player http://cdecas.free.fr/robots/wabot.php In 1984, Waseda University presented the Wabot-2. This machine was the first attemps of specializing a robot in domestic use. The chosen activity was music, and the Wabot-2 got worldwide famous as the first robot in the world which played piano. Ichiro Kato, W aseda University, Tokyo � 1980 � 84 � Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction Lawn Mowing ... V acuum Robot Roomba Robotic Floor V ac Friendly Robotics: Robomow Christian Jacob, University of Calgary Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction CPSC 433 � Arti � cial Intelligence: An Introduction
References • Menzel, P ., and D � Alusio, F. � 2000 � . Robo sapiens � Evolution of a New Species . Cambridge, MA, MIT Press. • Kurzweil, R. � 1990 � . The Age of Inte � igen � Machines . Cambridge, MA, MIT Press. Christian Jacob, University of Calgary CPSC 433 � Arti � cial Intelligence: An Introduction
Recommend
More recommend