COM2009-3009_L3_Sensing-Actuation-Control.pdf

© 2023 The University of Sheffield COM2009-3009 Robotics https://youtu.be/3xlm4nSHVx0 COM2009-3009 Robotics: Lecture 3 slide 1 1 © 2023 The University of Sheffield COM2009-3009 Robotics Lecture 3 Sensing, Actuation & Control COM2009-3009 Robotics: Lecture 3 slide 2 2 1 © 2023 The Univers...

© 2023 The University of Sheffield COM2009-3009 Robotics https://youtu.be/3xlm4nSHVx0 COM2009-3009 Robotics: Lecture 3 slide 1 1 © 2023 The University of Sheffield COM2009-3009 Robotics Lecture 3 Sensing, Actuation & Control COM2009-3009 Robotics: Lecture 3 slide 2 2 1 © 2023 The University of Sheffield What Makes Up a Robot? • Body (or bodies) – materials – morphology Computer Science • Sensor(s) => ‘perception’ • Effector(s) => ‘action’ • Control mechanism => ‘brain’ { • – intelligence – cognition – responsible for ‘behaviour’ (e.g. interaction) Energy/power source(s) – electricity – gas (pneumatic) – liquid (hydraulic) COM2009-3009 Robotics: Lecture 3 slide 3 3 © 2023 The University of Sheffield What Makes Up a Robot? body control sensing actuation environment Moore, R. K. (2016). Introducing a pictographic language for envisioning a rich variety of enactive systems with different degrees of complexity. Int. J. Advanced Robotic Systems, 13(74). COM2009-3009 Robotics: Lecture 3 slide 4 4 2 © 2023 The University of Sheffield What Makes Up a Robot? Body Control Sensing Actuation Interaction COM2009-3009 Robotics: Lecture 3 slide 5 5 © 2023 The University of Sheffield Body Body • Design principle – ‘mechanomorphic’ (machine-like) – ‘zoomorphic’ (animal-like, bio-inspired, biomimetic) – ‘anthropomorphic’ (human-like) • Materials – hard – soft • One/many (swarms) • Morphology – body layout – ‘degrees of freedom’ COM2009-3009 Robotics: Lecture 3 slide 6 6 3 © 2023 The University of Sheffield ‘Degrees of Freedom’ (DoF) Body • Equal to the number of independent parameters that define the configuration • A rigid object in 3D space has six DoF – three for positioning – three for orientation • A manipulator needs at least six joints to position an end-effector with an arbitrary orientation • Redundancy (i.e. more DoF than needed) facilitates optimization(s), such as … – energy minimization – obstacle avoidance • The human arm has seven DoF (shoulder = 3, elbow = 1, wrist = 3) • A robotic snake will have many DoF COM2009-3009 Robotics: Lecture 3 slide 7 7 © 2023 The University of Sheffield Degrees of Freedom 26 DoFs Body COM2009-3009 Robotics: Lecture 3 slide 8 8 4 © 2023 The University of Sheffield Sensing Sensing • Sensors ultrasonic distance sensor – cameras, microphones, force/pressure – infrared, ultrasound, LIDAR (“LIght Detection And Ranging”) – compass, gyroscope, GPS, etc. • Percepts – low-level (e.g. light level, colour, sound, temperature, proximity, contact, texture, smell, tilt, position, acceleration, voltage, current, etc.) – high-level (e.g. objects, people, scenes, etc.) – abstract (e.g. intentions, meanings, affective states, etc.) • External perception vs. internal proprioception • Direct perception vs. inference (e.g. orientation, effector joint angles, power level, etc.) (e.g. time-to-target using optic flow vs. object recognition using hypothesise-and-test) COM2009-3009 Robotics: Lecture 3 slide 9 9 © 2023 The University of Sheffield Actuation Actuation • Effectors – wheels, tracks, legs, wings, etc. – grippers, cutters, drills, etc. – lights, loudspeakers, lasers, etc. • Generate movement, light, sound, etc. • Action – locomotion – manipulation (e.g. gripping, welding, spraying, pushing, etc.) – signalling (simple low-level vs. coordinated high-level) • Actuators – electric motors, servos, etc. – pneumatic or hydraulic cylinders etc. – shape memory alloys, etc. A ‘mecanum’ wheel COM2009-3009 Robotics: Lecture 3 slide 10 10 5 © 2023 The University of Sheffield Interaction Interaction • Simultaneous Sensing + Actuation – interaction with the environment – interaction with other agents – interaction with itself • A robot may be able to sense its own actions – advantage … • it might be able to determine if it’s doing what it’s supposed to be doing (e.g. using its vision system to check that it’s successfully picked up an object) – disadvantage … • interference (e.g. noise from motors masks all other sounds) COM2009-3009 Robotics: Lecture 3 slide 11 11 © 2023 The University of Sheffield Control Control • The ‘brain’ of a robot • The source of its ‘intelligence’ • Responsible for its behaviour Computer Science • Concerned with … – – – – – reasoning planning learning perceiving doing • Approaches have been inspired by the changing paradigms in ‘Artificial Intelligence’ (AI) COM2009-3009 Robotics: Lecture 3 slide 12 12 6 © 2023 The University of Sheffield Control • 1940s: ‘Cybernetics’ – emphasis on control and communications Control • 1960s: ‘Computationalism’ (symbolic AI) – emphasis on mind and reasoning • ‘GOFAI’ (Good Old-Fashioned AI) 1980s: ‘Connectionism’ – mind as an emergent property of the workings of a physical machine – Parallel Distributed Processing (PDP) – Artificial Neural Networks (ANNs) • 1990s: ‘Embodied Cognition’ – – – – • relationship between brain, body, mind and world enactive systems developmental robotics morphological computing 2010s: ‘Deep (Reinforcement) Learning’ COM2009-3009 Robotics: Lecture 3 slide 13 13 © 2023 The University of Sheffield ‘Cybernetics’ “The scientific study of control and communication in the animal and the machine.” Control Norbert Wiener (1948) • The word comes from Greek: – κυβερνητική (“governor/pilot”) • Combination of … Lectures 5 & 6 – control theory – information science – biology • Example … – Grey Walter’s “Machina Speculatrix” COM2009-3009 Robotics: Lecture 3 slide 14 14 7 © 2023 The University of Sheffield Cybernetic Principles Control • Parsimony: – simpler is better (reflexes) • Exploration: – never stay still except when ‘feeding’ (recharging) • Attraction (+ve taxis/tropism): – motivated to move towards something • Aversion (-ve taxis/tropism): – moves away from negative stimuli • Discernment: – ability to distinguish between productive and unproductive behaviour COM2009-3009 Robotics: Lecture 3 slide 15 15 © 2023 The University of Sheffield GOFAI-based Robotics Control • Hierarchical “Deliberative” approach • Sense-Plan-Act (‘SPA’) – robot senses the world and constructs a global world map – robot plans all the directives needed to reach the goal – robot carries out the first directive – repeat (sensing the consequences of its action and re-planning the directives) COM2009-3009 Robotics: Lecture 3 slide 16 16 8 © 2023 The University of Sheffield Hans Moravec’s “Stanford Cart” (1979) • TV cameras took pictures of scenes • Robot planned path between obstacles • It moved in 1 metre spurts with 10-15 min. stops for image processing and planning • Successfully crossed a room full of obstacles in 5 hours ! COM2009-3009 Robotics: Lecture 3 slide 17 17 © 2023 The University of Sheffield Problems with the ‘SPA’ Approach Control • Closed world assumption – hard to include everything in the world model – huge world models – hard to keep track of all changes • Slowness – inability to react in a timely fashion • Impractical COM2009-3009 Robotics: Lecture 3 slide 18 18 9 © 2023 The University of Sheffield A Paradigm Shift Control ‘Cybernetics’ • Gradual move away from anthropocentric view and its emphasis on human intelligence • Greater awareness of abilities of non-human organisms, and their abilities to interact with and survive in the world • Idea of reactive responses to the world, instead of modelling and planning • Intelligence is determined by the dynamics of interaction with the world • Reintroduction of an old idea … “An ant, viewed as a behaving system, is quite simple. The apparent complexity of its behavior over time is largely a reflection of the complexity of the environment in which it finds itself” Herb Simon (1969) COM2009-3009 Robotics: Lecture 3 slide 19 19 © 2023 The University of Sheffield ‘Cybernetics’ Control Impact of GOFAI? Rediscovery? Kokol, P. (2018). Cybernetics: a bibliometric analysis snapshot. Cybernetics and Systems, 49(2), 95–102. COM2009-3009 Robotics: Lecture 3 slide 20 20 10 © 2023 The University of Sheffield Braitenberg’s “Vehicles” aitenberg Valentino Br 1) (1926-201 Braitenberg, V. (1984). Vehicles: Experiments in Synthetic Psychology. Cambridge, MA: MIT Press. COM2009-3009 Robotics: Lecture 3 slide 21 21 © 2023 The University of Sheffield Braitenberg’s “Vehicles” Control Sensor Body Connection Motor Vehicle 1 Braitenberg, V. (1984). Vehicles: Experiments in Synthetic Psychology. Cambridge, MA: MIT Press. COM2009-3009 Robotics: Lecture 3 slide 22 22 11 © 2023 The University of Sheffield Braitenberg’s “Vehicles” Control What will these vehicles do? COM2009-3009 Robotics: Lecture 3 slide 23 23 © 2023 The University of Sheffield Demo: Braitenberg’s “Vehicles” Control Sensor Body Connection Motor Braitenberg Vehicle-1 Vehicle 1 “Approach” COM2009-3009 Robotics: Lecture 3 Braitenberg, V. (1984). Vehicles: Experiments in Synthetic Psychology. Cambridge, MA: MIT Press. slide 24 24 12 © 2023 The University of Sheffield Demo: Braitenberg’s “Vehicles” Control Excitatory connections Vehicle 2b “Aggression” Braitenberg Vehicle-2 Vehicle 2a “Coward” COM2009-3009 Robotics: Lecture 3 Braitenberg, V. (1984). Vehicles: Experiments in Synthetic Psychology. Cambridge, MA: MIT Press. slide 25 25 © 2023 The University of Sheffield Demo: Braitenberg’s “Vehicles” Inhibitory connections Control Vehicle 3a “Love” Vehicle 3b “Explorer” Braitenberg, V. (1984). Vehicles: Experiments in Synthetic Psychology. Cambridge, MA: MIT Press. Braitenberg Vehicle-3 COM2009-3009 Robotics: Lecture 3 slide 26 26 13 © 2023 The University of Sheffield Braitenberg’s “Vehicles” Control 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. Getting around Fear & aggression Love Values & special tastes Logic Selection Concepts Space, things & movements Shapes Getting ideas Rules & regularities Trains of thought Foresight Egotism & optimism Braitenberg, V. (1984). Vehicles: Experiments in Synthetic Psychology. Cambridge, MA: MIT Press. COM2009-3009 Robotics: Lecture 3 slide 27 27 © 2023 The University of Sheffield ‘Behaviour-Based Robotics’ (BBR) • Layered “Reactive” approach • Pioneered by Rodney Brooks at MIT • New emphasis on (simple) living examples of intelligence • Threw out modelling and planning rooks Rodney B Brooks, R. A. (1991). Intelligence without representation. Artificial Intelligence, 47, 139–159. COM2009-3009 Robotics: Lecture 3 slide 28 28 14 © 2023 The University of Sheffield ‘Behaviour-Based Robotics’ (BBR) Brooks, R. A. & Flynn, A. M. (1989). Fast, cheap and out of control: a robot invasion of the solar system. Journal of The British Interplanetary Society, 42, 478-485. • Complex behaviour need not necessarily be the product of a complex control system • Intelligence is in the eye of the observer • “The world is its own best model” • Simplicity is a virtue • Robots should be cheap • Robustness in the presence of noisy or failing sensors is a design goal • Planning is just a way of figuring out what to do next • All on-board computation is important • Systems should be built incrementally • No representations, calibration, complex computation, high-bandwidth communication COM2009-3009 Robotics: Lecture 3 slide 29 29 © 2023 The University of Sheffield ‘Behaviour-Based Robotics’ (BBR) Layered ‘subsumption’ architecture (We’ll come back to this in Lecture 4) Brooks, R. (1986). A robust layered control system for a mobile robot. IEEE Journal of Robotics and Automation, 2(1), 14-23. COM2009-3009 Robotics: Lecture 3 slide 30 30 15 © 2023 The University of Sheffield ‘Behaviour-Based Robotics’ (BBR) https://youtu.be/YtNKuwiVYm0 Connell, J. H. (1988). Task Oriented Spatial Representations for Distributed Systems. PhD Thesis, MIT COM2009-3009 Robotics: Lecture 3 “Herbert” (first ‘behaviour-based’ robot) slide 31 31 © 2023 The University of Sheffield ‘Behaviour-Based Robotics’ (BBR) COM2009-3009 Robotics: Lecture 3 slide 32 32 16 © 2023 The University of Sheffield Deliberative vs. Reactive ‘sense ® plan ® act’ (deliberative) ‘subsumption’ (reactive) Brooks, R. (1986). A robust layered control system for a mobile robot. IEEE Journal of Robotics and Automation, 2(1), 14-23. COM2009-3009 Robotics: Lecture 3 slide 33 33 © 2023 The University of Sheffield Deliberative vs. Reactive Control Arkin, R. C. (1998). Behavior-Based Robotics. Cambridge, MA: The MIT Press. COM2009-3009 Robotics: Lecture 3 slide 34 34 17 © 2023 The University of Sheffield This lecture has covered … • What makes up a robot? • Body, sensing, actuation, interaction, control • Cybernetics • GOFAI robotics • Sense-Plan-Act • Braitenberg’s ‘vehicles’ • Behaviour-based robotics • Deliberative vs. reactive approaches COM2009-3009 Robotics: Lecture 3 slide 35 35 © 2023 The University of Sheffield Any Questions ? (or you can post on the ‘Lecture’ Discussion Forum) COM2009-3009 Robotics: Lecture 3 slide 36 36 18 © 2023 The University of Sheffield Next lecture … Autonomous Systems COM2009-3009 Robotics: Lecture 3 slide 37 37 © 2023 The University of Sheffield COM2009-3009 Robotics: Lecture 3 slide 38 38 19