CS 344R: Robotics CS 393R: Autonomous Robots Lecture 1: • Professor Benjamin Kuipers Introduction to the Course – 471-9561, kuipers@cs.utexas.edu – Office hours: TTh 10:00 - 11:00, CSA 1.120A • TA: Jeremy Stober (jstober@cs.utexas.edu) CS 344R: Robotics – Office hours: TBD, ENS 19N CS 393R: Autonomous Robots • Robot lab: ENS 19N Benjamin Kuipers • Wiki: http://z.cs.utexas.edu/wiki/cs344r.wiki Are these robots? Are these robots? Are these robots? What is a robot? • A robot is an intelligent system that Robot interacts with the physical environment sensors effectors through sensors and effectors. Environment – Program module? – Web crawling ‘bot? 1
Is a human a robot? Robbie • By our definition, yes. • from Forbidden Planet , 1956. – Humans interact with a complex physical environment via sensors and effectors. – We are not artificially manufactured, of course! • Does this diminish humans? No! – Understanding the difficulties of robotics helps us appreciate how amazing humans are. We will study robots that … What makes a good model of the environment? • … function in (mostly) unmodified human • A good model is a simplified description of environments. the environment such that … – (Well, in soccer fields, anyway.) – If the robot orients itself in the model , – and makes a plan using the model , • … that use, and perhaps even learn, useful – and executes that plan in the real environment , models of the environment. – then the plan has its intended effect. – They have knowledge , and act on it. Subject Material Areas What will we do in this course? • Our goal is to learn • Artificial Intelligence some methods for Robot • Computer Vision implementing this • Control Theory interactive loop. sensors effectors • Bayesian Probability Theory • We will spend a few • Mechanical Engineering Environment weeks each on topics • Cognitive and Developmental Psychology that often get entire graduate courses. 2
Control Laws and Behaviors Major Topics and Projects • Rules for behaving in a qualitatively uniform environment. • What is robotics? • “Hello, World!” (9/16) – Following walls, seeking open space or targets. • Control laws • Motor control (10/7) • Rich theory based on differential equations • Behavior architectures • Learning skill (10/28) and dynamical systems. • Observers and tracking • Metrical maps (11/13) • Reality outside the model is treated as noise. • Local metrical mapping • Localization (12/2) • Compose multiple control laws to make • Topological mapping behaviors. • Social implications • Grad student projects • Task: Approach and kick a ball to a target. – Learn to do it more accurately. Local Metrical Mapping Observers • Sensors don’t sense the world directly. • A map of the local environment is useful for local motion planning. – They just respond to its stimulation. • By gathering lots of sensor input over time, • Range sensors give distance to obstacles. we can estimate what the world is like. – Laser rangefinder is more accurate than sonar • Assumes models of the nature of the world, • Combine sensor returns to find obstacles. and of sensor properties, such as error types. • Robot must localize itself. • Task: Implement Kalman Filters to track • Tasks: Implement occupancy grid mapping. and block a rolling ball. – Next: Implement localization and SLAM. Topological Mapping Social Implications and Planning • Robots may change our world dramatically • Abstract local regions to “places”. – How? For better? Or for worse? • Abstract travel actions to “paths”. • Science fiction writers have thought about a lot of important possibilities. • Model the environment as a graph. • We will watch and discuss relevant clips • Transforms action planning to graph search. from movies and television shows. • Plans can be translated back to actions, and – Brief discussions. Few conclusions. to control laws. – Questions are more important than answers. 3
Robot Lab Assignments Robot Assignments 1, 2, 3 • There are five robot lab assignments. • Students will work in teams. – Due about every three weeks. – Each team has three people (10 teams). – (Once, it was six, due every two weeks!) – A single grade for each team. • Each team has one physical robot. • You demonstrate the techniques taught in class. – These are expensive, fragile, and irreplaceable ! – “ In theory, there’s no difference between theory and practice, but in practice, there is .” – Take care of them! Robot Assignments 4, 5 Previous robot: the Amigobot • Sonar sensors: • Students will work individually. front (6), back (2) – Each person gets their own grade. • Camera • Passive gripper • The “robot” is a recorded sensor trace. • Differential drive – A robot explores an area, using laser range-finder (right/left wheel) and measuring odometry. • Odometry • Build a map, given correct odometry. • Wireless communication – Then do simultaneous localization and mapping. Demo in the old Robot Lab This year: the Sony AIBO • Better sensors • More degrees of freedom • Onboard computing 4
Technical Details Entertainment Robot System 7 • CPU: 64 bit RISC • Image input: • Sony designed the AIBO as an entertainment – 64 mb RAM – 350,000 pixel CMOS robot, with sophisticated built-in behaviors. camera • LAN: 802.11b • Stereo microphones – We won’t be using those. • Degrees of freedom: • Infrared distance x 2 – You are welcome to explore them, but that’s not – Head: 3 dof part of the course. • Acceleration – Mouth 1 dof • We are using the AIBO as a platform for – Legs: 3 dof x 4 • Vibration – Ears: 1 dof x 2 implementing robotic capabilities. • Touch: head, back, – Tail: 2 dof chin, paw Shooting and Blocking Shooting and Blocking What Assignments Require An Illegal Strategy • The point of the assignments is to implement the methods taught in class. • To turn in an assignment: – Demonstrate the behavior to Jeremy before the due date. – Each team hands in a clear, concise memo describing the problem, your approach, and your results. • Append the code. – The memo describes the role of each individual on the team in accomplishing this assignment. • We will discuss each assignment in class on the due date. – Some teams will be selected to demonstrate the robots. – No assignments accepted after that class meeting. 5
Working in Teams Term Projects (CS 393R only) • Research one topic in greater depth. • One of the goals of this course is to give you experience at working in teams. • Select a topic (suggestions to be provided). – Robot assignments 1, 2, and 3. • Survey the related literature. • Your team can be stronger than any one • Describe the alternate approaches individual, but it is also vulnerable. – Discuss their strengths and weaknesses. • You are responsible for working effectively • Design and justify a project to advance the field. with your team – A novel experiment to discriminate approaches – not just for doing your own job, but also – A novel approach (and experiment) – for helping the team work well together. – A toolkit to build on mature successful methods Grading (344R/393R) Grading (344R/393R) 344R 393R • Exams (individual) • Robot Assignments Assignments 12% 8% – Mid-term 1 (15/15%) – Hello, World! (12/8%) 12% 8% October 14 – Motor control (12/8%) – Mid-term 2 (15/15%) 12% 8% – Learning skill (12/8%) November 25 12% 8% – Metrical maps (12/8%) 12% 8% – Localization (12/8%) • Projects (0/20%) Project 20% • These are never – CS 393R only Midterm exam 1 15% 15% accepted late! – Proposal (9/30) Midterm exam 2 15% 15% – Literature (10/21) Participation 10% 10% – Methods (11/6) • Participation (10/10%) Total 100% 100% – Research plan (12/4) This class is a lot of work. Robotics • Robotics includes many different concepts. • The topic is fundamentally important – Control theory, logic, probability, search, etc. scientifically and technologically. • Abstraction barriers are very strong in most – Building intelligent agents of Computer Science, but weak in Robotics. – Modeling the phenomenon of mind – Programs are vulnerable to sensor and motor • It will be very demanding on all of us. glitches. – Be prepared, and start work early. • Plan ahead, to put the time in to this course. • It’s also very exciting and lots of fun! – Your team will be depending on you. 6
Recommend
More recommend
