COMP150 Behavior-Based Robotics

Transcription

1 For class use only, do not distribute COMP150 Behavior-Based Robotics

2 Course Essentials This is not a standard lecture-type course, where the instructor flips slides and you take notes... This course will be essentially student-driven and project-based (there are great models of how instructors when acting as facilitators rather than lecturers can speed up student learning and help students retain new information better), hence You will get out of this course what you put into it both in terms of effort and time! Before you consider taking the course, ask yourself: what do I want to get out of it? Do not take this course unless you are willing to try hard to get robots to exhibit interesting behaviors If you decide to take it under these premises, I promise you that you will find it very rewarding and that you will learn a lot!

3 Rough Structure Much of the detailed course structure will depend on you, i.e., on your particular interests and projects Overall, we will go through the course text fairly quickly There will be regular readings assigned that everybody is expected to complete before class (otherwise you will have no basis for the discussion in class note that I will simply assume that everybody has read the chapters!) We will also read and discuss additional seminal papers in BBR and embodied situated cognitive science While I will start with lectures in the beginning (to give you the historic context, the basic ideas and principles, and the goals of BBR), there will be less lecturing and more discussions as the semester progresses

4 Rough Structure Cont'd Fairly early on, we will decide on what projects to purse based on your interests I am a big believer in collaborative work, so I would like to encourage the formation of groups to jointly work on projects At some point in late September/early October, you will get a detailed tutorial on the available software package for the robot (it will be helpful to know the basics of JAVA by then) We'll start with simple exercises and examples of behaviorbased systems and robotics architectures on simulated robots in simple simulated environments right away! Later, we will move to more faithful simulations of robots and environments, and ultimately we will try out our ideas on real robots everybody has to do a project on a real robot!

5 Rough Structure Cont'd Later in the semester, graduate students are expected to present a research paper to the class, preferably one that is being considered for or used in their project, and the class will then judge, discuss, and grade it We will also have regular very brief progress reports from each group on the status of their project later in the semester The final in the course is a research paper (possibly one that could be submitted to a conference somewhere) note that final papers will be individual papers! To avoid last minute stress, we will start working on the paper already during the second half of the semester (effectively, I would like to turn the progress reports slowly into papers) You will be provided with presentation and paper guidelines as well as paper templates

6 How to get help First and foremost, try hard yourself (this can be often the most rewarding if you figure it out on your own!) If things still don't work and/or you are still not able to figure something out or make sense of it, do the following, roughly in this order: Tip 1: consult the course web pages to see if there is information you might have missed ( Tip 2: talk to your classmates, they might be able to help Tip 3: contact the instructor -- is the best option!) or the teaching assistant especially for questions about ADE, your code or the robots Tip 4: if none of the above work, set up a meeting with the instructor or TA (again, use !)

7 Outline for the rest of this week's sesions Administrativa Questions? (move to Mondays?) JAVA crash course (tips on how to get started) First assignment (include a brief overview of the ADE system) Historical perspectives on robots and behavior-based robotics

8 Info and Materials on JAVA Best online resources from SUN, the source of JAVA! Online tutorial: Getting started on Linux: Getting started on Windows: Getting started with the NetBeans IDE: If you haven't programmed in an OO language, read For the first assignment, you will only need these basics: But make sure to read the rest of JAVA language description:

9 A Simple, Simple Example /** * This class implements my first attempt at running and compiling JAVA */ class MyFirstJavaClass { int myonlyvariable; public MyFirstJavaClass() { System.out.println("In the constructor..."); myonlyvariable = 1; } } public static void main(string[] args) { MyFirstJavaClass mfjc = new MyFirstJavaClass(); System.out.println("Class created!"); System.out.println("Attempt " + mfjc.myonlyvariable); // print it }

10 ADE: The Quick View ADE (agent development environment), an intelligent distributed real-time MAS infrastructure for robotic systems

11 ADE: The Quick View The basic component in ADE is a server (i.e., a computational unit representing a functional architectural component on a possibly computationally independent resource); for example: Robot server: a representation of a virtual or robotic agent within ADE with access to its sensors and effectors Laser server: a component that processes and provides information from a laser range finder sensor ADESim server: a 2D simulated environment that provides a robot and a laser server GUI server: a graphical user interface that allows the user to view all the servers in a running ADE system ADE registry: the main ADE control component that manages all servers (including authentication, access control, monitoring, error detection, etc.)

12 The robots are here...

13 The robots are here... The United Nations identified personal service robots, robots, which for the first time outnumbered industrial robots in 2005, as having the highest expected growth rate, i.e., to quadruple their number by 2008 (United Nations, 2005) Such robots are envisaged to help the elderly (Hirsch et al., 2000) and support humans in the house (Breemen, Yan, & Meerbeek, 2005) Popular robots include Aibo (Sony, 1999), Nuvo (ZMP, 2005) and Robosapien (WowWee, 2005) Robosapien has sold approximately 1.5 million units by January 2005 (Intini, 2005) What all these robots have in common is that they interact, in one way or another, with humans (be it for entertainment purposes or to achieve a task) No need to buy insurance yet...

14 The robots are here...

15 Early robots de Vaucanson's quacking duck 1738 Water clock 250 B.C. Da Vinci's humanoid 1495

16 Other old robots Jacquard' textile machine 1801 Writing robot 1773 Trumpetteer 1810

17 Not so old robots UNIMATE, first industrial robot by General Motors (1961) Walter Grey's Tortoise (1948/49) ELEKTRO could walk and talk (1937/38)

18 More modern robots Pathfinder (1997) Robot Tuna MIT (1996) Honda P2 (1997)

19 Sporty and entertaining robots

20 Robot musicians

21 Emotional Robots

22 Humanoids

23 Androids

24 What is Behavior-Based Robotics? Behavior-based robotics (BBR) bridges the fields of artificial intelligence, intelligence, engineering, engineering, and cognitive science. science. The behaviorbased approach is a methodology for designing autonomous agents and robots; robots; it is a type of intelligent agent architecture. architecture. Architectures supply structure and impose constraints on the way robot control problems are solved. The behavior-based methodology imposes a general, biologically inspired, inspired, bottom-up philosophy, philosophy, allowing for a certain freedom of interpretation. Its goal is to develop methods for controlling artificial systems (usually physical robots, but also simulated robots and other autonomous software agents) and to use robotics to model and better understand biological systems (usually animals, ranging from insects to humans). (from Maya Mataric' article in the MIT Encyclopedia of Cognitive Science)

25 Artificial Intelligence, Cognitive Science, and the Behavior-Based Approach To spell out what the BBR approach is an answer to, we'll first have to understand what behavior, and in particular, intelligent behavior could mean then we'll take a look at disciplines that deal with intelligence (i.e., cognitive science and artificial intelligence) and try to understand what models of intelligent behavior are So, what is intelligent behavior? What is intelligence? What does intelligent mean? (from Latin: interleggere meaning to understand, lit. to read between [the lines] ) When is a system intelligent? (ontological questions) How do we know when a system is intelligent? (epistemological question)

26 Artificial Intelligence, Cognitive Science, and the Behavior-Based Approach Question: intelligent = cognitive? E.g., is it possible to have intelligent systems that are not cognitive? Much of the initial BBR momemtum was based on this idea! Common sense notions include thinking, problem solving, learning and memory, emotions, consciousness, intuition and creativity, language capacity, etc. One possible answer: whatever the IQ test measures (cp. to the Bell Curve by Herrnstein and Murrey, or the recent discussions on the G-factor) Problem: limits intelligence to humans (this would not be unheard of, Descartes effectively claimed that only humans had souls and that other animals were possibly sophisticated automata)

27 Artificial Intelligence, Cognitive Science, and the Behavior-Based Approach Others attempts at defining intelligence : the ability to carry on abstract thinking the ability to adapt oneself to the environment the capacity to acquire capacity the capacity to learn or profit by experience a biological mechanism by which the effects of a complexity of stimuli are brought together and given a somewhat unified effect in behavior Turing's approach: avoid definition of intelligence (Turing test what does it measure?) Also: emotional IQ, street smart, (sporadic) good thinking, etc.

28 Artificial Intelligence, Cognitive Science, and the Behavior-Based Approach What about animals? Are they intelligent? How can we measure their intelligence? We assume there to be some sort of ordering relation... Maybe there is no general answer to the question (maybe intelligence is a cluster concept...) So, we'll take the following theoretical approach: first of all, avoid answer and focus on behavior, i.e., of behavior in particular circumstances (work with operational definition of intelligence) second, we are interested in how a given behavior can come about, how it can be implemented or how it can emerge from other behaviors (contrast different behaviors, different implementations, for different tasks in different environments)