Designing Robot Competitions That Promote AI Solutions Lessons
Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing Jeffrey R. Croxell Ross Mead Jerry B. Weinberg
Introduction • Robotics competitions are an excellent educational tool at the middle school, high school, and university levels • Gameplay of competition impacts emphasis for robot designs. . . – Rouse 2001 • DARPA, Robo. Cup, and AAAI competitions… – – – emphasize outstanding research issues in AI can be cost prohibitive requires significant human labor and team size • Other competitions typically require little use of AI – open-loop designs • Large gap between these types 2 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Introduction • Small-scale competitions can still promote AI solutions and closed-loop designs… – smaller physical size – less costly – easier to handle • Help bridge the gap from simpler to more advanced competitions 3 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Our Design • General purpose • Lynxmotion… • XBC • Light sensors • Sonar • IR rangefinders • Camera – 4 WD 1 – 5 Do. F arm 4 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
IEEE Region 5 Competition • Mini warehouse • Automated sorter • 4 colored cans placed at random locations • Sort into proper room 5 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
2006 Beyond Botball • Head-to-head • Remove “toxic waste” • Save Billy and Betty Botguy 6 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Competing Designs • Highly engineered designs • Movement based on littleto-no sensor feedback – no obstacle avoidance • Grabbed objects from known locations… – no searching • Simple and effective 7 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Competition Results • IEEE… – finalists recognized dominant strategy • Beyond Botball… – Highly engineered design → faster and more effective • Winner’s circles of these competitions consisted primarily of robots with little intelligence – more intelligent robots may be overkill for these competitions 8 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Designer’s Perspective • Annual SIUE Robotics Competitions (based on design in Martin 2001) roboti. cs. siue. edu 9 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Lessons Learned • Given game specifics, teams assume… – known layout of the arena • no mapping necessary – locations of objects and goals • no searching necessary – series of actions to achieve objectives • no localization or motion planning necessary – static arena state • closed-world assumption (Murphy 2000) 10 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
“Real Robots Don’t Drive Straight” “… students are unlikely to develop feedback-based approaches in their designs of mobile robots in contest events. ” — Martin 2007 12 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Emphasizing AI Solutions • Promote the use of closed-loop designs… – degrees of uncertainty must be included in the rules of the game • Encourage planning and re-planning based on physical interactions… – – – with the environment with objects with other robots • Intelligent decision-making takes time! – not much is offered by current competitions 13 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Opportunities in Mapping • Eliminate the closed-world assumption… – do not specify all characteristics of the field of play • Environment can be obstructive and interactive… – present robots with interesting situations and opportunities Robo. Fest www. robofest. net moveable wooden barrier 14 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
• Require robots to traverse much of the arena board… – dead reckoning is adequate for current competitions – error accumulates quickly Beyond Botball 2006 www. botball. org Opportunities in Localization • Provide a means for determining spatial position and orientation while navigating the world… – physical and visual landmarks Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing Fire. Bot Challenge roboti. cs. siue. edu 15
Opportunities in Object Recognition • Place game objectives at unspecified locations… – if the locations of these objects are given, there is no need for a search! – encourages the use of sensors to locate and approach game objects Robo. Soccer Shootout! roboti. cs. siue. edu 16 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Opportunities in Navigation and Planning • Unknown locations/order of objectives… – difficult to rely on a fixed navigational strategy • Navigation involves obstacle avoidance • More complex interactions with objects… – planning for object manipulation 17 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Opportunities in Interactions • Include obstacles! Fire. Bot Challenge roboti. cs. siue. edu – expected or unexpected – traversable or obstructive – game board itself • Allow for or encourage nondestructive interactions between competing robots… Beyond Botball 2006 www. botball. org – promotes use of sensory feedback to predict, avoid, or handle collisions 18 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Opportunities in Time • Sensing the environment and making decisions based on this feedback is a time -consuming process… – constrains AI-type strategies the most! • Provide enough time for robots to examine surroundings and calculate responses 19 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
An Example • Search-and-rescue… – 10’x 10’ arena representing an earthquake-damaged warehouse • “Blueprint” is given… – unknown conditions inside • Robots explore the arena and search for victims – all wearing red uniforms 20 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
An Example • As robot traverses arena, it must sense objects… – avoid obstacles – approach, confirm (signal), and map victims found • Dead reckoning is primary method of localization… – known colored landmarks and tone emitters for recalibration • 35º incline to second floor • 15 minute time period 21 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Conclusions • Rules must encourage AI solutions • Environment should reward sensory reaction and high-level decision-making • Important elements: – – – Unspecified dimensions Random placement Time constraints • Techniques utilized help provide an introduction to higher-level concepts and challenges 22 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
References • Arkin, R. 1998. Behavior-Based Robotics. The MIT Press. • Kumar, D. and Meeden, L. 1998. “A Robot Laboratory for Teaching Artificial Intelligence” in Proceedings of the Twenty-ninth SIGCSE Technical Symposium on Computer Science Education, D. Joyce, ed. ), Volume 30, Number 1, Pages 341 -344, ACM Press, March 1998. • Laird, J. E. and van Lent, M. 2001. “Computer Game Tutorial”, Tutorial Program at the Seventeeth International Joint Conference on Artificial Intelligence, Seattle, WA. • Martin, F. G. 2001. Robotic Explorations: A Hands. On Introduction to Engineering, Prentice Hall, Upper Saddle River, NJ. • Martin, F. G. 2007. “Real Robots Don’t Drive Straight” in Robots and Robot Venues: Resources for AI Education: Papers from the AAAI Spring Symposium, March 2007. • Mayer, G. , Weinberg, J. B. , and Yu, X. , 2004. “Teaching Deliberative Navigation Using the LEGO RCX and Standard LEGO Components”, Accessible Hands on Artificial Intelligence and Robotics Education: Working Papers of the 2004 AAAI Spring Symposium Series, March 2004. • Miller, D. and Stein, C. 2000. “‘So That’s What Pi is For!’ and Other Educational Epiphanies from Hands on Robotics”, Robots for Kids: Exploring New Technologies for Learning, A. Druin and J. Hendler, (Eds. ), Morgan Kaufmann, pp. 220 -243. • Murphy, R. 2000. An Introduction to AI Robotics. The MIT Press. • Rouse, R. 2001. Game Design: Theory and Practice. Wordware Publishing, Inc. , Plano, TX. • Weinberg, J. B. , W. White, C. Karacal, G. Engel, & A. Hu, “Multidisciplinary Teamwork in a Robotics Course”, The 36 th ACM Technical Symposium on Computer Science Education, February 2005, pp. 446 -450. 23 Designing Robot Competitions That Promote AI Solutions: Lessons Learned Competing and Designing
Questions?
- Slides: 23