Learning to Coordinate Behaviors Pattie Maes & Rodney A. Brooks Presented by: Javier Martinez.

Slides:



Advertisements
Similar presentations
Algorithm Design Techniques
Advertisements

Heuristic Search techniques
Planning
Chapter 5 Plan-Space Planning.
0 Solving Problems in Groups ©2008, University of Vermont and PACER Center Solving Problems in Groups PCL Module 9.
Problem Solving.
Evolving Edge detection Final project by Rubshtein Andrey ( )
Elephants Don’t Play Chess
Inconsistent Heuristics
Distributed Constraint Optimization Problems M OHSEN A FSHARCHI.
MOEAs University of Missouri - Rolla Dr. T’s Course in Evolutionary Computation Matt D. Johnson November 6, 2006.
Mining Compressed Frequent- Pattern Sets Dong Xin, Jiawei Han, Xifeng Yan, Hong Cheng Department of Computer Science University of Illinois at Urbana-Champaign.
SE263 Video Analytics Course Project Initial Report Presented by M. Aravind Krishnan, SERC, IISc X. Mei and H. Ling, ICCV’09.
COS 461 Fall 1997 Routing COS 461 Fall 1997 Typical Structure.
Comparing Effectiveness of Bioinspired Approaches to Search and Rescue Scenarios Emily Shaeffer and Shena Cao 4/28/2011Shaeffer and Cao- ESE 313.
an incremental version of A*
Dynamic Bayesian Networks (DBNs)
Rumor Routing in Sensor Networks David Braginsky and Deborah Estrin Presented By Tu Tran 1.
1 Sensor Relocation in Mobile Sensor Networks Guiling Wang, Guohong Cao, Tom La Porta, and Wensheng Zhang Department of Computer Science & Engineering.
Evolutionary Computational Intelligence Lecture 10a: Surrogate Assisted Ferrante Neri University of Jyväskylä.
1 Learning to Detect Objects in Images via a Sparse, Part-Based Representation S. Agarwal, A. Awan and D. Roth IEEE Transactions on Pattern Analysis and.
Sandra Wieser Alexander Spröwitz Auke Jan Ijspeert.
Modern Information Retrieval Chapter 2 Modeling. Can keywords be used to represent a document or a query? keywords as query and matching as query processing.
1 A Novel Binary Particle Swarm Optimization. 2 Binary PSO- One version In this version of PSO, each solution in the population is a binary string. –Each.
Planning operation start times for the manufacture of capital products with uncertain processing times and resource constraints D.P. Song, Dr. C.Hicks.
Universal Plans for Reactive Robots in Unpredictable Environments By M.J. Schoppers Presented by: Javier Martinez.
Classical Planning Chapter 10.
Fast Failover for Control Traffic in Software-defined Networks Globecom 2012 Neda B. & Ying Z. Presented by: Szu-Ping Wang.
Prepared by Barış GÖKÇE 1.  Search Methods  Evolutionary Algorithms (EA)  Characteristics of EAs  Genetic Programming (GP)  Evolutionary Programming.
Distributed Constraint Optimization Michal Jakob Agent Technology Center, Dept. of Computer Science and Engineering, FEE, Czech Technical University A4M33MAS.
06 - Boundary Models Overview Edge Tracking Active Contours Conclusion.
Artificial Intelligence Chapter 2 Stimulus-Response Agents
Artificial Intelligence Lecture 9. Outline Search in State Space State Space Graphs Decision Trees Backtracking in Decision Trees.
Stochastic Routing Routing Area Meeting IETF 82 (Taipei) Nov.15, 2011.
Chapter 11 – Neural Networks COMP 540 4/17/2007 Derek Singer.
What are the main differences and commonalities between the IS and DA systems? How information is transferred between tasks: (i) IS it may be often achieved.
Legged Robot Locomotion Control  Legged Robot Locomotion Control  CPG-and-reflex based Control of Locomotion.
A Framework for Elastic Execution of Existing MPI Programs Aarthi Raveendran Graduate Student Department Of CSE 1.
A Survey of Distributed Task Schedulers Kei Takahashi (M1)
Robotica Lecture 3. 2 Robot Control Robot control is the mean by which the sensing and action of a robot are coordinated The infinitely many possible.
Distribution of Student Mistakes between Three Stages of Solution Steps in Case of Action-Object-Input Solution Scheme Dmitri Lepp University of Tartu.
Data Mining Practical Machine Learning Tools and Techniques Chapter 4: Algorithms: The Basic Methods Section 4.6: Linear Models Rodney Nielsen Many of.
A User-Lever Concurrency Manager Hongsheng Lu & Kai Xiao.
Autonomic scheduling of tasks from data parallel patterns to CPU/GPU core mixes Published in: High Performance Computing and Simulation (HPCS), 2013 International.
A Study of Balanced Search Trees: Brainstorming a New Balanced Search Tree Anthony Kim, 2005 Computer Systems Research.
A new Ad Hoc Positioning System 컴퓨터 공학과 오영준.
Optimal Resource Allocation for Protecting System Availability against Random Cyber Attack International Conference Computer Research and Development(ICCRD),
Introduction to Planning Dr. Shazzad Hosain Department of EECS North South Universtiy
Decision Trees Binary output – easily extendible to multiple output classes. Takes a set of attributes for a given situation or object and outputs a yes/no.
Presented By, Shivvasangari Subramani. 1. Introduction 2. Problem Definition 3. Intuition 4. Experiments 5. Real Time Implementation 6. Future Plans 7.
Bundle Adjustment A Modern Synthesis Bill Triggs, Philip McLauchlan, Richard Hartley and Andrew Fitzgibbon Presentation by Marios Xanthidis 5 th of No.
1 Choosing a Computer Science Research Problem. 2 Choosing a Computer Science Research Problem One of the hardest problems with doing research in any.
Laboratory of mechatronics and robotics Institute of solid mechanics, mechatronics and biomechanics, BUT & Institute of Thermomechanics, CAS Mechatronics,
Top-K Generation of Integrated Schemas Based on Directed and Weighted Correspondences by Ahmed Radwan, Lucian Popa, Ioana R. Stanoi, Akmal Younis Presented.
Presented by: Idan Aharoni
A framework of safe robot planning Roland Pihlakas Institute of Technology in University of Tartu august 2008.
Adaptive Triangular Deployment Algorithm for Unattended Mobile Sensor Networks Ming Ma and Yuanyuan Yang Department of Electrical & Computer Engineering.
Vision-Guided Humanoid Footstep Planning for Dynamic Environments P. Michel, J. Chestnutt, J. Kuffner, T. Kanade Carnegie Mellon University – Robotics.
KAIS T Sensor Deployment Based on Virtual Forces Reference: Yi Zou and Krishnendu Chakarabarty, “Sensor Deployment and Target Localization Based on Virtual.
Control-Theoretic Approaches for Dynamic Information Assurance George Vachtsevanos Georgia Tech Working Meeting U. C. Berkeley February 5, 2003.
Robot Intelligence Technology Lab. 10. Complex Hardware Morphologies: Walking Machines Presented by In-Won Park
1 Last Time Buys and Reuse at Océ A last time buy decision tool for reusable service parts. Anke Verbaarschot 2011, October 19 th.
Bell Ringer: 8/17/15  Solve the following equation:
Sudoku Solutions Using Logic Equations Christian Posthoff The University of The West Indies, Trinidad & Tobago Bernd Steinbach Freiberg University of Mining.
4/22/20031/28. 4/22/20031/28 Presentation Outline  Multiple Agents – An Introduction  How to build an ant robot  Self-Organization of Multiple Agents.
Vision-Guided Humanoid Footstep Planning for Dynamic Environments
Deep Feedforward Networks
Distance Computation “Efficient Distance Computation Between Non-Convex Objects” Sean Quinlan Stanford, 1994 Presentation by Julie Letchner.
Today: Classic & AI Control Wednesday: Image Processing/Vision
Artificial Intelligence Chapter 2 Stimulus-Response Agents
Presentation transcript:

Learning to Coordinate Behaviors Pattie Maes & Rodney A. Brooks Presented by: Javier Martinez

Introduction Behavior-based system Learning using positive and negative feedback Behaviors decide when is time to activate Distributed algorithm Test the concept in a robot

Motivation Behavior control is a weak point initial Behavior-based systems Behavior control has to be prewired This approach doesn’t scale too well

New Ideas Behavior control is learned through experience Learning algorithm completely distributed Each behavior learns when to become active The solution maximizes positive feedback and minimizes negative feedback

The Learning Task What is needed: Vector of binary perceptual conditions Set of behaviors Positive feedback generator Negative feedback generator

The Learning Task The task: Change the precondition list from each behavior to maximize relevance and reliability

The Learning Task Constraints: Relevance: behavior correlated to positive feedback, not correlated with negative feedback Reliability: behavior receives consistent feedback

The Learning Task More constraints: Algorithm should deal with noise, Perform in real time, Support readaptation

The Learning Task Assumptions: At least one combination of preconditions is bounded Feedback is immediate Only combinations of conditions can be learned

Algorithm Measure: Number of times a positive/negative feedback did/didn’t happen when a behavior was/wasn’t active Calculate the correlation between positive/negative feedback and the status of the behavior

Algorithm Measure: Express relevance and reliability in terms of this correlation Relevance controls whether a behavior should be active or not Reliability decides whether the behavior should try to improve itself

Algorithm Measure: Improvement is done by monitoring a perceptual condition If reliability increases, the behavior is added to the list of preconditions Keep monitoring in a circle until reaching the threshold

Genghis Six-legged robot that walks forward 12 behaviors, 6 conditions, 8742 nodes 4 eight-bit microprocessors, 32 KB memory The challenge is to learn how to coordinate the legs to produce a forward movement

Results Convergence time Non-intelligent search during the monitoring stage: 10 minutes Intelligent search: 1min 45sec A “tripod” gait emerged which is common among six-legged insects

Conclusions A learning algorithm was developed which allows a behavior-based robot to learn when its behaviors should become active using positive and negative feedback

Comments + Impressive results + Global behavior (walking) emerges from coordinated Behaviors + Simple idea, powerful consequences. Robot learned how to walk, wasn’t taught

Comments – Dead behaviors don’t revive. They might be useful in other situations ? How to deal with concurrent actions? (i.e. walking and following a target)