MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Programming Cooperative Teams To Perform Global Science HOME TWO Enroute COLLECTION.

Slides:



Advertisements
Similar presentations
Robot Sensor Networks. Introduction For the current sensor network the topography and stability of the environment is uncertain and of course time is.
Advertisements

A vision-based system for grasping novel objects in cluttered environments Ashutosh Saxena, Lawson Wong, Morgan Quigley, Andrew Y. Ng 2007 Learning to.
CIS 540 Principles of Embedded Computation Spring Instructor: Rajeev Alur
IMagic Senior Project Emre AYDIN Asil Kaan BOZCUOĞLU Onur ÖZBEK Egemen VARDAR Onur YÜRÜTEN Project Supervisor: Asst. Prof. Uluç Saranlı.
Kinodynamic Path Planning Aisha Walcott, Nathan Ickes, Stanislav Funiak October 31, 2001.
DESIGN OF A GENERIC PATH PATH PLANNING SYSTEM AILAB Path Planning Workgroup.
EE631 Cooperating Autonomous Mobile Robots Lecture 5: Collision Avoidance in Dynamic Environments Prof. Yi Guo ECE Dept.
A Robotic Wheelchair for Crowded Public Environments Choi Jung-Yi EE887 Special Topics in Robotics Paper Review E. Prassler, J. Scholz, and.
David Hsu, Robert Kindel, Jean- Claude Latombe, Stephen Rock Presented by: Haomiao Huang Vijay Pradeep Randomized Kinodynamic Motion Planning with Moving.
Autonomous Robot Navigation Panos Trahanias ΗΥ475 Fall 2007.
Implementation of RRT based Path planner and conversion into Temporal Plan Network By: Aisha Walcott Final Project Presentation Dec. 10, J.
Brent Dingle Marco A. Morales Texas A&M University, Spring 2002
Experiences with an Architecture for Intelligent Reactive Agents By R. Peter Bonasso, R. James Firby, Erann Gat, David Kortenkamp, David P Miller, Marc.
Modeling and Planning with Robust Hybrid Automata Cooperative Control of Distributed Autonomous Vehicles in Adversarial Environments 2001 MURI: UCLA, CalTech,
Planning for Humanoid Robots Presented by Irena Pashchenko CS326a, Winter 2004.
Presented By: Huy Nguyen Kevin Hufford
Integrating POMDP and RL for a Two Layer Simulated Robot Architecture Presented by Alp Sardağ.
Design of Autonomous Navigation Controllers for Unmanned Aerial Vehicles using Multi-objective Genetic Programming Gregory J. Barlow March 19, 2004.
PEG Breakout Mike, Sarah, Thomas, Rob S., Joe, Paul, Luca, Bruno, Alec.
CS 326A: Motion Planning Kynodynamic Planning + Dealing with Moving Obstacles + Dealing with Uncertainty + Dealing with Real-Time Issues.
Dr. Shankar Sastry, Chair Electrical Engineering & Computer Sciences University of California, Berkeley.
Multirobot Coordination in USAR Katia Sycara The Robotics Institute
Executing Reactive, Model-based Programs through Graph-based Temporal Planning Phil Kim and Brian C. Williams, Artificial Intelligence and Space Systems.
Aeronautics & Astronautics Autonomous Flight Systems Laboratory All slides and material copyright of University of Washington Autonomous Flight Systems.
INTEGRATED PROGRAMME IN AERONAUTICAL ENGINEERING Coordinated Control, Integrated Control and Condition Monitoring in Uninhabited Air-Vehicles Ian Postlethwaite,
What is it? A mobile robotics system controls a manned or partially manned vehicle-car, submarine, space vehicle | Website for Students.
1 DARPA TMR Program Collaborative Mobile Robots for High-Risk Urban Missions Second Quarterly IPR Meeting January 13, 1999 P. I.s: Leonidas J. Guibas and.
Decentralised Coordination of Mobile Sensors School of Electronics and Computer Science University of Southampton Ruben Stranders,
World space = physical space, contains robots and obstacles Configuration = set of independent parameters that characterizes the position of every point.
Cooperating AmigoBots Framework and Algorithms
Tracking with Unreliable Node Sequences Ziguo Zhong, Ting Zhu, Dan Wang and Tian He Computer Science and Engineering, University of Minnesota Infocom 2009.
ANTs PI Meeting, Nov. 29, 2000W. Zhang, Washington University1 Flexible Methods for Multi-agent distributed resource Allocation by Exploiting Phase Transitions.
DARPA TMR Program Collaborative Mobile Robots for High-Risk Urban Missions Third Quarterly IPR Meeting May 11, 1999 P. I.s: Leonidas J. Guibas and Jean-Claude.
Model-based Programming of Cooperating Robots Brian C. Williams, Jonathan Kennell, I-hsiang Shu, and Raj Krishnan Artificial Intelligence and Space Systems.
Probabilistic Reasoning for Robust Plan Execution Steve Schaffer, Brad Clement, Steve Chien Artificial Intelligence.
Architecture for Autonomous Assembly 1 Reid Simmons Robotics Institute Carnegie Mellon University.
Embedding Constraint Satisfaction using Parallel Soft-Core Processors on FPGAs Prasad Subramanian, Brandon Eames, Department of Electrical Engineering,
Forward-Scan Sonar Tomographic Reconstruction PHD Filter Multiple Target Tracking Bayesian Multiple Target Tracking in Forward Scan Sonar.
MURI: Integrated Fusion, Performance Prediction, and Sensor Management for Automatic Target Exploitation 1 Dynamic Sensor Resource Management for ATE MURI.
16.412J/6.835 Intelligent Embedded Systems Prof. Brian Williams Rm Rm NE Prof. Brian Williams Rm Rm NE43-838
Aero/Astro Open House MERS Research Group Model-based Embedded and Robotic Systems Group Space Systems Laboratory Massachusetts Institute of Technology.
1 Distributed and Optimal Motion Planning for Multiple Mobile Robots Yi Guo and Lynne Parker Center for Engineering Science Advanced Research Computer.
Selection and Navigation of Mobile sensor Nodes Using a Sensor Network Atul Verma, Hemjit Sawant and Jindong Tan Department of Electrical and Computer.
Cooperative Air and Ground Surveillance Wenzhe Li.
Pervasive Self-Regeneration through Concurrent Model-Based Execution Brian Williams (PI) Paul Robertson MIT Computer Science and Artificial Intelligence.
Distributed Algorithms for Multi-Robot Observation of Multiple Moving Targets Lynne E. Parker Autonomous Robots, 2002 Yousuf Ahmad Distributed Information.
Real-Time Support for Mobile Robotics K. Ramamritham (+ Li Huan, Prashant Shenoy, Rod Grupen)
Aeronautics & Astronautics Autonomous Flight Systems Laboratory All slides and material copyright of University of Washington Autonomous Flight Systems.
System RMPL Activity plan model Kirk Planner / Scheduler Plan Runner Schedulable / consistent planre-plan request commandsstatus update Compiler Converter.
Autonomy for General Assembly Reid Simmons Research Professor Robotics Institute Carnegie Mellon University.
December 20, 2015 Decentralized Mission Planning for Heterogeneous Human-Robot Teams Sameera Ponda Prof. Jonathan How Department of Aeronautics and Astronautics.
Behavior-based Multirobot Architectures. Why Behavior Based Control for Multi-Robot Teams? Multi-Robot control naturally grew out of single robot control.
Model-based Programming of Cooperative Explorers Prof. Brian C. Williams Dept. of Aeronautics and Astronautics Artificial Intelligence Labs And Space Systems.
Programming Sensor Networks Andrew Chien CSE291 Spring 2003 May 6, 2003.
Discovery and Systems Health Technical Area NASA Ames Research Center - Computational Sciences Division Automated Diagnosis Sriram Narasimhan University.
Randomized Kinodynamics Planning Steven M. LaVelle and James J
Space Systems Laboratory Massachusetts Institute of Technology AUTONOMY MIT Graduate Student Open House March 24, 2000.
Scheduling Messages with Deadlines in Multi-hop Real- time Sensor Networks Wei Song.
Planning Tracking Motions for an Intelligent Virtual Camera Tsai-Yen Li & Tzong-Hann Yu Presented by Chris Varma May 22, 2002.
Dynamic Mission Planning for Multiple Mobile Robots Barry Brumitt and Anthony Stentz 26 Oct, 1999 AMRS-99 Class Presentation Brian Chemel.
Heterogeneous Teams of Modular Robots for Mapping and Exploration by Grabowski et. al.
Beard & McLain, “Small Unmanned Aircraft,” Princeton University Press, 2012, Chapter 12: Slide 1 Chapter 12 Path Planning.
Scarab Autonomous Traverse Carnegie Mellon December 2007 David Wettergreen.
Ocean Observatories Initiative OOI CI Kick-Off Meeting Devils Thumb Ranch, Colorado September 9-11, 2009 Observation Planning and Autonomous Mission Execution.
University of Pennsylvania 1 GRASP Control of Multiple Autonomous Robot Systems Vijay Kumar Camillo Taylor Aveek Das Guilherme Pereira John Spletzer GRASP.
Automation as the Subject of Mechanical Engineer’s interest
Autonomous Operations in Space
Market-based Dynamic Task Allocation in Mobile Surveillance Systems
Self-Managed Systems: an Architectural Challenge
Overview: Chapter 2 Localization and Tracking
Presentation transcript:

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Programming Cooperative Teams To Perform Global Science HOME TWO Enroute COLLECTION POINT RENDEZVOUS Diverge SCIENCE AREA 1’ SCIENCE AREA 3 Landing Site: ABC Landing Site: XYZ ONE SCIENCE AREA 1 Mission controller specifies abstract set of goals for a robot team System must handle: Low-level planning and execution Dynamic environment Inter-vehicle communication

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Heterogeneous Robots Orbiter: –Earth Com link, –Large scale feature detection –Science observation Tethered Blimp: –Reconnaissance: Rover tracking, feature detection, local map generator –Sensor network deployment –Rover Com link Smart Mobile Lander –Slow mobile base station –Orbiter Com link –Large science package Scout Rovers –Fast agile rovers –Sensor package for identifying science objectives –Terrain mapping functionality Sensor Network –Highly constrained sensing/effecting communication array –Science sensing –Aids in robot localization High Tier Low Tier Mid Tier

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Kirk System Overview Graph-based Temporal Planner TPN Planner Control Program Plant Model distributed plan runner w/ strong/weak controllability RMPL Program DTPN Planner Distributed Runner Distributed Kirk Cooperative Path Planning Localization Incremental STN Consistency Control Program Plant Model Tiny RMPL Cooperative Kirk Kirk plans by selecting among alternate strategies (choose). Fast planning is achieved using a novel graph representation. Kirk executes by selecting consistent execution times, monitoring outcomes and re planning on failure.

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Programming Teams in RMPL (Group-Enroute() [l,u] ( (sequence choose ( (do-watching (PATH1=OK) ((Group-Traverse-Path(PATH1_1,PATH1_2,PATH1_3,RE_POS))[l*90%,u*90%]) ) (do-watching (PATH2=OK) ((Group-Traverse-Path(PATH2_1,PATH2_2,PATH2_3,RE_POS))[l*90%,u*90%]) )) (parallel ((Group-Transmit(OPS,ARRIVED))[0,2]) (do-watching(PROCEED=SIGNALLED) ((Group-Wait(HOLD1,HOLD2))[0,u*10%])) ))) Rendezvous Rescue Area Corridor 2 Corridor 1 Enroute RMPL Programs Describe concurrent sensing, actuation and movements activities. Choose specifies redundant strategies and contingencies. [A,B] Specifies timing constraints.

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Fast Planning Using Temporal Plan Networks (TPNs) [405,486] Ask(PATH1=OK) 12 7 Ask(PATH2=OK) 8 [405,486] [450,540] Ask(PROCEED) [0,54] [0,2] [0,  ] 1415 Tell(PATH1=OK) [450,450] 1617 Tell(PROCEED) [200,200] s e [500,800] [10,10] [0,  ] Group-Enroute Group Traverse Group Wait Group Transmit Science Target Planning searches out consistent trajectories. Satisfies all Asked conditions Checks Schedulability A TPN is created, representing all possible executions of an RMPL program.

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Kirk research challenges How do we generalize the plan runner to a more rich capability for plan execution monitoring, and how does this relate to mode estimation? How do we preserve dynamic scheduling with path planning? How do we achieve global optimality in the integration of TPN planning, generative activity planning and cooperative path planning? What is the relationship between hybrid estimation and localization? How do we fold the different elements of SLAM (Simultaneous localization and mapping) into the Kirk architecture?

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Distributed Planning & Execution mission Why distributed? to avoid mission being dependent on a single planning and execution robot distributed execution is essential when communication is limited resource constrained robots or sensor networks can’t run a big planning system

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Distributed Planning Solution Cooperative activity program is reformulated into data structures that can be distributed to robots Robust distributed coordination of activities: choose between functionally redundant procedures coordinate temporally flexible actions Research Challenges Minimize computation & communication Limited communication environment Plan failures Robot failures RMPL Distributed CSP Rover Rovers Blimp Rover Sensors

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Distributed Execution Goals –Provide methods to allow temporally flexible distributed plan execution in the presence of limited or unreliable communication due to physical isolation or communication requirements –Provide likelihood of successful execution of distributed plans Methods –Efficient methods to parse the plan while identifying regions of high interdependence –Use flexible contracts to coordinate separated plans and allow retrograde constraint propagations to update local temporal constraints –Implement efficient algorithms of for communication and contract negotiation –Develop calculate and continuously maintain stochastic measures of successful plan execution give probabilistic measures of activity durations and communication models

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Globally Optimal Planning for Agile Cooperative Autonomous Air Vehicles Apartment Building on Fire Example Scenario: Urban Search and Rescue Scenario Trapped Residents Autonomous Air Vehicles (AAVs): Do not have to account for the pilots’ safety. Can be pushed to their physical limits by performing aggressive maneuvers. Improve mission costs, such as fuel, by using multiple vehicles.

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Agile Vehicles Plan Optimally How can we plan for cooperative agile air vehicles while minimizing resource costs? –Extend the Kirk model-based activity planning system to plan collision-free trajectories to various locations. –Extend the model to include associated costs with activities. –Apply a best first search strategy to find the globally optimal plan (minimizing costs such as power, fuel, etc.)

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Agile Teams Search for Optimal Activity and Path Plans Concurrently Integrated Activity Planning and Path Planning: Search a plan representation, called a temporal plan network, in best-first order Dynamically compute collision-free paths for those plan activities that require moving between locations and the estimated cost of flying along this path Continuously interleave activity and path planning to pursue the most promising plan. Collision-free path Cost estimate = 10 units of fuel Cost estimate = 20 units of fuel Path Planning Strategy: Uniformly explore state space with randomized kinodynamic path planner based on Rapidly-exploring Random Trees (RRTs) Maneuver Automaton: Describes a set of agile maneuvers with respect to the vehicle’s dynamics RRT Location A: start state Location B: goal state

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Cooperative Planning Objectives: –Elevate role of human controller to that of a coach or wing commander, using autonomy to control individual autonomous vehicles –Use rich plan structure so that computer-generated plans can be explained back to human controllers with contingencies, alternative options, and choice justification –Support efficient distribution and communication for team missions –Include control-vector assignment with obstacle and collision avoidance

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Generative Deductive Controller Deductive Controller Generative Planner plant model environment and action data state configuration goals schedulable plan with low-level commands The deductive controller takes the high-level goal specification from the sequencer and outputs a hardware-executable plan network Challenge: Need to design a planning algorithm that supports rich actions (including time and resources), returns an optimal result, and is still fast enough for use in real-time, embedded systems Control Vector Assignment

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Control Vector Assignment Vehicle Waypoint Obstacle Control Vector Assignment Waypoint visitation Ordering Constraints Movement Commands For each Vehicle Generative Planner allocates vehicles to each waypoint and specifies ordering constraints on waypoint visitation This must be transformed into low-level commands for vehicles that take into account obstacle avoidance and vehicle- vehicle collision avoidance

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Dealing with Obstacles O1 O2 O3 O1 = L O2 = B O3 = A A simpler example: Mathematically solving the problem of vehicle control normally involves straightforward Linear Programming But the addition of obstacle avoidance introduces an Integer Programming element This makes the problem difficult to solve “online”: fast enough for actual vehicles in motion To resolve this we transform obstacles into Constraint Satisfaction Problem elements: For each obstacle, the domain is split into four regions (above/below/left/right), one of which is selected Integrating the selection of domains with the standard vehicle control leads to a an algorithm that can be used as a Hybrid CSP/LP Solver

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House Robonaut potential future MERS testbed Robonaut can perform or assist in extravehicular activities, reducing the need for humans to work in dangerous space environments.

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House 3D Simulation test-bed Simulates heterogeneous robots in a 3D world Compatible with hardware interface Extensions include: RoboSoccer RoboQuidditch

MIT Dept of Aeronautics and Astronautics March 21, 2003 Graduate Open House The rover test-bed Rover test-bed in MIT building 41 Rover Sensors –Stereo camera head –Sonar array –Laser range scanner –DGPS –Wheel encoders –Digital compass 1 ATRV 3 ATRV jr. A ceiling mounted stereo camera “the blimp” Motes Sensor Network