Robotic Capabilities David Kortenkamp (NASA Johnson ) Liam Pedersen (NASA Ames) Trey Smith (Carnegie Mellon University) Illah Nourbakhsh (Carnegie Mellon University) David Wettergreen (Carnegie Mellon University) Dan Clancy (NASA Ames)
Motivation Human/Robots Human/Robots Human/Robots Robots Robots Science Objectives Mission Concepts Robots Return all data Return selected data Select targets Characterize site Recognize unforeseen scientific opportunities Robot capability METRICS 10 year forecast! Breakthrough 12/18/2001 Robotics State-of-Art 2
Methodology How do we measure the state-of-the-art in space robotic capabilities? What is important? Functionalities, e.g., mobility, assembly How do you measure it? State-of-the-art metrics (qualitative) Performance metrics, e.g., distance traveled readiness metrics What is the state of the art? Flown robotic systems, e.g., Sojourner Fielded robotic systems, e.g., Nomad Laboratory demonstrations What is the future? Projections, breakthroughs and roadmaps 12/18/2001 Robotics State-of-Art 3
Mission Scenarios Planetary Surface Missions In- Missions Exploration Work Operations 12/18/2001 Robotics State-of-Art 4
In- Assembly, Inspection, and Maintenance Assembly Inspection Maintenance and repair Human EVA assistance Planetary Surface Exploration Surface reconnaissance In-depth site survey Sample acquisition and analysis Human exploration assistance 12/18/2001 Robotics State-of-Art 5
Robotic Functionalities Derived from mission scenario requirements Provide means for organizing and evaluating various robotic technologies Deliberately limited: robotics, not robotics Two mission scenarios Motivated by existing space robotics research 12/18/2001 Robotics State-of-Art 6
Mars Surface Exploration Functionalities Mobility Mobility Autonomy Terrain assessment, path planning, visual servoing Mobility Mechanism Extreme terrain access, energy efficiency Science Operations Perception, Planning, Execution On-board and ground tools; data analysis, target selection, operations planning and execution Sample Manipulation Position sensors, collect and process samples Human Interaction Human-Robot Interaction Tele-operation to human supervision; robot/eva astronaut teams 12/18/2001 Robotics State-of-Art 7
In- Assembly, Inspection, and Maintenance Functionalities Assembling structures Transporting and mating Move self and other massive elements; path planning, coverage patterns Making connections Manipulate small objects and tools; hand-eye coordination; fine motion planning Maintenance and Repair Component change-out Manipulation and sensing; grasping; turning bolts Accessing components Opening covers; removing blankets Inspecting structures Locomotion Path planning to cover an area; visual servoing on an anomaly Data analysis Recognizing and characterizing anomalies; taking appropriate action Human EVA Assistance Monitoring Tracking crew members; video archiving Teaming Physical interaction; sensing of human intention 12/18/2001 Robotics State-of-Art 8
Metrics State-of-the-art metrics (qualitative) Precise definitions Generalize to many systems Performance measures (quantitative) Resist temptation to use many easy to measure but uninformative numbers Cannot be reported for some fielded systems, but will hopefully set the bar for future reporting of results readiness metrics Mass, power, size, computation, etc. 12/18/2001 Robotics State-of-Art 9
What is the current state-of-art? Provide a list of functionalities and metrics to rate our progress in that particular functionality Ask experts to check off the metric that corresponds to the state-ofthe-art and the metric that will be the state-of-the-art in 10 years Experts currently being polled The following slides are preliminary assessments of the state-of-the-art that could change as we get more input 12/18/2001 Robotics State-of-Art 10
Surface Exploration Metrics Planetary Surface Exploration Surface reconnaissance In-depth site survey Sample acquisition and analysis Human exploration assistance 12/18/2001 Robotics State-of-Art 11
Surface Mobility Metric Traverse distance per command cycle 1 m 10 m 100 m 1000+ m Flight SOA Fielded SOA 10 year Forecast 12/18/2001 Robotics State-of-Art 12
Surface Mobility Metric Autonomous mobility in terrain types Level Consolidated Boulder Field Dunes Escarpment Flight SOA Fielded SOA 10 year Forecast 12/18/2001 Robotics State-of-Art 13
Sample Approach and Instrument Placement Metric Remote measurements Simple surface contact measurements Precision surface contact measurements Multiple targets in single cycle, highly robust Command cycles / operation : Multiple Multiple Single Highly autonomous Flight SOA Fielded SOA 10 year forecast 12/18/2001 Robotics State-of-Art 14
Whole Sample Manipulation Imprecise and unpredictable manipulation Precise and predictable manipulation Manipulate complex shapes Operate in complex environment w/ clutter, constraints and occlusions Command cycles / operation : Multiple Multiple Single Highly autonomous Example manipulators: Scoops, clamshell Gripper Dexterous gripper Human hand 10 year forecast Breakthrough Flight SOA 12/18/2001 Robotics State-of-Art 15
Onboard Science Perception and Science Plan Execution Execution: None (teleoperation) Time stamped sequence Flexible time, contingencies Prioritized task list with constraints High level science goals 10 years Return all data Return selected data Select targets 10 years Characterize site Recognize unforeseen scientific opportunities Breakthrough Perception: 12/18/2001 Robotics State-of-Art 16
Human Exploration Assistance Sensing of humans Generic obstacle avoidance Tracking of humans Tracking of human body parts (i.e., gestures) Recognition of humans and their activities Recognition of human physical and mental state Fielded SOA 10 year forecast Breakthrough Gesture recognition Simple, static gestures Fielded SOA Dynamic gestures Hand signals Gestures linked to natural language 10 year forecast 12/18/2001 Robotics State-of-Art 17
State-of-the-art example EVA Robotic Assistant at NASA Johnson 12/18/2001 Robotics State-of-Art 18
In- Robotic Operations Assembly Inspection Maintenance Human EVA Assistance 12/18/2001 Robotics State-of-Art 19
In- Robotic Assembly Payload capture Teleoperated capture of fixed component Autonomous capture of fixed component Teleoperated capture of free-flying component Autonomous capture of free-flying component Flight SOA Fielded SOA 10 year forecast Gross assembly One or more basic elements Flight SOA Fielded SOA Multiple components and orientations Large mass or flexible components 10 year forecast Complex assembly; gossamer components Breakthrough Mating connnectors Flight SOA Teleoperated mating of robot friendly connectors Autonomous mating of robot friendly connectors Fielded SOA Teleoperated mating of EVA connectors Autonomous mating of EVA connectors 10 year forecast Autonomous mating of arbitrary connectors Breakthrough 12/18/2001 Robotics State-of-Art 20
State-of-the-art Example Skyworker from Carnegie Mellon University 12/18/2001 Robotics State-of-Art 21
State-of-the-art Example Dira from Carnegie Mellon University and NASA JSC 12/18/2001 Robotics State-of-Art 22
In- Robotic Maintenance Locating a component Open loop control Closed loop control using special markers A priori model of undamaged component A priori model of damaged component Flight SOA Fielded SOA 10 year forecast Grasping a component Teleoperated grasping of robot friendly component Flight SOA Autonomous grasping of robot friendly component Fielded SOA Teleoperated grasp of component w/ handle Autonomous grasp of component w/handle 10 year forecast Autonomous grasp of arbitrary component Breakthrough 12/18/2001 Robotics State-of-Art 23
State-of-the-art Example Robonaut NASA Johnson Humanoid Robot 12/18/2001 Robotics State-of-Art 24
In- Inspection Inspecting structure Visual inspection of specific site -- teleop Visual inspection of a large area -- teleop Visual inspection of a large area -- autonomous Visual inspection of complex structure -- autonomous Flight SOA Fielded SOA 10 year forecast Analyzing data No data analysis Moaicing of images Filtering of data Detecting modeled anomalies automatically Detecting unmodeled anomalies automatically Flight SOA Fielded SOA 10 year forecast Breakthrough 12/18/2001 Robotics State-of-Art 25
Human EVA Assistance No commands -- teleoperation Text-based commands Speech-based commands Multi-model interaction Interactive dialogue Human-Robot Communication Flight SOA Fielded SOA 10 year forecast Physical Interaction Holding object for human Handing objects to human Taking objects from human Carrying/rescuing human Flight SOA Fielded SOA 10 year forecast 12/18/2001 Robotics State-of-Art 26
In- Assembly Overall Evaluation Teleoperated robots that move large components and mate parts Closely supervised, semi-autonomous robots that move large components and mate parts Teleoperated robots that can mate parts and make fine connections between parts Closely supervised, semi-autonomous robots that mate parts and make fine connections between parts Autonomous robots that move large components and mate parts with minimal human intervention Autonomous robots that mate parts and make fine connections between parts with minimal human intervention Autonomous robots that perform complete assembly of complicated structure (e.g., large telescope) from start to finish with substantial support from groundbased or in-space humans Autonomous robots that perform complete assembly of complicated structures (e.g., large telescope) from start to finish with minimal human intervention 12/18/2001 Robotics State-of-Art 27
Conclusions -fielded robotic systems lag far behind the current state-of-the-art In-space assembly lags behind surface exploration Not as much of an agency initiative Requirements for space robotics are growing Planetary exploration In-space assembly of next generation space telescopes at the LaGrange points little human capability 12/18/2001 Robotics State-of-Art 28