Robotic Technology for USAR Illah Nourbakhsh Robotics Lead NASA/Ames Research Center

1
Robotic Technology for USAR Illah
NourbakhshRobotics LeadNASA/Ames Research Center
2
Background Key Secondary Affiliations

• University
• Carnegie Mellon Associate Professor (on leave)
• University of California Santa Cruz Adjunct
  Professor
• University of Pittsburgh, Stanford, USF
  collaborations
• Government
• NSF Co-PI on Agent Architectures for Urban Search
  Rescue
• NIST PI on Instrumented Facilities for USAR
  standardization
• Chair, 2003 National Robocup Rescue exhibition
• Industry
• Intel Corporation embedded robotics processor
  development
• Intel Pittsburgh robotics university lead
• Robotics Engineering Task Force
• Evolution Robotics advisor

3
Role of Robotics in USAR

• Lower latency of first entry
• HAZMAT scheduling, preparation
• Structural analysis and approval
• Lower very high human risk
• Increase accessible domain
• Broaden operating conditions (heat, lack of
  oxygen)
• Human sensing augmentation
• Sensing Infrared imaging, Environmental modeling
• Force multiplier

4
Role of Robotics in USAR

• Lower latency of first entry
• Lower human risk
• Human sensing augmentation
• Increase survival chance and outcome for
  victims, decrease risk exposure and hazards to
  first responders.

5
Role of Robotics in Space Exploration

• Lower latency of first entry
• Lower human risk
• Human sensing augmentation
• Space Exploration and USAR share common goals
  and therefore common technological trajectories.
  - synergy

6
Barriers to Success

• Effective human-robot interaction
• USAR robot operating system standardization
• Mechatronic robot innovation
• Robot sensor interfaces
• Systems-level field testing and validation

7
Barriers to Success

• Effective human-robot interaction
• Current interfaces fragile, inefficient and need
  extensive training
• Human factors analysis must be applied to USAR
  case
• Iterative interaction design of interfaces
• This investment has high payoff, imagine 16
  humanrobot
• USAR robot operating system standardization
• Mechatronic robot innovation
• Robot sensor interfaces
• Systems-level field testing and validation

8
ROBOTS AT GROUND ZERO
Photos courtesy of University of South Florida
9
Barriers to Success

• Effective human-robot interaction
• USAR robot operating system standardization
• To maximize effectiveness across research
  efforts, we need standardized integrations of
  heterogeneous robot platforms.
• Decouple physical robot structure, embedded
  processing and high-level interaction design
• USAR O.S. will lower barrier to entry for
  industry and research partners
• Mechatronic robot innovation
• Robot sensor interfaces
• Systems-level field testing and validation

10
Barriers to Success

• Effective human-robot interaction
• USAR robot operating system standardization
• Mechatronic robot innovation
• No single robot design serves all search rescue
  needs.
• Consider USAR robotics as a set of tools
  dynamically assembled based on real-time demands
• Robustness and price point essential to
  commercial viability
• Robot sensor interfaces
• Systems-level field testing and validation

11
Barriers to Success

• Effective human-robot interaction
• USAR robot operating system standardization
• Mechatronic robot innovation
• Robot sensor interfaces
• Retooling existing USAR sensors for robot use.
• Human-readable sensors must be dual-use
• Overcome mechanical, electronic and AI obstacles
• Systems-level field testing and validation

12
Barriers to Success

• Effective human-robot interaction
• USAR robot operating system standardization
• Mechatronic robot innovation
• Robot sensor interfaces
• Systems-level field testing and validation
• We must build foundational knowledge rather than
  individual engineered solutions.
• Design gt Implementation gt Testing gt Evaluation gt
  Refinement gt Dissemination
• Instrumented test facilities, standards and
  evaluation methodologies are required

13
ARC and Partner Roles

• ARC has the on-base physical and intellectual
  resources and collaborators to be a prime center
  for rapidly maturing research and development for
  USAR robotics.

14
ARC Instrumented Test Facility

• Instrumentation for robotics per NIST direction
• Additional instruments for human factors
• From NIST 3-level to ARC spectrum of realism and
  continuity
• No test facility for robotics would be complete
  without real first responder training

15
Cooperative Test Facilities
16
USAR Test Arena ProliferationFOSTERING
COLLABORATION THROUGH STANDARDS
PREVIOUS COMPETITIONS AAAI Conference
2000 AUSTIN, TEXAS, USA IJCAI/AAAI Conference
2001 SEATTLE, WASHINTON, USA RoboCupRescue
2002 FUKUOKA, JAPAN AAAI Conference
2002 EDMONTON, ALBERTA, CANADA American Open
2003 PENNSYLVANIA, USA Japan Open 2003 NIIGATA,
JAPAN RoboCupRescue 2003 PADUA, ITALY
IJCAI/AAAI Conference 2003 ACAPULCO, MEXICO
2004 COMPETITIONS American Open German
Open Japan Open RoboCupRescue LISBON, PORTUGAL
AAAI Conference CALIFORNIA, USA
YEAR-ROUND ARENAS NIST MARYLAND, USA
(2000) Museum of Emerging Science TOKYO, JAPAN
(2002) Carnegie Mellon University PENNSYLVANIA,
USA (2003) Istituto Superiore Antincendi ROME,
ITALY (2003) University of New Orleans LOUSIANA
USA (2004) Bremen University BREMEMN GERMANY
(2004) Portugal TBD LISBON, PORTUGAL (2004)
17
USAR Interaction Design Testing

• ARC Human Factors division
• ARC Human-centered Computing Area
• JPL MER operation lessons
• U. Pitt. Psychology test instruments

18
Modeling and Visualization

• Migration of ARC tools used by scientists at MER
  mission ops in JPL
• Model visualization
• Image refinement

19
USAR Robotics Operating System

• ARC, Intel and CMU cooperation
• Modular robot architecture developed by JPL and
  ARC
• Embedded robot control board produced by Intel
• Robot control API for Intel XScale released by CMU

20
USAR Robotics Operating System

• ARC, Intel and CMU cooperation
• Modular robot architecture developed by JPL and
  ARC
• Embedded robot control board produced by Intel
• Robot control API for Intel XScale released by CMU

21
Sensor Integration

• USF expertise in first responder sensor needs
• CMU expertise in embedded sensor interfacing
  electronics and software
• ARC expertise in local reasoning and
  interpretation

22
Sensor Integration

• USF expertise in first responder sensor needs
• CMU expertise in embedded sensor interfacing
  electronics and software
• ARC expertise in local reasoning and
  interpretation

23
Sensor Integration

• USF expertise in first responder sensor needs
• CMU expertise in embedded sensor interfacing
  electronics and software
• ARC expertise in local reasoning and
  interpretation

24
Sensor Integration

• USF expertise in first responder sensor needs
• CMU expertise in embedded sensor interfacing
  electronics and software
• ARC expertise in local reasoning and
  interpretation

25
Mechatronic Robot Innovation

• ARC robot control development
• CMU rapid prototyping facilities

26
Mechatronic Robot Innovation

• ARC robot control development
• CMU rapid prototyping facilities

27
Mechatronic Robot Innovation

• ARC robot control development
• CMU rapid prototyping facilities

28
Mechatronic Robot Innovation

• ARC robot control development
• CMU rapid prototyping facilities

29
Mechatronic Robot Innovation

• ARC robot control development
• CMU rapid prototyping facilities

30
Simulation and Training

• U. Pitt. / CMU Unreal simulation, chosen for
  broad dissemination as the NIST standard
• Sensor model characterization only possible in
  most realistic possible environments ARC
• Scorpion EarBot dynamic modeling and neural net
  controlled V.O.R. research

31
Simulation and Training

• U. Pitt. / CMU Unreal simulation, chosen for
  broad dissemination as the NIST standard
• Sensor model characterization only possible in
  most realistic possible environments ARC
• Scorpion EarBot dynamic modeling and neural net
  controlled V.O.R. research

32
Simulation and Training

• U. Pitt. / CMU Unreal simulation, chosen for
  broad dissemination as the NIST standard
• Sensor model characterization only possible in
  most realistic possible environments ARC
• Scorpion EarBot dynamic modeling and neural net
  controlled V.O.R. research

33
Simulation and Training

• U. Pitt. / CMU Unreal simulation, chosen for
  broad dissemination as the NIST standard
• Sensor model characterization only possible in
  most realistic possible environments ARC
• Scorpion EarBot dynamic modeling and neural net
  controlled V.O.R. research

34
Systems-level Field Testing

• ARC Robotics acts as a whole-system hub
• Continuity extending beyond a student or academic
  year
• Test facility combined with evaluation
  methodology
• Industry, government and academic partnerships

35
Conclusion

• ARC has the on-base physical and intellectual
  resources and collaborators to be a prime center
  for rapidly maturing research and development for
  USAR robotics.
• First milestones for success
• Training center for USAR robotics
• Quantitative measurement of team effectiveness
• Robot prototypes that have viable commercial
  potential

36
Questions
Thank you