CREATING COMMON PRESENCE FOR A MULTIENTITY RESCUE TEAM

1
CREATING COMMON PRESENCE FOR A MULTIENTITY RESCUE
TEAM
Jussi Suomela, Jari Saarinen and Aarne
Halme Helsinki University of Technology

• Publication of PeLoTe-project

Project funded by the European Community under
the IST programme Future and Emerging
Technologies
2
IntroductionPeLoTe Building Presence through
Localization for Hybrid Telematic Systems

• Scenario
• Human and Robotic entities (HE, RE) explore
  common area
• Both provide continuous mapping data from
  environment
• Mapping information is processed to a common
  presence for both entities
• Applications Rescue, military, planetary, etc.

3
Motivation

• Multiple entities working in the same environment
  for a common goal
• Entities are autonomous, but different in nature
• How to fuse the information from entities in a
  way that would benefit all entities?
• ? Common Presence

4
Presence - Being where?

• Different entities have different locations and
  tasks
• Physical Presence Sense of being here/there
• Social Presence Sense of being together
• Common Presence Sense of being here/there
  together

5
Common Presence

• Common presence is a dynamic model - shared by
  diverse entities - that represents knowledge
  about a working environment, objects in it and
  their mutual relationships, as well as goals and
  the status of task.
• Common presence is an interface, not only an
  environmental model
• Transferring information between entities
• Providing a common space - presence, which all
  entities can understand

6
Common Presence in PeLoTe

• Centralized spatial model that includes concepts
  related to the fire fighting
• In PeLoTe all working entities are sharing the
  same frame of reference
• In essence, the common presence in PeLoTe creates
  and maintains the shared situational awareness

7
Operator

• Telepresent in the rescue site
• Plans and supervises the task
• Controls all the entities
• Updates the Common Presence model
• GUI for mission planning, entity control and
  telepresence. Verbal communication with human
  entites

PeLoTe GUI
8
Human entity

• Present and locally telepresent in the rescue
  site
• Explores the site and provides verbal information
  to the operator
• Wearable sensor system provides location and
  automatic mapping data to the common presence
  system

Human interface with map including the
navigated route (left) and real-time scanner data
with proximity alarm (right).
9
Fitting human in CP

• Common presence and the reality have to match ?
  POSITION!
• Personal Assistance System
• Communication
• Display
• PeNa for human positioning
• inertial navigation
• stride length measurement
• laser scanner for odometry and SLAM
• own US/RF based localization beacons

10
Robot entity

• Present and locally telepresent in the rescue
  site
• Robots presence is the virtual model of the CP
• Follows autonomously the planned track
• Provides continuous mapping data to the CP
• Can carry localization beacons

11
Building the presence

• Based on Standard Rescue Map (SRM), a 2D polygon
  map and an object database
• Future objective all buildings with straight
  alarm connection have one
• Shows the navigable area and important objects
• Objects in dynamic layers based on type and
  importance

• Continuous updating
• Sensor mapping
• Verbal comments, video clips
• Presence model is filtered to entity specific
  form
• Zoomable map for humans
• geometric polygon map for robots
• Updated CP model provided realtime to the entities

12
System telematics

• In real rescue task the communication is the most
  critical system
• In PeLoTe communication was built on TCP/IP based
  client server architecture
• The core is PeLoTe server located in the operator
  room
• WLAN supported with direct radio modem and walkie
  talkie communication

Pelote communication, TC means the operator and
PAS is the personal assistance system meaning the
support system of the exploring humans
13
PeLoTe experiments

• End user demonstrations
• Introducing the system to professionals

• User tests
• -6 teams with system, 6 without
• - Evaluation with questionnaires, observations
  and performance measures

14
Results - questionnaire

• In all cases the PeLoTe teams reported higher
  feeling of presence than control teams
• Exception was suspension of disbelief (the people
  trust more to own senses than information
  provided by system)
• SSM spatial situation model
• SoD Suspension of disbelief
• SP_SL Spatial presence self localisation
• VSI Visual spatial imagery

15
Results - Performance
Performance table for control teams

• Performance measures are in average much better
  for PeLoTe teams than for control teams.
• Only exception is time, which is due to the
  different actions they made during mission (such
  as teleoperating robot through dangerous areas)

Performance table for PeLoTe teams
16
Results - Observations

• People with no or little background on
  computers were able to use the system, control
  robots and successfully execute the mission
• Operator was able to track and map the
  situation, give control paths to robot and human.
• PeNa user was able to stay a way from the
  dangerous areas, find and rescue victims,
  teleoperate robot locally and follow instructions
  given by operator.
• The stress level of PeLoTe teams were much lower
  than with traditional teams.
• The PeLoTe teams had much better memory of what
  happened during the mission (situational
  awareness)

17
Conclusions

• Common presence is a tool to change information
  between cooperative human, robotic and possible
  other entities
• Presence provides communication, support and
  virtual world model for entities exploring in
  changing and partially unknown environment where
  conditions may limit the perception
• System was demonstrated in a rescue task but can
  be applied in military, construction etc.
• System was evaluated to increase the shared
  situational awareness
• The future work is to continue the theoretical
  formulation of the common presence