Online Interactive Building of Presence

1
Online Interactive Building of Presence
Suomela J., Saarinen J., Halme A., Harmo P.

• Description of PeLoTe-project

Project funded by the European Community under
the IST programme Future and Emerging
Technologies
2
Partners

• CTU Czech Technical University, Gerstner
  Laboratory (CZ) Coordinator
• Certicon CertiCon a.s. (CZ)
• HUT Helsinki University of Technology,
  Automation Technology Lab.(FI)
• JMUW Bayerische Julius-Maximilians Universität
  Würzburg (DE)
• ARS Steinbeis GmbH Co. für Technologietransfer
  (DE)

3
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.

4
Methods to be developed

• Personal Navigation (Human dead reckoning)
• human localization indoors
• Human SLAM
• human localization and mapping
• Cooperative localization
• Beacon based localization with the help of robots
• Robot SLAM
• Localization and mapping
• Forming the presence
• Common presence from different type of data

5
Human Entities (HE)

• Verbal mapping information
• Automatic localization with human dead reckoning
  and robot beacons
• Human SLAM based onboard laser scanner, if
  feasible

6
Personal Navigation (PeNa)

• In rescue situation the location of all men has
  to be known
• To map the environment the position has to be
  know
• ? human positioning needed!! (as accurate as
  possible)
• In indoor conditions ? commercial beacons (GPS,
  Cellular, etc.) cant be used
• ? Human dead-reckoning needed
• ? Own beacon system to correct the accumulated
  error
• Dead reckoning robot aided localization

7
PeNa Dead reckoning

• DRM (dead reckoning module, commercial)
• pedometer (inertial), compass, gps
• promises 5-10 error to distance
• very sensitive to pre-defined step length
• indoors problems with heading
• IMU (inertial measurement unit, commercial)
• three vibration gyroscopes
• three silicon-chip accelerometers
• SiLMU (Stride length measuring unit, lab-made)
• US based ankle distance meter
• 30Hz continuous measurement
• To be developed in the project

8
Dead reckoning, 1st tests

• Unaccurate heading data ? accumulating position
  error
• Distance accuracy of DRM is depending on the
  preset step length ? accumulating odometry error
• ? dead reckoning has to be supported

9
PeNa Robot aided localization

• Beacons on robots or dropped by the robots
• Observer unit measures distances to beacons

• Radio and US based low cost beacon system
• Simple system for small distances (lt10m)

10
Human SLAM

• To improve the dead-reckoning
• To provide accurate measurement data
• The first feasibility tests with Sick-scanner and
  IMU (heading only)
• Main problems are swinging and placement in human
  body, especially floor and ceiling echoes are a
  problem
• In future more compact scanners will be available

11
Human SLAM, feasibility tests

• Raw laser data and position estimate (integrated
  from IMU) as inputs
• Matching between scans with particle type
  algorithm
• Nearest neighbour method for each particle to
  find out the best match
• Scan to map matching by optimizing point
  distances to lines (geometric map)

12
Scanner based localization

• Comparison between
• scan to scan matching
• scan to map matching
• Heading estimation from gyroscope

13
Human SLAM, 1st results

• Map based localization succeeded
• Human SLAM seems feasible
• Used algorithm is non recoverable in case of
  failing
• Other algorithms (probabalistic approach) and
  tests in more difficult conditions are under work

14
Robot Entities

• Controlled by the operator
• Mapping continuously the environment
• Sending the mapping and visual information to
  operator
• Carry and/or drop localization beacons

15
Building Presence

• Model(SRM) database
• Static structures included in model
• Objects in database
• Geometric polygon model
• For robots geometric map is converted to OCM
• For Humans geometric model is visualized
• Final updating by operator

16
Conclusions

• Cooperative mapping of environment with robot and
  human entities
• Project in the start phase
• All methods will be developed, tested, integrated
  and demonstrated before end of 2004
• Methods (partially) tested until now
• Human dead reckoning
• Feasibility of human slam
• 3D modelling with augmented data
• Stride length measurements

17
Human SLAM demo
18
Standard Resque Map (SRM)

• SRM is the apriori map from the environment (2 or
  3D)
• ECDIS type multi- layer structure
• Model(map) database
• Structure in model
• Objects in database
• Alarms and alarm areas
• Sprinklers
• Dangerous materials
• Etc.