Cooperative Navigation

1
Cooperative Navigation

• Towards Efficient Navigation of AUVs in a
  Multi-Vehicle Environment

Alexander Bahr
2
Overview

• Introduction, background and research interest
• Challenges of multi-AUV operation
• Cooperative navigation scenarios
• Three experiments (setup and results)
• Outlook
• Application of results in future experiments
• Conclusion

3
Why AUVs?

• Compared to ship/ROV survey
• Cheaper
• Safety
• Mission duration
• Access to previously unreachable places (under
  ice)
• and when using many
• Larger area coverage
• Robustness
• New types of missions

2004 2002 1995 1990
4
My AUV projects
SERAFINA A small and inexpensive test platform
for AUV swarm experiments GOATS 2004 Mine
detection and classification using 2 AUVs to
collect multi-static sonar data ASC
kayaks Equipped with Woods Hole Oceanographic
Institution (WHOI) modems for experiments with
multi-AUV configurations
5
Multi-AUV Challenges

• More complex mission
• Increased use of acoustic spectrum through
  navigation and communication equipment
• Collision avoidance

• Higher degree of autonomy, meta tasks
• Efficient communication schemes
• Efficient use of all navigation information

• The introduction of multiple AUVs affects all
  areas
• Mission planning, task assignment, execution
• Navigation/Communication

6
Multi-vehicle navigation scenarios

• Centralized navigation
• Special AUVs serve as Moving Long Baseline (MLBL)
  navigation beacons
• Absolute position is relevant
• Navy scenario
• Where am I?

• De-centralized navigation
• Homogenous group of vehicles
• Geometry of group relevant
• Where are the others?

7
Navigation state of the art

• Doppler Velocity Log (DVL) and Inertial
  Navigation System (INS) provide dead-reckoned
  position
• AUV queries Long Baseline (LBL) navigation
  beacons and receives response
• AUV records time of flight, knows beacon
  positions and computes absolute position
• Extended Kalman (EKF) filter is used to update
  dead-reckoned position estimates with absolute
  position readings

8
AOFNC 2003 Experiment Setup

• Two boats serve as Moving Long Baseline (MLBL)
  beacons
• AUV gets range information every 3 seconds
• Range ping reception is synchronously logged on
  both boats
• Position of AUV is determined by post-processing
  with the combined logs of the two boats and the
  AUV

9
AOFNC 2003 Experiment Results
10
AOFNC 2003 Experiment Conclusions

• This experiment relied on
• Reliable, high bandwidth communication
• High update rate for range measurements
• Pros
• Simple
• No Ambiguities
• Cons
• High bandwidth (communication and range
  measurements)
• Reliable communication is needed
• Does not scale well

11
AOFNC - Kayak Experiment Setup

• All kayaks have perfect, absolute position
  information through GPS
• Range measurements between kayaks using the WHOI
  modems

12
AOFNC - Kayak Experiment Setup

• Kayak 2 and kayak 3 serve as MLBL beacons
  transmitting their traveled vector via acoustic
  modem
• Kayak 1 uses several range and vector
  measurements from kayak 2 and kayak 3 to estimate
  its absolute position
• Assumptions High INS/DVL-update O(1/s)
  Few modem transmissions
  O(2/min)

13
AOFNC - Kayak Experiment Setup
14
AOFNC - Kayak Experiment range errors
15
AOFNC - Kayak Experiment Results
16
AOFNC - Kayak Experiment Results
17
AOFNC - Kayak Experiment Conclusions

• Tracking, using range measurements over time
  combined with traveled vector information is
  possible
• Pros
• More effective use of communication bandwidth
• Scalable
• Cons
• Position ambiguities
• Possible improvements
• Different approach to selection of cluster center
• Different clustering techniques

18
Intra-Vehicle-Range Experiment Setup

• Kayak 1 tracks kayak 2
• Kayak 2 broadcasts themotion vector between
  tworange pings
• Kayak 1 combines rangemeasurements and
  motionvectors to determine its relative
  position to kayak 2

19
Inter-Vehicle-Range Experiment Results
20
Conclusions and Outlook

• Range-only measurements over time combined with
  information about traveled vector can be used for
  tracking
• Combination of data transmission (communication)
  and range measurements (navigation) effective way
  to increase use of the acoustic spectrum
• Future Research
• Improve processing of raw solutions
• Acquire more data sets with different setups
  kayaks/AUVs
• Move from post-processing to real time
• Apply results in upcoming experiments

21
PLUS project
22
ASAP Experiment (Monterey Bay 2006)
23
(No Transcript)
24
Slotted Data Transmission Time Sync
25
Slotted Data Transmission Data Transfer
26
LBL range error gaussian ?
27
LBL range error gaussian ?
28
AOFNC 2003 Experiment Results (EKF)
29
AOFNC 2003 Experiment Results
30
AOFNC - Kayak Experiment range errors
31
How to compute the solutions
32
Challenges for Cooperative AUV Navigation

• Communication challenges
• Acoustic Communication Only
• Low Range (200m 2km)
• Low Bandwidth (GOATS 2004 10 Bytes/sec)
• High Power Consumption (100W TxPower)
• Unreliable Channel
• Interference with Navigation Equipment
• Not Very Stealthy
• Navigation challenges
• No GPS (Underwater GPS SystemRelies on
  Infrastructure that Needs tobe Deployed)
• Error of Onboard Dead-Reckoning Sensors
  Growswithout Bound
• Limited Navigation Bandwidth (AcousticNavigatio
  n Sensors may Interfere with Each Other)
• Interference with Communication Equipment
• Not Very stealthy

33
Pianosa 2004 (PLUS)
34
AOFNC - Kayak Experiment Setup
35
(No Transcript)