Multiple-UUV Approach for Enhancing Connectivity in Underwater - - PowerPoint PPT Presentation

Multiple-UUV Approach for Enhancing Connectivity in Underwater Ad-hoc Sensor Networks

Winston Seah Institute for Infocomm Research, Singapore

winston@i2r.a-star.edu.sg http://www1.i2r.a-star.edu.sg/~winston

2

Outline

• Introduction & Research Aim
• Reference Architecture
• Research Issues

– Identification of communication gaps – Searching of communication gaps – Bridging communication gaps

• Simulation Results and Analysis
• Conclusion

3

Underwater Ad-hoc Sensor Networks

• Large scale deployment for data collection,

monitoring and surveillance

• Adverse environment

– Ambient noise – Multipath and fading effects – Temporal variations in channel

• Acoustic communication

– Low bit rates – Long and variable propagation delays

4

Research Aim

• To improve network connectivity by bridging

temporal disconnections

– Using cooperative robotics approach – Relies on cooperation between mobile (robot) and static sensor nodes

• Hop count information
• Clustering information
• For post-deployment topology optimization

5

Reference Architecture

• Underwater collection points or sinks

– Deployed at the corner/boundary of the interested area

• Sensor nodes:

– Deployed in the area of interest – Communicate via multi-hop acoustic links to send data to collection points or sinks – Clustering and localization – Use simple ALOHA protocol for medium access

• UUV:

– Transmission equipment – Mission sensors (e.g., actuators) – Motion sensors (e.g., sonar) – A payload of sensor nodes that can be deployed

6

Reference Architecture (2)

sink S3

• ptical fibre

monitoring centre UUV S2 S1 S0 cluster C1 C0 C2 C3 C4 C5

7

Reference Architecture (3)

• UUV

– Directional antenna (distance/orientation) – Mission sensors – Motion sensors – Payload of static nodes to be dropped

front rear 1 2

3

4 5 6

8

Research Issues

• Identification of critical communication

gaps in underwater ad-hoc sensor networks

• Search algorithm to find the critical

communication gaps

• Methodology to bridge the communication

gaps

9

Identification of Communication Gaps

• Related work

– Clustering by ID – [Corke et al, 2004]

• Pros: simple and complete
• Cons: overhead, information only used for gap identification,

scalability issues

• Our approach – by hop count (HC)

– Each static node calculates its HC to sinks – Each static node broadcast this HC – A robot can detect communication gaps by receiving and analyzing the received static HC information

10

Propagation of Hop Counts

S0

1 1 2 2 2 3 3 3 3 3 4 4 4 5 5 5 5 5

*only hop counts to sink S0 are shown

11

Detection/Analysis of Neighbors’ HC Information

S3 X Y D E F Z C (10,10,10,3) (9,9,9,3) (16,13,6,16) (11,11,7,4) (12,12,6,5) (10,10,8,3) S0 S1 S2

12

Searching of Communication Gaps

• Exploration
• Related work

– With mapping – [Yamauchi, 1998]

• Pros: complete
• Con: complex and difficult, not scalable, expensive, not applicable

– Without mapping – [Bandyopadhyay et al, 2005]

• Pros: simple and scalable
• Cons: cannot guarantee 100% coverage
• Our approach – without mapping or with minimal mapping

– Predefined searching (need simple map and rough location info) – Perimeter searching (no mapping, need clustering info) – Swarm intelligence (no mapping) – Advanced Potential field based searching (no mapping)

13

Predefined Searching

• Needs map and location information
• Pros: complete
• Cons: not optimal

Reference Node UUV Static Node

14

Perimeter Searching

• Search around the boundary of clusters
• Pros: more promising to find inter-cluster gaps

suitable for non-uniform networks;

• Cons: requires a clustering algorithm

16

Advanced Potential Field Based Search Searching (Swarm Intelligence)

• Potential Field Based Search - Disperse UUVs in the

environment with obstacle avoidance capabilities

– Virtual repulsive force among UUVs – Virtual repulsive force between UUV and obstacles – Virtual attractive force to let UUVs move

• Multi-Robots collaboration

– broadcast headings neighboring robots choose different

• rientation
• Pros: simple and scalable
• Cons: low level cooperation

18

Bridging Communication Gaps

• If link is disconnected temporarily, the UUV will

move away when the connectivity is restored

• If the disconnection is permanent, the UUV can:

– Deploy sensor from its carried payload to bridge the gap; or – Remain as a bridge (or move away, depending on decision made at the monitoring centre)

19

Performance Evaluation

• Connectivity number of nodes that are

connected to at least one sink

• Average hop count average hop count of

each nodes to a sink

• k-connectivity average number of sinks

that nodes are connected to, where 0≤k≤n (n=4 in our simulation studies)

20

Simulation Environment

• Simulator

– Qualnet network simulator (packet loss, delay, …) – Player/Stage robotics simulator (sensor/actuator error, …) – Semaphore synchronize Qualnet and Player/Stage

• Size of network 2.5 km × 2.5 km
• Number of sinks 4 (at four corners)
• Number of static nodes initially 80
• Robots 4
• Static nodes to be dropped 20
• Transmission range 250 metres
• MAC protocol ALOHA (with no retransmissions)

21

Simulation Scenarios

• Number of UUV

1 (single-UUV) or 4 (multi-UUVs)

• 2 scenarios:

– Node failure due to corrosion, energy depletion, etc.. And may lead to network partitions – Intermittent link failure common phenomenon in underwater environments

22

Simulation Results – Node Failure

k-connectivity (multi-UUVs) 0.5 1 1.5 2 1 3 5 7 9 11 13 15 time step avg k-connectivity static APFS predefined perimeter hopcount (multi-UUVs) 20 22 24 26 28 30 1 3 5 7 9 11 13 15 time step avg hopcount static APFS predefined perimeter connectivity (multi-UUVs) 20 40 60 80 1 3 5 7 9 11 13 15 time step connected nodes static APFS predefined perimeter

23

Simulation Results –Link Failure

hopcount (multi-UUVs) 18 20 22 24 26 28 30 1 3 5 7 9 11 13 15 time step avg hopcount static APFS predefined perimeter connectivity (multi-UUVs) 10 20 30 40 50 60 70 80 1 3 5 7 9 11 13 15 time step connected nodes static APFS predefined perimeter k-connectivity (multi-UUVs) 0.5 1 1.5 2 1 3 5 7 9 11 13 15 time step avg k-connectivity static APFS predefined perimeter

24

Conclusion

• Harsh environment and adverse underwater

communication channels poses many challenges to deployment of sensor networks

• Poor channel conditions can further deteriorate

due to multipath fading and ambient noise

• Use of multiple UUVs with collaborative search

strategies can identify and bridge communication impairments in underwater sensor networks

25

References

Corke, et. al., “Deployment and Connectivity Repair of a Sensor Net with a Flying Robot”, in Proceedings of the 9th International Symposium

• n Experimental Robotics 2004 (ISER04), Singapore, Jun 2004

Yamauchi, B., “Frontier-Based Exploration Using Multiple Robots”, in Proceedings of the Second International Conference on Autonomous Agents (Agents '98), Minneapolis, MN, USA, 1998 Bandyopadhyay, T., Liu, Z., Ang, M. H. Jr., and Seah, W. K. G., “Visibility-based Exploration in Unknown Environment Containing Polygonal Obstacles”, in Proceedings of the 12th International Conference on Advanced Robotics (ICAR05), 2005

26

Thank you

http://www1.i2r.a-star.edu.sg/~winston http://www1.i2r.a-star.edu.sg/~winston