92846 CSCI696501 Special Topics on WIRELESS AD HOC NETWORKS

1
92846 CSCI-6965-01 Special Topics on WIRELESS
AD HOC NETWORKS

• S. Yang

2
Connected Dominating Set (CDS)

• Virtual network backbone
• Energy efficiency / interference reduction
• A node set is the DS if every node is
• in the set, or
• has at least one neighbor in the set
• If all nodes in the DS is connected, the set is a
  CDS

3
Research Efforts

• Many works seek a minimum connected dominating
  set (MCDS) in unit-disk graphs as their major
  design goal
• Performance bounds is their primary design
  parameter

4
MCDS for CDS (Das et al, UIUC, 1997)

• Select a node with global maximal degree as root
• Grow a tree by adding nodes with maximal
  effective degrees
• Centralized
• size at most 2(1 H(d)) OPT
• d is the maximum degree of the input graph and H
  is the harmonic function

5
Clustering Method for CDS (LinGerla, UCLA, 1996)

• Min-ID
• Select nodes with minimal IDs in 1 hop neighbors
  as cluster heads
• Select gateways to connect cluster heads
• localized

6
Pruning Based CDS Construction (Wu and Lis)

• Marking process
• Initially all vertices are unmarked
• any vertex having two unconnected neighbors is
  marked as a dominator
• Pruning
• Rule 1u can withdrew if there exists a neighbor
  v with higher ID be a neighbor to all neighbors
  of u
• Rule 21u can withdrew if there exist two
  connected neighbors v and w with higher IDs be
  neighbors to all neighbors of u
• Rule k

7
Research Scope
8
I. Extended CDS in CC Model
9
Cooperative Communication Model

• Cooperative communication (CC)
• physical layer design
• combine several partial signals to achieve the
  original signal
• Combine the advantages
• power savings
• spatial diversity
• increased data rates

10
Extended CDS (ECDS) in CC Model

• A node set is an EDS if every node is
• in the set,
• a regular neighbor of a node in the set, or
• a quasi neighbor of k nodes in the set
• Connectivity in CC model
• Strongly connectivity ECDS
• Weakly connectivity EWCDS
• EDS, ECDS, and EWCDS problems are NP-complete

11
(a) A sample network with CDS u, v, w and
ECDS u, v. (b) Another sample with EWCDS x,
u, v. (k2)
12
Solutions for ECDS Problems

• E-MCDS for EWCDS
• E-Clustering for EDS/ECDS
• E-AWF for EWCDS
• E-Rule K for ECDS

13
MCDS for CDS (Das et al, UIUC, 1997)

• Select a node with global maximal degree as root
• Grow a tree by adding nodes with maximal
  effective degrees

14
Extended MCDS for EWCDS

• Basic idea
• each node contributes 1 to all its neighbors and
  1/k to all quasi neighbors if selected
• Global solution
• Select a node with maximal contribution as the
  root
• Grow a tree by adding nodes with maximal
  effective contribution

15
20 nodes, 4 selected, k2
16
(No Transcript)
17
II K-Coverage CDS for Multiple Coverage
Requirement
18
Sensor Coverage

• Coverage models
• area coverage
• point coverage
• point coverage -gt area coverage (If the number of
  points are large enough)
• Select a subset of sensors to cover all the
  sensors
• Construct CDS/DS
• Multiple coverage (k-coverage)
• Reliability and security

19
Problem Definitions

• k-Coverage Set (k-CS) Problem
• a subset of nodes
• each node is covered by at least k different
  nodes in this subset
• K-connected coverage set (k-CCS) if they are
  connected

20
Local Solutions for K-CS/K-CCS

• Linear programming-based approach
• Cluster-based approach
• Rule-K based approach

21
Clustering Method for CDS (LinGerla, UCLA, 1996)

• Min-ID
• Select nodes with minimal IDs in 1 hop neighbors
  as cluster heads
• Select gateways to connect cluster heads

22
Cluster-based Solution

• Apply clusterhead finding k times
• clusterheads marked each time are removed from
  the network
• Find gateways
• connect the first set of the clusterheads
• For each marked node (clusterhead or gateway)
• if it does not have k marked neighbors, it
  designates some unmarked neighbors to be marked

23
15 nodes, 9 selected, k2
24
III Directional CDS using Directional Antennas
25
Directional Systems

• Directional antennas
• with the help of switched beam and steerable beam
  techniques
• each node divides its omnidirectional
  transmission range into K sectors
• Save energy and reduce interference
• Directional Network Backbone
• Contains a subset of nodes and their selected
  sectors to provide
• Domination
• connectivity

26
(No Transcript)
27
Directional CDS

• A special case of the directional network
  backbone where K is infinite
• In a directed graph, a DCDS is a set of selected
  nodes and their associated selected edges
• Each selected node can reach all other nodes in
  the network via only nodes or edges in DCDS
• The minimum DCDS problem is NP-complete

28
Coverage Condition for CDS (WuDai, IEEE TC04)

• A node is unmarked, if any pair of its neighbors
  are connected via high priority nodes

29
A Localized Solution

• Node Coverage Condition
• P(t) gt p(v) / P(u) gt p(v)
• Edge Coverage Condition
• P(v-gtu), p(v-gtw) gt p(v-gtw), p(v-gtw) (p(v),
  p(w)

(Dominating neighbor)
(Absorbant neighbor)
30
30 nodes, 13 nodes, 47 edges.
31
IV Super-node CDS in WSANs
32
Wireless Sensor and Actor Network

• Distributed hybrid sensor network consisting of
• resource-rich sensor devices (called actors)
• resource-impoverished sensor devices (called
  sensors)
• Actors provide real-time distributed management
• coordinate among themselves
• communicate with the sink for further instructions

33
WSAN
34
Super-node CDS

• Actors and selected sensors (G)
• G is persistent actor-connected
• if a sensor is connected to an actor through
  nodes in G, then it is still connected to the
  actor through nodes in G
• G is at-least-one actor-connected
• if each sensor is connected to at least one actor
  through nodes in G

35
Local Solutions

• persistent actor-connectivity
• any pair of its neighbors are connected via high
  priority nodes
• at-least-one actor-connectivity
• if any pair of its neighbors are connected via
  high priority nodes, or
• Each of the two neighbors is connected to an
  actor, and intermediate nodes in these two
  connections have higher priorities

(a)
(b)
36
Conclusions Future Work

• Conclusions
• Four new problems based on CDS
• Definition, solution, simulation
• Future work
• Related to the variations of CDS
• Directional transmission and reception
• Asynchronized environment
• Other topics
• Movement-assisted data communication
• Wireless security

37
Selected Publications

• Extended CDS
• J. Wu, M. Cardei, F. Dai, and S. Yang, Extended
  Dominating Set in Ad Hoc Networks Using
  Cooperative communication, Proc. of IFIP
  Networking 2005 (poster paper), June 2005.
• J. Wu, M. Cardei, F. Dai, and S. Yang, Extended
  Dominating Set and Its Applications in Ad Hoc
  Networks Using Cooperative Communication, IEEE
  Transactions on Parallel and Distributed Systems,
  Vol. 17, No. 8, August 2006.

38
Selected Publications (Cont.)

• K-Coverage CDS
• S. Yang, F. Dai, M. Cardei, and J. Wu, On
  Multiple Point Coverage in Wireless Sensor
  Networks, Proc. of the 2nd IEEE International
  Conference on Mobile Ad-hoc and Sensor Systems
  (MASS), November 2005. (Received the best paper
  award.)
• S. Yang, F. Dai, M. Cardei, J. Wu, and F.
  Patterson On Connected Multiple Point Coverage
  in Wireless Sensor Networks, accepted to appear
  in International Journal of Wireless Information
  Networks.

39
Selected Publications (Cont.)

• Directional CDS
• S. Yang, J.Wu, and F. Dai, Efficient Backbone
  Construction Methods in MANETs using Directional
  Antennas, accepted to appear in International
  Conference on Distributed Computing Systems
  (ICDCS), 2007.
• Super-node CDS
• J.Wu and S. Yang, On Maintaining Sensor-Actor
  Connectivity in Wireless Sensor and Actor
  Networks, INFOCOM 2008.

40
Questions/Comments

• http//cs.rpi.edu/yangs6