A Distributed Framework for Correlated Data Gathering in Sensor Networks

1

A Distributed Framework for Correlated Data Gathering in Sensor

Networks

Kevin Yuen, Ben Liang, Baochun LiIEEE Transactions on Vehicular Technology 2008

2

OutlineIntroductionProblem FormulationLocalized Slepian-Wolf CodingDistributed Solution: A Price-Based FrameworkImplementation IssuesPerformance Evaluation

3

IntroductionRecent technological advances have enabled the

production of low-cost sensors.Usually sensors are densely deployed in sensor

networks. (Overlapping sensing ranges)Find a transmission structure to minimize total

energyThis framework should be compatible

e.g. multi-sink, distributed solution, asynchronous network settings, sink mobility, duty schedules

4

Problem FormulationModel the WSN as a directed graph G=(V,E)

V = Assign every node i with rate Transmission range and exists if Each link(i,j) has a weight represents the flow rate of link(i,j)

We can minimize the optimization objective by adjusting and

5

Problem FormulationUse rate distortion theory to analyze the problem

Let S be a spatially correlated random Gaussian vector

Σ𝑖𝑗=𝑊𝑑𝑖𝑗

2

6

Problem FormulationGoal : Minimize transmission energy

ConstraintsFlow Conservation

Channel Contention

Rate Admissibility

7

Problem FormulationThe constraints and the correlated data-gathering

problem can be modeled as an exponential-constraint linear programming formulation

8

Localized Splepian-Wolf CodingDisadvantages of the optimization formulations

Difficult to solveRequire global knowledge of the correlation structure

Use Slepian-Wolf coding to relax the rate admissibility constraints such that only local correlation information is required.Each sensor node i should encode its data at a rate equal to

the conditioned entropyConsider the data correlation with one-hop neighbors in

9

Localized Splepian-Wolf CodingSupports multiple sinks:a subset of sensors within the neighborhood of

sensor that are closer to sensor ’s sink

10

Distributed Solution:A Price-Based

Framework

11

Lagrangian Dualization(1/2)Goal: allocate the limited capacity of the wireless

shared mediumPrice-based resource allocation

Each wireless link is a basic resource unitA price can reflect the relation between the traffic load of

the link and its bandwidth capacity

Relax the channel contention constraints with Lagrangian dualization

12

Lagrangian Dualization(2/2)

The weight of each link is equal to the sum of its energy and capacity cost.

energy capacity cost

13

Subgradient AlgorithmAn efficient iterative algorithm to solve the

Lagrangian dual problem.Solve the Lagrangian sub problem by finding the

shortest path from each sensor node to its nearest sink node with current Lagrangian multiplier during each iteration

Update the Lagrangian multiplier

𝛽𝑖𝑗 [𝑘+1 ]

14

Distributed Algorithm(1/2)

15

Distributed Algorithm(2/2)The algorithm requires 3 control packets

Flow rates of all links within the clusterPrices for all clusters that are inherent to itThe identities of other sensor nodes in its neighborhood

and their distance to destination sink node

90sensors,10sinks,Transmission rage=30m

16

Asynchronous Network ModelSynchronous network model

Every node simultaneously execute at every time instanceIt is expensive to synchronize local clocks across the

entire networkPartial-asynchronous network model

The time between consecutive updates is bounded by BAt time t, instead of the most recent information, a node

may receive a sequence of recent updatesCompute the average of the sequence of updates from

time to

17

Implementation IssuesPrimal Recovery

Guarantee to generate feasible primal solutionThe network must remain static

: step size : the weights of convex combination

18

Implementation IssuesCapacity Reservation

The rate allocation generated by subgradient algorithm often violate the channel contention constraints

Generate feasible solutions by reserving a suitable amount of capacity (e.g. 10%)

Handling Network DynamicsNodes retrieve up-to-date topology in their neighborhood

19

Performance Evaluation

20

Simulation EnvironmentsImplement with C++Experiments are performed on the random

topology with 90 sensor nodes and 10 sink nodesTransmission range & interference range are 30mThe capacity of wireless shared medium is 150

bitsCorrelation parameter Per node distortion

21

Converge SpeedChose 10% as sink nodesThe algorithm is executed in synchronous

environment with 500 iterationsPrimal Sub gradient

22

Impact of Asynchronous Network Settings

• Run 500 iterations with different time bounds B = 1,5,10,25

• The convergence speed is associated with the time bound B.

Primal

Sub gradient

23

Effect of Data CorrelationCompare the effect of data correlation between

synchronous and independent environment.D = 0.001, 0.01 and 0.1W = 0.9 to 0.9999

Implementation I : localImplementation II: global

24

Adaptation to Sink M

obility

Adaptation to Duty SchedulesModel duty schedules as a 2-state Markov chain and are state transition probabilitiesSet the simulation environment for 300s

25

𝜶𝜷=𝟓 𝜶+𝜷=𝟎 .𝟎𝟏