ºÝºÝߣ

ºÝºÝߣShare a Scribd company logo

Clustering in wireless sensor networks with compressive sensing

•Download as PPTX, PDF•
•
S
shivani Shivanichou1

Clustering is commonly used for data routing in wireless sensor networks. This document proposes using compressive sensing (CS) with clustering to further reduce data transmissions. Simulation results show that clustering with hybrid CS can significantly reduce transmissions compared to clustering without CS, shortest path tree routing without CS, and shortest path tree routing with CS. Specifically, when the compressive ratio is 10, clustering with hybrid CS reduced transmissions by 60% compared to clustering without CS and 50% compared to the other methods. When the ratio is 5, reductions were 50%, 40%, and 20% respectively. Thus the proposed method of using CS with clustering minimizes data and helps maximize the lifetime of resource-constrained sensor networks.

1 of 14
Downloaded 30 times
CLUSTERING IN WIRELESS SENSOR NETWORKS USING COMPRESSIVE SENSING
Wireless sensor network • Sensor nodes distributed for observing physical and environmental conditions
Clustering for data routing •Clustering is a data aggregation method Clustering in wireless sensor networks
Compressive sensing • Mathematical technique • Works on the logic of : ‘information’ bandwidth<‘total’ bandwidth • States that it is possible to acquire same amount of data from fewer measurements than we take conventionally. • Applicable as WSN data is largely of sparse nature.
Data sensed by the sensor nodes M<<N samples are transmitted Y=φx projections are transmitted Obtain x by using l-1 minimalization Accurate recovery of signals Compressive ratio = M/N
Analysis and comparison We compare routing using • Clustering with no CS • clustering with Hybrid CS • SPT with no CS • SPT with hybrid CS Parameters of comparison • No. of transmissions vs. no. of nodes • Reduction ratio of transmissions
Clustering with Hybrid CS
Simulation of the project TOPOLOGY FORMATION CLUSTER FORMATION SINK 20m 10 m
Clustering in simulation CLUSTER HEAD’S SINK AT (0,0)
Results WHEN THE COMPRESSIVE RATIO IS 10
if no. of measurements =1/10 (no. of nodes ) Reduction Ratio Method 1 Method 2 60 Clustering with hybrid CS Clustering without CS 50 Clustering with hybrid CS SPT without CS 30 Clustering with hybrid CS SPT with Hybrid CS
WHEN COMPRESSIVE RATIO IS 5
if no. of measurements =1/5(no. of nodes ) Reduction Ratio Method 1 Method 2 50 Clustering with hybrid CS Clustering without CS 40 Clustering with hybrid CS SPT without CS 20 Clustering with hybrid CS SPT with Hybrid CS
Results • Our method of using CS with clustering in WSN can significantly reduce data transmissions compared with conventional data collection methods of Clustering without CS,SPT without CS,SPT with CS. • Using clustering with hybrid CS minimize data transmissions and help maximize lifetime of network with the resource constrained sensor nodes.

More Related Content

Clustering in wireless sensor networks with compressive sensing

  • 1. CLUSTERING IN WIRELESS SENSOR NETWORKS USING COMPRESSIVE SENSING
  • 2. Wireless sensor network • Sensor nodes distributed for observing physical and environmental conditions
  • 3. Clustering for data routing •Clustering is a data aggregation method Clustering in wireless sensor networks
  • 4. Compressive sensing • Mathematical technique • Works on the logic of : ‘information’ bandwidth<‘total’ bandwidth • States that it is possible to acquire same amount of data from fewer measurements than we take conventionally. • Applicable as WSN data is largely of sparse nature.
  • 5. Data sensed by the sensor nodes M<<N samples are transmitted Y=φx projections are transmitted Obtain x by using l-1 minimalization Accurate recovery of signals Compressive ratio = M/N
  • 6. Analysis and comparison We compare routing using • Clustering with no CS • clustering with Hybrid CS • SPT with no CS • SPT with hybrid CS Parameters of comparison • No. of transmissions vs. no. of nodes • Reduction ratio of transmissions
  • 8. Simulation of the project TOPOLOGY FORMATION CLUSTER FORMATION SINK 20m 10 m
  • 9. Clustering in simulation CLUSTER HEAD’S SINK AT (0,0)
  • 11. if no. of measurements =1/10 (no. of nodes ) Reduction Ratio Method 1 Method 2 60 Clustering with hybrid CS Clustering without CS 50 Clustering with hybrid CS SPT without CS 30 Clustering with hybrid CS SPT with Hybrid CS
  • 13. if no. of measurements =1/5(no. of nodes ) Reduction Ratio Method 1 Method 2 50 Clustering with hybrid CS Clustering without CS 40 Clustering with hybrid CS SPT without CS 20 Clustering with hybrid CS SPT with Hybrid CS
  • 14. Results • Our method of using CS with clustering in WSN can significantly reduce data transmissions compared with conventional data collection methods of Clustering without CS,SPT without CS,SPT with CS. • Using clustering with hybrid CS minimize data transmissions and help maximize lifetime of network with the resource constrained sensor nodes.