Load-Balanced Clustering in Wireless Sensor Networks Gaurav Gupta and Mohamed Younis IEEE International Conference on Communications, (ICC 2003) Bao-Hua Yang
Outline Introduction Introduction Network model Network model Load-Balanced clustering Load-Balanced clustering Simulation Simulation Conclusion Conclusion
Introduction In sensor network, sensor node has constrained energy In sensor network, sensor node has constrained energy When the topology is cluster, cluster head When the topology is cluster, cluster head Act as the manager Act as the manager May die quickly May die quickly Objective Objective Balance the load between cluster heads Balance the load between cluster heads Prolong the lifetime of cluster heads Prolong the lifetime of cluster heads
Introduction Basic idea Basic idea Balance the number of the sensor nodes in each cluster Balance the number of the sensor nodes in each cluster Consider on communication cost between gateway and sensors Consider on communication cost between gateway and sensors
Network model Assumption Assumption Two kind of nodes Two kind of nodes Sensor node Sensor node Sensing data Sensing data Energy-constrained Energy-constrained Gateway node Gateway node Cluster manager Cluster manager High-energy High-energy All gateways form their own subnet All gateways form their own subnet All nodes have All nodes have Location information---GPS Location information---GPS Not mobile Not mobile
Network model
Sensor energy model Sensor energy dissipation model transmitterreceiver
Load-Balanced Clustering Define Range set(Rset), Exclusive set(Eset), Define Range set(Rset), Exclusive set(Eset), the reach of sensor node the reach of sensor node G1G1 G2G2
Load-Balanced Clustering Define Range set(Rset), Exclusive set(Eset), Define Range set(Rset), Exclusive set(Eset), the reach of sensor node the reach of sensor node G1G1 G2G2 Rset G 1 {1,2,3,4,5,6,7} 1.Location energy reserve 2.energy reserve
Load-Balanced Clustering Define Range set(Rset), Exclusive set(Eset), Define Range set(Rset), Exclusive set(Eset), the reach of sensor node the reach of sensor node G1G1 G2G2 Rset G 1 {1,2,3,4,5,6,7} Rset G 2 {6,7,8,9,10,11}
Load-Balanced Clustering Define Range set(Rset), Exclusive set(Eset), Define Range set(Rset), Exclusive set(Eset), the reach of sensor node the reach of sensor node G1G1 G2G2 Rset G 1 {1,2,3,4,5,6,7} Rset G 2 {6,7,8,9,10,11}
Load-Balanced Clustering Define Range set(Rset), Exclusive set(Eset), Define Range set(Rset), Exclusive set(Eset), the reach of sensor node the reach of sensor node G1G1 G2G2 Rset G 1 {1,2,3,4,5,6,7} Eset G 1 {1,2,3,4,5} Rset G 2 {6,7,8,9,10,11} Eset G 2 {8,9,10,11}
Load-Balanced Clustering Define Range set(Rset), Exclusive set(Eset), Define Range set(Rset), Exclusive set(Eset), the reach of sensor node the reach of sensor node G1G1 G2G2 Rset G 1 {1,2,3,4,5,6,7} Eset G 1 {1,2,3,4,5} Reach: reach 1 =1,reach 6 =2 Rset G 2 {6,7,8,9,10,11} Eset G 2 {8,9,10,11}
Load-Balanced Clustering G2G2 G3G G1G Rset{1,2,3,5,4,6,10 } Rset{11,14,15,12, 13,10} Rset{7,8,9,4,6,12,13,10}
Load-Balanced Clustering G2G2 G3G G1G Rset{1,2,3,5,4,6,10 } Rset{11,14,15,12, 13,10} Rset{7,8,9,4,6,12,13,10}
Load-Balanced Clustering Define critical distance Define critical distance The median of distances in Eset The median of distances in Eset G1G Eset {2,3,1,5 } Rset{1,2,3,5,4,6,10 }
Load-Balanced Clustering Define critical distance Define critical distance The median of distances in Eset The median of distances in Eset G1G Eset {2,3,1,5 } Rset{1,2,3,5,4,6,10 }
Load-Balanced Clustering Define critical distance Define critical distance The median of distances in Eset The median of distances in Eset Eset {2,3,1,5,4} G1G Rset{1,2,3,5,4,6,10 } 10 Critical distance
Load-Balanced Clustering G2G2 G3G Rset{11,14,15 12,13,10} Rset{7,8,9,12,13,6,10} G1G Rset{1,2,3,5,6,4,10 }
Load-Balanced Clustering The cardinality of per cluster The cardinality of per cluster Objective function minimizes the variance of the cardinality Objective function minimizes the variance of the cardinality The cost between gateway and sensor node The cost between gateway and sensor node Where G: number of gateways X: cardinality of gateway G i X ’ : average cardinality including the node under consideration
Load-Balanced Clustering G2G2 G3G Rset{11,14,15 12,13,10} Rset{7,8,9,6,12,13,10} G1G Rset{1,2,3,5,6,4,10 }
Load-Balanced Clustering G2G2 G3G Rset{11,14,15 12,13,10} Rset{7,8,9,12,13,6,10} G1G Rset{1,2,3,5,6,4,10 } If add 6 to G 1
Load-Balanced Clustering G2G2 G3G Rset{11,14,15 12,13,10} Rset{7,8,9,12,13,6,10} G1G Rset{1,2,3,5,6,4,10 } If add 6 to G 1 If add 6 to G 2
Load-Balanced Clustering G2G2 G3G Rset{11,14,15 12,13,10} Rset{7,8,9,12,13,6,10} G1G Rset{1,2,3,5,6,4,10 }
Simulation Environmental setup Environmental setup Sensing area:1000*1000 square meter Sensing area:1000*1000 square meter Number of sensors: 100 to 500 Number of sensors: 100 to 500 Number of gateways: 2 to 10 Number of gateways: 2 to 10 Each node has initial energy: 5 joules Each node has initial energy: 5 joules
Simulation Gateways:5 Sensor nodes:100~500
Simulation
Conclusion This paper introduced an approach to balance the load among cluster head This paper introduced an approach to balance the load among cluster head Future work Future work Fault tolerance by providing backup gateway Fault tolerance by providing backup gateway WCNC2003 WCNC2003 Dynamic gateway Dynamic gateway