This preview shows page 1. Sign up to view the full content.
Unformatted text preview: Clustering in Ad hoc and Sensor Networks Why Clustering?
– – The data collected by each sensor is communicated through the The network to a single processing center that uses the data network Clustering groups nodes into groups such that each node communicate information only to clusterheads and then the clusterheads communicate the aggregated information to the processing center, saving energy and bandwidth saving The cost of transmitting a bit is higher than a computation; therefore, it The may be beneficial to organize the sensors into clusters may Clusterbased control structures provides more efficient use of Clusterbased resources for large dynamic networks resources – – Clustering can be used for – Transmission management Transmission – – Backbone formation Routing Efficiency Routing LinkClustered Architecture
[Baker+ 1981a, 1981b, Ephremides+ 1987]
– – – – – Reduces interference in multipleaccess broadcast environment Distinct clusters are formed to schedule transmissions in a contentionfree way Each cluster has a clusterhead, one or more gateways and zero or Each more ordinary nodes more Clusterhead schedules transmission and allocates resources within its Clusterhead cluster cluster Gateways connect adjacent clusters To establish linkclustered control structure 1. 2. 3. Discover neighbors Select clusterhead to form clusters Decide on gateways between clusters LinkClustered Architecture
[Baker+ 1981a, 1981b, Ephremides+ 1987]
Cluster Clusterhead Gateway Ordinary node Clusterheads
– Resemble base stations in cellular networks, but dynamic – Responsible for resource allocation – Maintains network topology – Acts as routers – forwards packets from one node to another – Aware of its cluster members – Aware of its onehop neighboring clusterheads Since clusterheads decide network topology, Since election election of clusterheads optimally is critical Previous Work
HighestDegree Heuristic [Gerla+ 1995, Parekh 1994] Computes the degree of a node based on the distance Computes (transmission range) between the node and the other nodes (transmission The node with the maximum number of neighbors (maximum The degree) is chosen to be a clusterhead and any tie is broken by the node ids by Drawbacks: A clusterhead cannot handle a large number of nodes due to clusterhead resource limitations resource Load handling capacity of the clusterhead puts an upper Load bound on the nodedegree bound The throughput of the system drops as the number of nodes The in cluster increases in Previous Work
LowestID Heuristic [Baker+ 1981ab, Ephremides+ 1987] The node with the minimum nodeid is chosen to be a clusterhead A node is called a gateway if it lies within the transmission range of node two or more clusters two Distributed gateway is a pair of nodes that reside within different clusters, but they are within the transmission range of each other clusters, Drawbacks: Since it is biased towards nodes with smaller nodeids, leading to Since battery drainage It does not attempt balance the load for across all the nodes Previous Work
NodeWeight Heuristic [Basagni 1999a, 1999b] Nodeweights are assigned to nodes based on the suitability of a node being a clusterhead of The node is chosen to be a clusterhead if its nodeweight is The higher than any of its neighbor’s nodeweights and any tie is broken by the minimum node ids broken Drawbacks: No concrete criteria of assigning the nodeweights Works well for “quasistatic” networks where the nodes do Works not move much or move very slowly not Optimizing Clustering Algorithm in Mobile Ad hoc Networks Using Genetic Algorithmic Approach
[Turgut+ 2002] [Turgut+ Weighted Clustering Algorithm (WCA) Weighted A clusterhead can ideally support δ nodes ideally
– Ensures efficient MAC functioning – Minimizes delay and maximizes throughput A clusterhead uses more battery power clusterhead
– Does extra work due to packet forwarding – Communicates with more number of nodes A clusterhead should be less mobile
– Helps to maintain same configuration Helps – Avoids frequent WCA invocation A better power usage with physically closer nodes
– More power for distant nodes due to signal attenuation Weighted Clustering Algorithm (WCA) Steps
1. Compute the degree dv each node v degree d v =  N (v )  = v∈ V , v ≠v ' ∑ { dist ( v, v ) < tx }
' range
' Coordinate distance, predefined transmission range. 1. Compute the degreedifference for every node 1. Compute degreedifference for ∆v =  d v − δ 
For efficient MAC (medium access control) functioning. Upper bound on # of nodes a cluster head can handle. Weighted Clustering Algorithm (WCA) Steps
3. Compute the sum of the distances Dv with all neighbors 3. sum Dv = v ∈ N (v )
' ∑ { dist ( v, v ) }
' 2 12 1 7 17 3 13 14 15 5 4 16 6 Energy consumption; more energy for greater dist. Energy communication. communication. Power required to support a link increases faster than Power linearly with distance. (For cellular networks) linearly (For Weighted Clustering Algorithm (WCA) Steps
4. Compute the average speed of every node; gives a measure of 4. mobility Mv mobility 1T Mv = T ∑ t =1 ( X t − X t −1) + (Y t − Y t −1)
2 2 Yt Yt1 time Xt1 Xt where where ( X t,Y t ) and v ( X t −1,Y t −1)
at time coordinates of the node coordinates t are the and ( t −1) Component with less mobility is a better choice for clusterhead. Component Weighted Clustering Algorithm (WCA) Steps
1. Compute the total (cumulative) time Pv a node acts as Compute time clusterhead clusterhead Battery drainage = Power consumed 6. Calculate the combined weight Wv for each node combined Wv = w1Δv + w2Dv + w3Mv + w4Pv for each node for 7. Find min Wv; choose node v as the cluster head, remove all Find neighbors of v for further WCA 1. Repeat steps 2 to 7 for the remaining nodes Load Balancing Factor (LBF) It is desirable to balance the loads among the clusters Load balancing factor (LBF) has defined as (should be high) LBF =
where, where, nc ∑i ( x i − µ ) 2 nc xi iis the number of clusterheads s iis the cardinality of cluster i s and µ N − n c is the average number of neighbors of a clusterhead = nc (N being the total number of nodes in the system) being Connectivity For clusters to communicate with each other, it is assumed that For clusterheads are capable of operating in dual power mode A clusterhead uses low power mode to communicate with its immediate clusterhead low neighbors within its transmission range and high power mode is used for high communication with neighboring clusters communication Connectivity is defined as (for multiple component graph) connectivity = size of largest component N Probability that a node is reachable from any other node ( 0 – 1; 1 being most desirable) 1; Demonstration Scattered nodes in the network Demonstration Clusterheads are identified Demonstration Clusters are formed Demonstration Clusters are connected Features of WCA Invocation of WCA is ondemand Invocation ondemand – Reduces information exchange by less system updates Reduces – Reduces computation/communication costs – Manages mobility by reaffiliations Manages reaffiliations – Delays (avoids) invocation of clustering as far as possible WCA is distributive WCA distributive – No clusterhead is over loaded – Balances load by limiting the cluster size Performance Metric
1. Number of clusterheads 2. Number of reaffiliations – a process where a node detaches from one clusterhead and process attaches attaches to another to
1. Number of dominant set updates – when a node can no longer attach to any of the existing when clusterheads clusterheads These parameters are studied for the varying These number of nodes number transmission range transmission maximum displacement maximum Simulation Environment System with N nodes on a 100x100 grid N was varied between 20 and 60 Nodes moved in all directions randomly Nodes Velocity of nodes were varied uniformly between 0 and 10 Transmission range of nodes was varied between 0 and 70 Ideal degree was fixed at δ = 10 Weighing factors: w1 = 0.7, w2 = 0.2, w3 = 0.05 and w4 = 0.05 0.7, Experimental Results Max displacement = 5 (const) Transmission range = 0  70 Number of nodes = 20  60 Ideal degree = 10 Experimental Results Max displacement = 1  10 Transmission range = 30 (const) Number of nodes = 20  60 Ideal degree = 10 Load Balancing Connectivity Performance of WCA Genetic Algorithms Map the possible solutions of the problem to symbolic space Possible solutions form a pool of solutions – population Possible population Solution strings – chromosomes and components of chromosomes – and genes genes Genetic Algorithm operations: – Selection Selection – Crossover – Mutation – Replacement – Elitism Encoding of the Chromosome N = # of nodes in the network of each with unique node id [1..N] used to encode each the chromosome by integer permutation integer all the ids should be included without any duplication, all and without order. and For instance: N = 100 , node ids [1..100] For Pool size = 50 (50 strings of integers/chromosomes) Pool 1 2 5 3 15     1 99 3 7 5     77 88 . . 50 6 8 3     55 44 Mapping WCA to GA
WV Values Node Ids of all nodes 5 3 7
. . 1 2 3
. . 7 10 25 5 63 22
… … 1 2
. . . 1 3 7    25 35 22 12 3 7 5    34 64 45 25 12
… 50 6 8 3    55 44 34 56 8
…
Data Encoded into chromosomes 4 8 99 100 Neighbors list WCA intermediate results Mapping WCA to GA
WV Values Node Id of a ClusterHead 5 7
. . . . 1 3
. . . . 7 10 25 15 22
… … 12 55
… 2 76 ClusterHead Set for a single chromosome GA Steps
1. Choose Initial Population 1. Choose
Randomly generate the initial population. Randomly Pool size = 50 (means 50 chromosomes) Pool While (new_pool_size < old_pool_size) repeat step 3 to 6 (repeat step 2 until the number of generation or the convergence is met) generation 2. Selection 2. Compute the fitness value for each chromosome by WV . Compute Roulette Wheel method is used based on the fitness values. Roulette 3. Crossover 3. X_Order1 method is used. X_Order1 Crossover rate = 0.8 GA Steps
4. Mutation 4. Mutation Swap method is used; randomly selecting two gene at positions i and j. positions Mutation rate = 0.1 Mutation 5. Replacement Append method is used. The new children will be appended Append into the new pool. into 6. Elitism  Check if the new children are better than the best, then replace the best by the child  Avoid being stuck on local optima Cfit Value Algorithm
FitnessValue = 0; 1. For each gene in chromosome repeat step 2 to 3 2. node = gene[I]; 3. if node is not clusterH and is not a member of the other clusterH and is Nodedegree <= MAX_DEGREE ( const ) Nodedegree Then it is a clusterH, Compute WV for this node Compute insert it into clusterHSet insert fitnessValue += WV; fitnessValue Cfit Value Algorithm
4. For each remaining node I from the network 4. If (it is not a clusterH and member of other clusterH, and and NodeDegree <= MAX_DEGREE) and then then Compute WV for this node Compute insert it into clusterHSet insert fitnessValue += WV; fitnessValue Performance Metric
1. Number of clusterheads 2. Number of reaffiliations – a process where a node detaches from one clusterhead and process attaches to another attaches These parameters are studied for the varying These number of nodes number transmission range transmission maximum displacement maximum 1. Load Distribution Load Simulation Environment System with N nodes on a 100x100 grid N was varied between 20 and 60 Nodes moved in all directions randomly Nodes Velocity of nodes were varied uniformly between 0 and 10 Transmission range of nodes was varied between 0 and 70 Ideal degree was fixed at δ = 10 Weighing factors: w1 = 0.7, w2 = 0.2, w3 = 0.05 and w4 = 0.05 0.7, Experimental Results WCA Optimized WCA Max displacement = 5 (const) Transmission range = 0  70 Number of nodes = 20  60 Ideal degree = 10 Experimental Results WCA Optimized WCA Max displacement = 1  10 Transmission range = 30 (const) Number of nodes = 20  60 Ideal degree = 10 Experimental Results WCA Optimized WCA Max displacement = 1  10 Transmission range = 30 (const) Number of nodes = 20  60 Ideal degree = 10 Load Balancing with WCA Load Balancing with GA The load balancing factor has improvement ten times with GA An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
– This paper proposes a distributed, randomized clustering algorithm to This organize the sensors in a wireless sensor network into clusters to minimize the energy used to communicate information from all nodes to the processing center processing By the generation of hierarchy of clusterheads, the energy savings increase By with the number of levels in the hierarchy with Sensor detects events and then communicate the collected information to a Sensor central location where parameters characterizing these events are estimated estimated In the clustered environment, the data gathered by the sensors is In communicated to the data processing center through a hierarchy of clusterheads The processing center determines the final estimates of the parameters The using information communicated by the clusterheads using – – – – An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
– – The processing center can be a specialized device or one of the sensors The itself itself In such clustered environment, sensor data is communicated over smaller In distances, the energy consumed in the network will be much lower than the energy consumption when every sensor communicates directly to the information processing center information The results in stochastic geometry are used to derive values of parameters The for the algorithm that minimize the energy spent in the sensor network for – An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
A New, EnergyEfficient, SingleLevel Clustering Algorithm New,
– Each sensor becomes a clusterhead (CH) with probability p and advertises Each itself as a clusterhead to the sensors within its radio range – these clusterheads are called volunteer clusterheads volunteer This advertisement is forwarded to all the sensors that are no more than k This hops away from the clusterhead hops Any sensor node that is not clusterhead itself receiving such advertisement Any joins the cluster of the closest clusterhead joins Any sensor node that is neither a clusterhead nor has joined any cluster Any itself becomes a clusterhead – called forced clusterheads forced Since the advertisement forwarding has been limited to k hops, if a sensor Since does not receive a CH advertisement within time duration t (where t is the time required for data to reach the CH from any sensor k hops away), it means that the sensor node is not within k hops of any volunteer CHs means – – – – An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
A New, EnergyEfficient, SingleLevel Clustering Algorithm New,
– – Therefore, the sensor node becomes a forced clusterhead The CH can transmit the aggregated information to the processing center The after every t units of time since all the sensors within a cluster are at most k hops away from the CH hops The limit on the number of hops allows the CH to reschedule their The transmissions transmissions This is a distributed algorithm and does not demand clock synchronization This between the sensors between The energy consumed for the information gathered by the sensors to reach The the processing center will depend on the parameters p and k Since the objective of this work is to organize sensors in clusters to Since minimize the energy consumption, values of the parameters (p and k) must minimize must be found to ensure the goal be – – – An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
A New, EnergyEfficient, SingleLevel Clustering Algorithm New,
Assumptions made for the optimal parameters are as follows: – – – – – – – The sensors are distributed as per a homogeneous spatial Poisson process The of intensity λ in 2dimensional space All sensors transmit at the same power level – have the same radio range r All Data exchanged between two communicating sensors not within each others’ Data radio range is forwarded by other sensors radio A distance of d between any sensor and its CH is equivalent to d / r hops distance Each sensor uses 1 unit of energy to transmit or receive 1 unit of data A routing infrastructure is in place; when a sensor communicates data to routing another sensor, only the sensors on the routing path forward the data another The communication environment is contention and errorfree; sensors do not The have to retransmit any data have An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
A New, EnergyEfficient, Hierarchical Clustering Algorithm New,
– – – – – This algorithm is extension of the previous one by allowing more than one This level of clustering in place level Assume that there are h levels in the clustering hierarchy with level 1 being Assume the lowest level and level h being the highest The sensors communicate the gathered data to level1 clusterheads (CHs) The level1 CHs aggregate this data and communicate the aggregated data The to level2 CHs and so on to Finally, levelh CHs communicate the aggregated data or estimates based on Finally, this aggregated data to the processing center this An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
A New, EnergyEfficient, Hierarchical Clustering Algorithm New,
– The cost of communicating the information from the sensors to the The processing center is the energy consumed by the sensors to communicate the information to level1 CHs, plus the energy consumed by the level1 CHs to communicate the aggregated data to level2 CHs, …., plus the energy consumed by the levelh CHs to communicate the aggregated data to the information processing center information Algorithm Details – – The algorithm works in a bottomup fashion First, it elects the level1 clusterheads, then level2 clusterheads, and so on An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
A New, EnergyEfficient, Hierarchical Clustering Algorithm New,
Algorithm Details – Level1 clusterheads are chosen as follows: o Each sensor decides to become a level1 CH with certain probability p1 Each and advertises itself as a clusterhead to the sensors within its radio range range o o – – This advertisement is forwarded to all the sensors within k1 hops of the This advertising CH advertising Each sensor receiving an advertisement joins the cluster of the closest Each level1 CH; the remaining sensors become forced level1 CHs level1 Level1 CHs then elect themselves as level2 CHs with a certain probability Level1 p2 and broadcast their decision of becoming a level2 CH and This decision is forwarded to all the sensors within k2 hops This An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
A New, EnergyEfficient, Hierarchical Clustering Algorithm New,
Algorithm Details – The level1 CHs that receive the advertisement from level2 CHs joins the The cluster of the closest level2 CH; the remaining level1 CHs become forced level2 CHs level2 Clusterheads at level 3, 4, 5,…,h are chosen in similar fashion with Clusterheads 3, probabilities p3, p4, p5,...,ph respectively to generate a hierarchy of CHs, in which any leveli CH is also CH of level (i1), (i2),…,1. which – An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
Advantages: Advantages:
– It is considered one of the earliest clustering algorithms in sensor It networks that incorporates energy efficiency into the design of the algorithm algorithm Since it is distributed algorithm, there is no need for clock Since synchronization between sensor nodes synchronization It achieves not only better energy efficiency, but also better time It complexity compared to previous work complexity The sensor nodes considered are simple nodes with fixed power level of The transmissions transmissions Since the algorithm is run periodically, the probability of becoming a Since clusterhead for each period is chosen to ensure that every node will get a chance to become clusterhead – providing the functionality for load balancing – – – – An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
Advantages: Advantages:
– – – Another approach to ensure load balancing is to trigger the algorithm Another when the energy levels fall below a certain threshold when Energy savings increases as the density of the sensor nodes increases Energy for single level clustering for For the hierarchical clustering algorithm, the energy savings increase for For (i) networks of sensors with lower communication radius, (ii) lower density of sensors in the network, and (iii) increase in the number of hierarchy levels hierarchy An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
Disadvantages: Disadvantages:
– The energy consumption of clusterheads has not been addressed since these The nodes will involve with more computation and communication of data to higher level clusterheads – consequence of nonuniform power consumption on the performance of the overall sensor network in the long run on An ideal network is assumed (contention and errorfree) which may not An reflect the real life scenarios reflect Possible load imbalance between different clusters Overhead associated with the clusterheads selection is not considered How does the network cope with sensor node failures? How is detected and How remedied? remedied? How does the network handle information sent by faulty sensors? – – – – – An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
Disadvantages: Disadvantages:
– How many forcedclusterheads can the sensor network handle? What is the How upper bound? What are the guarantees that forcedclusterhead will be able to communicate with the neighboring clusterheads? to Similarly, what is the upper bound on the number of sensor nodes within Similarly, one cluster? one Energy is wasted by those sensor nodes closer to the processing center Energy than their CH, but still need to go through their CH than – – An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
Suggestions/Improvements/Future Work: Suggestions/Improvements/Future
– What happens if a sensor node receives several join advertisements What from multiple nearby clusterheads? How does the sensor node decides which one to join? Possible solution: the decision can be made to join to the cluster with the Possible minimum number of members such that sensor nodes are evenly distributed among the clusters distributed – Error and contention in communication is not considered Error Possible solution: results may be verified with the real MAC protocol and traffic conditions under a simulator or a testbed traffic – The capabilities of the processing center should be more than the The regular sensor nodes regular An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks [Bandyopadhyay+, 2003]
Suggestions/Improvements/Future Work: Suggestions/Improvements/Future
– – Further energy efficiency can be achieved if the clusterheads can be in Further active or inactive mode (energy saving mode) active Depending on the distance from the clusterheads, the sensor nodes may Depending choose to transmit data towards clusterhead in various power levels (for instance, low vs. high) instance, In multihop mode, the sensor nodes closest to the clusterhead have the In most energy drainage due to data forwarding most Possible solution: a scheme allowing the sensor nodes to alternate between singlehop and multiplehop mode periodically between – EnergyEfficient Communication Protocol Architecture for Wireless Microsensor Networks (LEACH Protocol) [Heinzelman+ 2000, 2002]
– LEACH (LowEnergy Adaptive Clustering Hierarchy) is a clusteringbased protocol that utilizes the randomized rotation of local cluster base stations to evenly distribute the energy load within the network of sensors to It is a distributed, does not require any control information from base station It (BS) and the nodes do not need to have knowledge of global network for LEACH to function LEACH The energy saving of LEACH is achieved by combining compression with The data routing data Key features of LEACH include: Localized coordination and control of cluster setup and operation Randomized rotation of the cluster base stations or clusterheads and their Randomized clusters clusters Local compression of information to reduce global communication – – – LEACH [Heinzelman+ 2000, 2002]
– Considered microsensor network has the following characteristics: – – The base station is fixed and located far from the sensors All the sensor nodes are homogeneous and energy constrained Communication between sensor nodes and the base station is expensive and no Communication high energy nodes exist to achieve communication high By using clusters to transmit data to the BS, only few nodes need to transmit for By larger distances to the BS while other nodes in each cluster use small transmit distances distances LEACH achieves superior performance compared to classical clustering algorithms LEACH by using adaptive clustering and rotating clusterheads; assisting the total energy of the system to be distributed among all the nodes the By performing load computation in each cluster, amount of data to be transmitted to By BS is reduced. Therefore, large reduction in the energy dissipation is achieved since communication is more expensive than computation since – – LEACH [Heinzelman+ 2000, 2002]
Algorithm Overview Algorithm
– – – The nodes are grouped into local clusters with one node acting as the local base The station (BS) or clusterhead (CH) station The CHs are rotated in random fashion among the various sensors Local data fusion is achieved to compress the data being sent from clusters to the Local BS; resulting the reduction in the energy dissipation and increase in the network lifetime lifetime Sensor elect themselves to be local BSs at any any given time with a certain Sensor probability and these CHs broadcast their status to other sensor nodes probability Each node decided which CH to join based on the minimum communication energy Upon clusters formation, each CH creates a schedule for the nodes in its cluster Upon such that radio components of each nonclusterhead node need to be turned OFF always except during the transmit time always The CH aggregates all the data received from the nodes in its cluster before The transmitting the compressed data to BS transmitting – – – – LEACH [Heinzelman+ 2000, 2002]
Algorithm Overview Algorithm
– – – – – The transmission between CH and BS requires high energy transmission In order to evenly distribute energy usage among the sensor nodes, clusterheads In are selfelected at different time intervals are The nodes decides to become a CH depending on the amount of energy it has left The decisions to become CH are made independently of the other nodes The The system can determine the optimal number of CHs prior to election procedure The based on parameters such as network topology and relative costs of computation vs. communication (Optimal number of CHs considered is 5% of the nodes) vs. It has been observed that nodes die in a random fashion No communication exists between CHs Each node has same probability to become a CH – – – LEACH [Heinzelman+ 2000, 2002]
Algorithm Details Algorithm
– – The operation of LEACH is achieved by rounds The rounds Each round begins with a setup phase (clusters are selected) followed by steadystate phase (data transmission to BS occurs) Advertisement Phase: Advertisement – Initially, each node need to decide to become a CH for the current round based Initially, on the suggested percentage of CHs for the network (set prior to this phase) and the number times the node has acted as a CH and The node (n) decides by choosing a random number between 0 and 1 If this random number is less than T(n), the nodes become a CH for this round The threshold is set as follows: 1. – – – P T(n) =
1 – P * (rmod 1 ) P If n C G Otherwise 0 P = desired percentage of CHs r = current round G = set of nodes that have not been CHs in the last 1/P rounds LEACH [Heinzelman+ 2000, 2002]
Algorithm Details Algorithm
1. Advertisement Phase: 1. – – – – Assumptions are (i) each node starts with the same amount of energy and (ii) Assumptions each CHs consumes relatively same amount of energy for each node each Each node elected as CH broadcasts an advertisement message to the rest During this “clusterheadadvertisement” phase, the nonclusterhead nodes During hear the ads of all CHs and decide which CH to join hear A node joins to a CH in which it hears with its advertisement with the highest node signal strength signal 2. Cluster SetUp Phase: – Each node informs its clusterhead that it will be member of the cluster 3. Schedule Creation: – Upon receiving all the join messages from its members, CH creates a TDMA Upon schedule about their allowed transmission time based on the total number of members in the cluster members LEACH [Heinzelman+ 2000, 2002]
Algorithm Details Algorithm
4. Data Transmission: 4. – – – – Each node starts data transmission to their CH based on their TDMA schedule The radio of each cluster member nodes can be turned OFF until their The allocated transmission time comes; minimizing the energy dissipation allocated The CH nodes must keep its receiver ON to receive all the data Once all the data is received, the CH compresses the data to send it to BS Multiple Clusters – In order to minimize the radio interference between nearby clusters, each CH In chooses randomly from a list of spreading CDMA codes and it informs its cluster members to transmit using this code cluster The neighboring CHs radio signals will be filtered out to avoid corruption in the The transmission transmission – LEACH [Heinzelman+ 2000, 2002]
Advantages: Advantages:
– – – Localized coordination to enable scalability, and robustness for dynamic Localized networks networks Incorporates data fusion into the routing protocol in order to reduce the Incorporates amount of information transmitted to BS amount Distributes energy dissipation evenly throughout the sensors, thus increasing Distributes the system lifetime of the network the LEACH [Heinzelman+ 2000, 2002]
Disadvantages: Disadvantages
– – – How to decide the percentage of cluster heads for a network? The topology, How density and number of nodes of a network could be different from other networks density No suggestions about when the reelection needs to be invoked The clusterheads farther away from the base station will use higher power and The die more quickly than the nearby ones die LEACH [Heinzelman+ 2000, 2002]
Suggestions/Improvements/Future Work: Suggestions/Improvements/Future
– Extensions can be included to have hierarchical clustering where each CH Extensions will communicate with “superclusterhead” until the top layer of hierarchy in which the data needs to be sent to BS which The degree and remaining energy of a node may be considered as The parameters to decide a clusterhead in a round. If a clusterhead with a limited power used up its power in a round, the data to be transmitting may be lost power Since TDMA schedule is used, a large delay may be introduced between Since event detection and notification at base station. Therefore, the protocol is not suitable for a realtime application suitable – – TAS: Topology Adaptive Clustering for Wireless Sensor Networks [Virrankoski+, 2005]
– TASC is a distributed algorithm that partitions the network into a set of locally isotropic, nonoverlapping clusters without prior knowledge of the number of clusters, cluster size and node coordinates number Spatial grouping of nodes with respect to regions of close proximity and Spatial similar deployment density benefits similar – – Improving the ease of network management Efficient data aggregation and compression of sensor data Formation of hierarchies and node localization – The set of weights that encode distance, connectivity, and density The information within the locality of each node are derived information These weights form the terrain for holding a coordinated leader election in These which each node selects the node closer to the center of mass of its neighborhood to become its leader neighborhood TAS: Topology Adaptive Clustering for Wireless Sensor Networks [Virrankoski+, 2005]
– – The algorithm employs a dynamic density reachability criterion which allows The the grouping of nodes according to their neighborhood density properties the Assumptions made: Nodes are aware of their 2hop neighborhood Distances between nodes Clustering objectives: – A clustering algorithm should partition the network so that the nodes inside clustering each cluster have high correlation in sensor measurements and are evenly spaced in order to maximize gains and reduce errors due to ill geometric positioning as in the case of node localization positioning TASC requires only minimum number of nodes in a cluster The goal is to partition networks with density nonuniformities, into a set of The smaller locally isotropic clusters by grouping nodes with similar density attributes attributes – – TAS: Topology Adaptive Clustering for Wireless Sensor Networks [Virrankoski+, 2005]
Distributed Leader Election Algorithm Distributed – – Two main components: node weights and density reachability Two node density Two phases: nomination and voting followed by a merging phase Two nomination merging In first phase, each node considers weights of 2hop neighbors, In nominates the node with maximum weight as an election candidate and notifies the nodes in its neighborhood of this nomination and In second phase, each node elects the closest candidate as its leader. In Nodes that end up in clusters that are smaller than a prespecified minimum cluster size are dismantled and their node members join bigger existing clusters. It includes all shortest paths between all pairs of nodes that are located in path S. of TAS: Topology Adaptive Clustering for Wireless Sensor Networks [Virrankoski+, 2005]
Distributed Leader Election Algorithm Example Distributed
A
4 B
7 4 B 0.86 3 C
8 D
7 E
4 F 0.49
5 4 A 1.29 G1 D 10.15 C 0.84 E 11.46 H 0.51 – – – Define the weights to be the number of times a node is found on a shortest path Define when computing a weight for node when Node A can be found on the paths AB, AC, AD, and AE, its weight = 4 Node Node C receives a weight of 8 References
[Baker+ 1981a] D.J. Baker and A. Ephremides, A Distributed Algorithm for Organizing Mobile Radio Telecommunication Networks, Proceedings of the 2nd International Conference on Distributed Computer Systems, April 1981, pp. 476483. [Baker+ 1981b] D.J. Baker and A. Ephremides, The Architectural Organization of a Mobile Radio Network via a Distributed Algorithm, IEEE Transactions on Communications COM29(11), 1981, pp. 16941701. [Bandyopadhyay+ 2003] S. Bandyopadhyay and E.J. Coyle, An Energy Efficient Hierarchical Clustering Algorithm for Wireless Sensor Networks, IEEE INFOCOM 2003, San Francisco, CA, March 30 – April 3, 2003. [Basagni 1999a] S. Basagni, Distributed Clustering for Ad hoc Networks, Proceedings of International Symposium on Parallel Architectures, Algorithms and Networks, June 1999, pp. 310315. [Basagni 1999b] S. Basagni, Distributive and MobilityAdaptive Clustering for Multimedia Support in Multihop Wireless Networks, Proceedings of Vehicular Technology Conference, VTC, Vol. 2, 1999Fall, pp. 889893. [Ephremides+ 1987] A. Ephremides J.E. Wieselthier and D.J. Baker, A Design Concept for Reliable Mobile Radio Networks with Frequency Hopping Signaling, Proceedings of IEEE, Vol. 75(1), 1987, pp. 5673. References
[Gerla+ 1995] M. Gerla and J.T. Tsai, Multicluster, mobile, multimedia radio network, Wireless Networks, Vol. 1, No. 3, 1995, pp. 255265. [Heinzelman+ 2002] W. Heinzelman, A.P. Chandrakasan and H. Balakrishnan, An ApplicationSpecific Protocol Architecture for Wireless Microsensor Networks, IEEE Transactions on Wireless Communications, Vol. 1, No. 4, October 2002, pp. 660670. [Heinzelman+ 2000] W. Heinzelman, A.P. Chandrakasan and H. Balakrishnan, EnergyEfficient Communication Protocol for Wireless Microsensor Networks, IEEE Proceedings of the Hawaii International Conference on System Sciences, January 47, 2000, Maui, Hawaii. [Parekh 1994] A.K. Parekh, Selecting Routers in Adhoc Wireless Networks, Proceedings of the SBT/IEEE International Telecommunications Symposium, August 1994. [Turgut+ 2002] D. Turgut, S. K. Das, R. Elmasri, and B. Turgut, Optimizing Clustering Algorithm in Mobile Ad hoc Networks Using Genetic Algorithmic Approach, Proceedings of IEEE GLOBECOM 2002, Taipei, Taiwan, November 1721, 2002. [Virrankoski+ 2005] R. Virrankoski, D. Lymberopoulos, and A. Savvides, TASC: Topology Adaptive Spatial Clustering for Sensor Networks, IEEE INFOCOM 2005. ...
View
Full
Document
 Spring '08
 Boloni,L

Click to edit the document details