Relaying Node Placement in Wireless Sensor Networks

1
Relaying Node Placement in Wireless Sensor
Networks

• By
• Waleed Alsalih

2
Outline

• Introduction to Wireless Sensor Networks.
• Relaying node placement problem.
• Relaying node placement algorithms.

3
Wireless Sensor Networks (WSNs)

• WSN is a network of tiny devices called sensor
  nodes.
• A huge number of sensor nodes are densely
  deployed in a physical environment to sense a
  phenomena of interest.
• Sensor nodes sense the environment and send data
  to one or more Base Stations (BS) using wireless
  communication.
• Examples
• Environmental monitoring (e.g., temperature,
  humidity, habitat).
• Battlefield support (e.g., target tracking).

4
Properties of sensor nodes

• Sensor nodes are small in size.
• A sensor node consists of one or more sensors, a
  processor, memory, a radio, and a battery.
• Sensor nodes are battery-operated (hence, limited
  lifetime).
• Usually, sensor node is not recharged once its
  battery energy is drained.
• The main energy consumer is the radio.

Intel Mote (original size 3x3 cm). This picture
is taken from http//www.intel.com/research/
5
Multi-hop Relaying

• The energy consumed by wireless communication of
  a sensor node is a function of its transmission
  range and its traffic.
• Due to the limited amount of battery energy, the
  transmission range of a sensor node is limited,
  too.
• So, sensor nodes work as routers to help each
  other to reach the BS.

n1
n4
n2
n3
Base Station
Sensor node
6
Hierarchical Relaying

• Devices are of two types
• Sensor nodes.
• Relaying nodes.
• Each sensor node can reach at least one relaying
  node in one hope.
• Sensor node monitors the environment and
  generates data to be sent to its corresponding
  relaying node.
• Relaying nodes are responsible of relaying the
  received data to the base station.
• Advantages Energy efficient , distributed
  control, and localized effect of failures and
  changes.

Base Station
Sensor node
Relaying Node
7
Lifetime and connectivity

• Definitions
• The lifetime of a device is the time until it
  runs out of energy.
• The lifetime of the whole network is the time
  until one of the participating nodes run out of
  energy.
• Connectivity of a network is satisfied if every
  node has at least one path to the base station.

8
Communication energy consumption model

• With a fixed transmission range, the amount of
  energy consumed to send one date unit is
  constant. Let us call it Esend.
• Similarly, the amount of energy consumed to
  receive one data unit is constant. Let us call it
  Ereceive.
• Let T denote the desired lifetime,
• Let tr denote the traffic rate of a relaying
  node, and
• Let J denote the initial energy of a relaying
  node.
• Then we must have
• So, the capacity of a relaying node (the maximum
  traffic rate in order to stay alive for the
  desired lifetime T) is given by

9
Problem definition

• General problem formulation
• Input
• The number of sensor nodes.
• Location of each sensor node.
• Location of the base station.
• Data generation rate of each sensor node.
• A desired lifetime for the whole network.
• Output
• The minimum number of relaying nodes and their
  placement such that the connectivity and the
  required lifetime of the network are satisfied.

10
Minimum set cover

• Given a set A and a set S A2, where A2 is the
  set of all subsets of A, find set C S, which is
  the minimum cardinality set that satisfies the
  following condition
• Minimum set cover problem is NP-Hard.
• There are several approximation polynomial-time
  algorithms for this problem.

11
Case 1 Relaying nodes without energy constraints

• Sensor nodes have limited energy supply and,
  hence, limited transmission range.
• Relaying nodes have ample energy supply (e.g.,
  wall powered or high capacity battery) and,
  hence, can reach the base station directly (i.e.,
  in a one-hop fashion).
• This problem can be mapped to a minimum set cover
  problem.

12
Case 1 Relaying nodes without energy constraints

• Definitions
• Let X v1, v2, ,vn be the set of sensor
  nodes.
• A subset s of X is called viable if the
  intersection of the corresponding transmission
  disks of the sensor nodes in s is not empty.
• A viable subset s of X is called densest if it
  is not strictly contained in any other viable
  subset.

13
Case 1 Relaying nodes without energy constraints

• Example

n5
n6
n1
n3
n4
n2
Densest subset
Sensor node
14
Case 1 Relaying nodes without energy constraints

• Algorithm
• Find the set D of all densest subsets of X.
• Find a minimum cover set of D with respect to X
  (using one of the available approximation
  algorithms).
• If a densest subset d is in the minimum cover
  set, put one relaying node in the intersection of
  the corresponding transmission disks of sensor
  nodes in d.

15
Case 2 Relaying nodes with limited energy

• Sensor nodes and relaying nodes both have limited
  energy supply and, hence, limited transmission
  range.
• Relaying nodes communicate with the base station
  in a multi-hop fashion.
• Case 2 is more realistic than case 1.
• The solution space is infinitely large.
• Two-phase algorithms have been proposed for this
  problem.

16
Case 2 Relaying nodes with limited energy

• First phase
• The number and placement of the first phase relay
  nodes (FPRNs) aim at satisfying the connectivity
  of sensor nodes.
• Again, the problem of FPRNs placement can be
  viewed as a minimum cover set problem as in case
  1 version of the problem with minor
  modifications.

17
Case 2 Relaying nodes with limited energy

• Definition
• Let X v1, v2, ,vn be the set of sensor
  nodes.
• A viable subset s of X is energy-feasible if a
  relaying node deployed in the corresponding
  region can relay all traffic of all sensor nodes
  in s and still satisfy the lifetime constraint.
• An energy-feasible subset s of X is called
  densest-energy-feasible if it is not strictly
  contained in any other energy-feasible subset.

18
Case 2 Relaying nodes with limited energy

• FPRNs placement algorithm
• Find the set D of all energy-feasible-densest
  subsets of X.
• Find a minimum cover set of D with respect to X
  (using one of the available approximation
  algorithms).
• If an energy-feasible-densest subset d is in the
  minimum cover set, put one relaying node in the
  intersection of the corresponding transmission
  disks of sensor nodes in d.

19
Case 2 Relaying nodes with limited energy

• Second phase
• The number and placement of the second phase
  relay nodes (SPRNs) aim at satisfying the
  connectivity and lifetime constraints of relaying
  nodes.
• A Far-Near Max-Min algorithm has been proposed
  for this problem.
• Far-Near Start from the farthest relaying node
  from the base station and evolve step by step to
  ones closer to the base station.
• Max-Min Maximally utilize the energy capacity
  of existing relaying nodes to minimize the number
  of SPRNs.

20
Case 2 Relaying nodes with limited energy

• Definitions
• RNi is relaying node number i.
• The capacity of a relaying node is the amount of
  traffic it can handle without violating the
  lifetime constraint.
• The workload of a relaying node is the traffic
  assigned to it until now.
• The residual capacity of a relaying node is the
  difference between its capacity and its workload.
  
• The network of RNs is modeled as a digraph
  G(V,E), where
• VRN0, RN1, RN2 ,, RNN and RN0 is the
  base station.
• E(RNi, RNj) i gt 0, RNj is within the
  transmission range of RNi and RNj is closer to
  the base station.
• rci is the sum of the residual capacities of
  RNis adjacent relaying nodes, which are closer
  to the base station (i.e., having an incoming
  edge from RNi).

21
Case 2 Relaying nodes with limited energy

• SPRNs placement algorithm
• U V RN0
• Repeat the following steps until stop.
• Pick a relaying node RNi in U which is farthest
  from RNi .
• If (RNi, RN0) E, stop.
• Calculate rci.
• If rci gt the workload of RNi, Then
• RNi distributes its workload to the neighboring
  relaying nodes by first filling up the relaying
  nodes closer to the base station.
• Delete RNi from U
• Else
• A new RN is deployed along the line between RNi
  and the base station, and as close to the base
  station as posible.
• Add the new RN to U and delete RNi from U.

22
References

• A list of references is given in
• Q. Wang, K. Xu, G. Takahara, and H.
  Hassanein, Locally Optimal Relay Node Placement
  in Heterogeneous Wireless Sensor Networks, Proc.
  of the IEEE GLOBECOM, 2005.
• --------------------------------------------------
  ------------------------
• Questions ???