«Carles Gómez Josep Paradells José E. Caballero Edita: Fundación Vodafone España Autores: Carles Gómez Montenegro* Universitat Politècnica de ...»
i) topology control techniques based on the use of hierarchical networks, and ii) topology control techniques based on the use of dynamic transmission range adjustment. It must be noted, however, that while topology control may be integrated with other mechanisms (e.g. MAC or routing protocols), it is actually a cross-layer function and is not related with any layer in particular.
In the first case, there exists a hierarchy within the network whereby certain mechanisms do not employ all the nodes in the network. Instead, such mechanisms involve only a subset of nodes according to the network hierarchy. In addition, some special nodes may alleviate certain tasks of other nodes. In the second case, whether or not a network hierarchy exists, the transmission range is appropriately (and dynamically) tuned according to a particular goal. Since transmission and reception are two dominant components of energy consumption in a sensor node, transmission range adjustment is a relevant technique for meeting performance requirements in terms of parameters such as energy consumption and connectivity. Note that combinations of the two categories of topology control techniques may exist.
Topology control requires the use of topology control protocols, which aim at building and maintaining network topologies. Ideally, these protocols
should be asynchronous, fully distributed, and should use locally available information (i.e. information about the network should be obtained by a node by communicating only with its neighbours).
This chapter is organized as follows: Section 6.1 presents topology control techniques based on the use of hierarchical networks. Section 6.2 focuses on topology control techniques aimed at dynamically adjusting the transmission range of the WSN nodes. Finally, Section 6.3 focuses on topology control solutions for the construction and maintenance of faulttolerant topologies.
6.1. Topology control techniques based on hierarchical networks
The mechanisms for data transmission on a given network can make use of a flat approach (whereby all nodes are at the same level) or a hierarchical one (whereby nodes are organized into different hierarchical levels).
In a typical WSN for data collection, a flat network approach may lead to a reduced network lifetime. Let us assume that a multi-hop routing protocol (see Chapter 7) is used in all the WSN nodes for data delivery from the sensor nodes towards the sink. Since the nodes close to the sink would relay more traffic than those far from the sink, the former ones would rapidly become unavailable due to battery depletion. An opposite approach would be to assume that all sensor nodes can directly transmit their data to the sink. In this case, the nodes distant from the sink would run out of battery quickly, due to the use of higher transmission power.
One solution is to establish a hierarchy in the network based on the use of clusters. In a cluster-based network, nodes organize themselves into local clusters. One of the nodes of the cluster is selected as the cluster-head, which acts as the gateway of the cluster for communication between nodes that do not belong to the same cluster (see Fig. 6.1). The cluster-head may communicate with other cluster-heads (of other clusters) or may directly transmit data to a particular destination (e.g. a sink node). In data collection WSNs, a clusterhead may also perform data processing in order to minimize the redundancy of the information collected by the sensor nodes of the cluster.
Fig. 6.1. Nodes of a network organized into clusters. The cluster-head acts as a gateway for the nodes that belong to its cluster In WSN applications where the traffic is mainly between arbitrary nodes (e.g. as in building automation), clustering can also be useful for limiting the scope of certain mechanisms in order to avoid involving the whole network.
One example may be the transmission of routing protocol messages. In effect, clustering favours network scalability.
The Low Energy Adaptive Clustering Hierarchy (LEACH) protocol  is an adaptive clustering protocol for data collection WSNs16. In this protocol, sensor nodes collect information and transmit it to the corresponding clusterhead. The latter carries out data fusion operations and then transmits the information directly to a sink node (see Fig. 6.2). While transmission from a cluster-head to the sink may require high transmission power (especially for cluster-heads distant to the sink), only a small percentage of nodes are elected as cluster-heads. The overall result is that network lifetime is significantly longer than that obtained without the use of clustering mechanisms. A key Note that the description of the MAC mechanisms used in LEACH can be found in
element in LEACH is that in order to avoid fast battery depletion of clusterheads, these cluster-heads are randomly chosen, taking into account the remaining energy at the nodes. On the other hand, sensor nodes attach themselves to the cluster that minimizes communication energy.
Fig. 6.2. Example of a network with LEACH. The cluster-heads receive data from the nodes of their clusters, and after being processed the data are transmitted by the cluster-heads directly to the sink LEACH operation is organized into rounds. Each round starts with an advertisement phase in which nodes elect themselves as cluster-heads with a certain probability. The algorithm used for cluster-head election takes into account the percentage of cluster-heads required for the network (which has to be set a priori) and whether a node has recently been a cluster-head or not. Once a node has become a cluster-head, it broadcasts an advertisement message to the other nodes. On the basis of the received signal strength of the advertisement, the sensor nodes decide to which cluster they will belong. Subsequent to this decision, each sensor node informs the cluster-head that it will be a member of the cluster. The cluster-head then creates a TDMA schedule for transmission of the sensor nodes (see Chapter 4). After receiving all the data, the cluster-head performs signal processing operations to avoid redundant information and transmits the data to the sink node. After a certain period of time, a new round starts.
6.2. Topology control techniques based on dynamic transmission range adjustment A taxonomy of topology control techniques based on dynamic transmission range adjustment is proposed in  (see Fig. 6.3). First, these techniques can be divided into homogeneous and non-homogeneous approaches. In the first case, nodes are assumed to use the same transmission range, while in the second one nodes may use different transmission ranges. In nonhomogeneous topology control, the topology can be computed using location information (assuming that the positions of the nodes are known), direction information (whereby nodes do not know their position, but they can estimate the relative direction of their neighbours) and neighbourbased information (where nodes know the identities of their neighbours and can order them according to parameters such as distance and link quality).
A brief overview of some types of the three categories is shown below.
Fig. 6.3. A taxonomy of topology control techniques based on dynamic transmission range assignment (reproduced from ) 6.2.1. Location-based topology control protocols The authors in  propose a distributed topology control algorithm that uses location information obtained by low-power GPS receivers. The algorithm minimizes the energy needed for communication with a central node.
Local Minimum Spanning Tree (LMST)  is a protocol that generates a strongly connected communication graph with a node degree17 bounded by 6, that is, high network connectivity is achieved with a maximum number of neighbours per node equal to 6. In LMST, each node builds its own Minimum Spanning Tree (MST) (i.e. the minimum topology that connects all the network nodes) and keeps one-hop on-tree nodes as its neighbours in the topology. LMST outperforms other protocols, such as , in terms of average node degree and node transmitting range. LMST assumes that the location of nodes is known (e.g. thanks to the use of a system like GPS).
6.2.2. Direction-based topology control protocols
Cone Based Topology Control (CBTC)  is a distributed topology control based on directional information. Each node transmits with minimum power such that there is at least one neighbour in every cone of angle α centred at the node. The authors demonstrate that connectivity is ensured if α ≤ 2π/3 and, if α ≤ π/2, every node in the final communication graph has a node degree of at most 6.
6.2.3. Neighbour-based topology control protocols
Neighbour-based topology control protocols are based on connecting each node to its k closest neighbours. The MobileGrid protocol  and the Local Information No Topology (LINT) protocol  try to keep the number of neighbours of a node within a range centred on an optimal value. The transmission range is tuned accordingly. These protocols estimate the number of neighbours by overhearing control and data messages, which does not generate control overhead but rather depends on the transmission activity of the nodes.
The k-NEIGH protocol  maintains the node degree at a value smaller than or equal to a given value k. A simulation-based study has shown that k=9 is sufficient to obtain high network connectivity for a range of node densities . The same simulation results show that the topology generThe node degree is defined as the number of neighbours of a node.
ated by k-NEIGH is on average 20% more energy efficient than that generated by CBTC.
6.3. k-connectivity One of the main goals of topology control is the construction and maintenance of a connected network topology. This includes the problem of assuring a k-connected network topology, where k is the number of different paths between any two nodes. Such a topology would still assure connectivity between any two nodes, even if k-1 nodes became unavailable.
Path redundancy is indeed an important feature required in WSNs, which are subject to node unavailability (e.g. due to battery depletion), node mobility (in some scenarios) and the dynamics of radio propagation. Another relevant factor to take into consideration is the active and sleep time schedules that may be used by the WSN nodes, which may also affect network connectivity.
There exist some extensions to protocols like LMST and CBTC that offer k-connectivity [9, 10].
 P. Santi, “Topology Control in Wireless Ad hoc and Sensor Networks”, ACM Computing Surveys, Vol. 37, No. 2, pp. 164–194, June 2005.
 W. Rabiner Heinzelman, A. Chandrakasan, H. Balakrishnan, “EnergyEfficient Communication Protocol for Wireless Microsensor Networks”, in proceedings of the 33rd Hawaii International Conference on System Sciences, Island of Maui, Hawaii, USA, January 2000.
 V. Rodoplu, T. Meng, “Minimum energy mobile wireless networks”, IEEE J. Selected Areas Communications. 17, 8, pp.1333–1344.
 N. Li, J. Hou, L. Sha, “Design and analysis of an mst-based topology control algorithm”, in Proceedings of the IEEE Infocom, pp. 1702–1712, 2003.
 R. Wattenhoffer, L. Li, P. Bahl, Y. Wang, “Distributed topology control for power efficient operation in multihop wireless ad hoc networks”, in Proceedings of IEEE Infocom., pp. 1388–1397, 2001.
 J. Liu, B. Li, “Mobilegrid: Capacity-aware topology control in mobile ad hoc networks”, in Proceedings of the IEEE International Conference on Computer Communications and Networks, pp. 570–574, 2002.
 R. Ramanathan, R. Rosales-Hain, “Topology control of multihop wireless networks using transmit power adjustment”, in Proceedings of IEEE Infocom, pp. 404–413, 2000.
 D. Blough, M. Leoncini, G. Resta, P. Santi, “The k-neighbors protocol for symmetric topology control in ad hoc networks”, in Proceedings of the ACM MobiHoc 03, pp. 141–152, 2003.
 N. Li, J. Hou, “Flss: a fault-tolerant topology control algorithm for wireless networks”, in Proceedings of ACM Mobicom, pp. 275–286, 2004.
 M. Bahramgiri, M. Hajiaghayi, V. Mirrokni, “Fault-tolerant and 3-dimensional distributed topology control algorithms in wireless multihop networks”, in Proceedings of the IEEE International Conference on Computer Communications and Networks, pp. 392–397, 2002.
7. Routing As in other types of wireless multi-hop networks, in WSNs, if the receiver is not within the transmission range of the sender, the sender can take advantage of intermediate nodes which can route the data towards the receiver. This reduces the amount of energy required to transmit data between two nodes. In addition, communication path redundancy is commonly present in WSNs to some extent, which provides reliability.
Network routing is the process of selecting a path for the relaying of a message from a source device to the intended destination.
Ideally, data should be routed through good paths. Routing protocols are in charge of finding and maintaining such paths. The constraints of WSNs pose a set of requirements for the routing protocols, which should aim at finding a good trade-off between a number of performance parameters, such as delivery ratio, latency and energy consumption. Scalability is particularly important, given the potential large number of nodes in a network and their memory and energy limitations.
This chapter focuses on routing protocols for WSNs. Section 7.1 presents various types of routing protocols which were designed for or are particularly suitable for WSNs. Section 7.2 focuses on adaptations of IETF Mobile Ad Hoc Network (MANET) routing protocols for WSNs.