**3. Broadcast problem and system model**

In this section, the broadcast problem, the system model and the basic assumptions are presented.

#### **3.1. Broadcast problem in CR networks**

To further illustrate the challenges associated with broadcast in CR ad hoc networks, consider simple single-hop broadcast topology for traditional and CR network shown in **Figures 1** and **2**, respectively, where *node A* is the source node with *N* neighbours.

In traditional ad hoc networks, all nodes can tune to the same channel due to the uniformity of channel availability. Therefore, broadcasting a message can be easily implemented over a single common channel as all nodes receive messages from the same channel. As shown in **Figure 1**, *node A* only needs to broadcast over a single channel to deliver the broadcast message to all its neighbouring nodes.

However, in CR ad hoc networks, the opportunity of a common channel available for all CR nodes may not exist. In addition, different CR users might acquire different channels at different times. Therefore, broadcasting in cognitive radio ad hoc networks is a much more challenging task. As shown in **Figure 2**, to deliver the broadcast message to all the neighbouring

**Figure 1.** Single-hop broadcast topology for traditional network.

Reliable Broadcast over Cognitive Radio Networks: A Bipartite Graph-Based Algorithm http://dx.doi.org/10.5772/intechopen.69216 11

**Figure 2.** Single-hop broadcast topology for CR network.

nodes, node A needs to transmit the broadcast message to different channels. In the worst case, each neighbouring node may tune onto a different channel. Consequently, the source node has to broadcast over all the channels.

In fact, the reliable broadcasting in CR ad hoc networks depends on connecting different local topologies. Hence, the broadcast channel(s) should be carefully and dynamically allocated in order to secure a reliable communication in CR networks.

#### **3.2. Network model**

where a set of channels and nodes are selected to convey a message from the source node to its neighbours. The authors in Ref. [19] propose a simple heuristic algorithm to transmit the messages between CR nodes in CR ad hoc networks. In this work, CR nodes are assumed to

In this section, the broadcast problem, the system model and the basic assumptions are presented.

To further illustrate the challenges associated with broadcast in CR ad hoc networks, consider simple single-hop broadcast topology for traditional and CR network shown in **Figures 1** and **2**,

In traditional ad hoc networks, all nodes can tune to the same channel due to the uniformity of channel availability. Therefore, broadcasting a message can be easily implemented over a single common channel as all nodes receive messages from the same channel. As shown in **Figure 1**, *node A* only needs to broadcast over a single channel to deliver the broadcast mes-

However, in CR ad hoc networks, the opportunity of a common channel available for all CR nodes may not exist. In addition, different CR users might acquire different channels at different times. Therefore, broadcasting in cognitive radio ad hoc networks is a much more challenging task. As shown in **Figure 2**, to deliver the broadcast message to all the neighbouring

be equipped with multiple transceivers to broadcast to multiple channels.

respectively, where *node A* is the source node with *N* neighbours.

**3. Broadcast problem and system model**

**3.1. Broadcast problem in CR networks**

10 Cognitive Radio

sage to all its neighbouring nodes.

**Figure 1.** Single-hop broadcast topology for traditional network.

A CR ad hoc network with no centralised coordinator is considered. Hence, network environment tasks like channel selection, neighbour discovery and spectrum sensing are individually accomplished by the CR users.

We consider a set of *N* cognitive radio (CR) nodes {*CR*<sup>1</sup> , *CR*<sup>2</sup> , …, *CRn* } and a set of *M* primary radio (PR) nodes {*PR*<sup>1</sup> , *PR*<sup>2</sup> , …, *PRm*} in the same geographical area. Primary radio nodes are the licenced users, and they can access their respective licenced bands without any restriction. While CRs can access licenced bands opportunistically, i.e. they are allowed to use the idle licenced bands only if they do not interfere with ongoing PR transmissions.

Note that an idle state describes the temporal availability of a channel. To prevent interference, CR users are capable of sensing spectrum opportunities using energy detectors, cyclostationary feature extraction, pilot signals or cooperative sensing [5].

A set of *K* nonoverlapping orthogonal frequency channels (*Cglobal = C*<sup>1</sup> , *C*<sup>2</sup> , …, *Ck* ) is considered, which may be freely occupied by the PR users. Each CR node knows the global channel set *Cglobal* and can operate on a subset *Clocal* of this global channel set depending on the local channel availability at that node, where *Clocal* ⊆ *Cglobal*. For simplicity, it is assumed that all channels have the same capacity. However, the proposed protocol can be easily extended to channels of different capacities.

In this model, it is assumed that CR nodes are equipped with half-duplex transceivers that can either receive or transmit (not both) on a single channel at any given time. Each CR can swiftly hop between channels using software-defined radio (SDR) technology. The utilisation of a single transceiver reduces the operational cost of the CR device [6], as well as avoiding any potential interference between adjacent transceivers due to their close proximity [7].

Throughout this chapter, it is assumed that the channel availability is relatively stable (i.e. during a short period of time, channel status does not change). Therefore, the proposed protocol is more suited to the case of temporal underutilisation and spatial spectrum underutilisation when the activity of PR user is not very dynamic. The main notations used in the chapter are summarised in **Table 1** for easy reference.

### **3.3. Sensing spectrum holes**

Spectrum sensing aims to identify the available spectrum and prevent any harmful interference to the primary users. It is assumed that CR nodes periodically perform spectrum sensing to ensure up-to-date information regarding the PR activity and identify the available channels.

In addition, it is assumed all CR nodes are synchronised and follow the same sensing cycles. In the sensing period, no transmission is allowed, and all CR nodes must be silent. Therefore, the time needed to deliver a packet in the network may be influenced when the CR nodes are banned from transmission due to the imposition of the silent duration.

The transmission time and the spectrum sensing for every CR user are *Tt* and *Ts* , respectively, where *Tt* is the effective duration of time for which transmission is allowed for any CR node


**Table 1.** Symbols used for OBA description.

on any choice of free spectrum, while *Ts* is the duration of time that all CR nodes must be silent for the purpose of sensing. *Ts + Tt* gives the frame time for each user when considered together.

#### **3.4. Discovering CR neighbouring nodes**

availability at that node, where *Clocal* ⊆ *Cglobal*. For simplicity, it is assumed that all channels have the same capacity. However, the proposed protocol can be easily extended to channels

In this model, it is assumed that CR nodes are equipped with half-duplex transceivers that can either receive or transmit (not both) on a single channel at any given time. Each CR can swiftly hop between channels using software-defined radio (SDR) technology. The utilisation of a single transceiver reduces the operational cost of the CR device [6], as well as avoiding any potential interference between adjacent transceivers due to their close proximity [7].

Throughout this chapter, it is assumed that the channel availability is relatively stable (i.e. during a short period of time, channel status does not change). Therefore, the proposed protocol is more suited to the case of temporal underutilisation and spatial spectrum underutilisation when the activity of PR user is not very dynamic. The main notations used in the chapter

Spectrum sensing aims to identify the available spectrum and prevent any harmful interference to the primary users. It is assumed that CR nodes periodically perform spectrum sensing to ensure up-to-date information regarding the PR activity and identify the available channels. In addition, it is assumed all CR nodes are synchronised and follow the same sensing cycles. In the sensing period, no transmission is allowed, and all CR nodes must be silent. Therefore, the time needed to deliver a packet in the network may be influenced when the CR nodes are

is the effective duration of time for which transmission is allowed for any CR node

and *Ts*

, respectively,

banned from transmission due to the imposition of the silent duration.

The transmission time and the spectrum sensing for every CR user are *Tt*

of different capacities.

12 Cognitive Radio

are summarised in **Table 1** for easy reference.

**Symbols Descriptions** *N* Set of CR nodes

*G*(*X*, *Y*, *E*) Bipartite graph

**Table 1.** Symbols used for OBA description.

*Cglobal* Total number of channels *Ci* The available channel set of *CRi Ni* Set of single-hop neighbours of *CRi Ts* Spectrum sensing time for CR users *Tt* Transmission time for CR users

*PR* Transmission range of PR users

*CR* Transmission range of CR users

*BCSi* Broadcast channel set of *CRi TCi* Tuning channel of *CRi*

**3.3. Sensing spectrum holes**

where *Tt*

*Crr*

*Crr*

To successfully deliver the broadcast messages to all the CR nodes in each neighbourhood, CRs must discover the network topology and the common idle channels that can be used to communicate among neighbours; these tasks are typically undertaken during the neighbour discovery.

In the absence of a common control channel, discovering neighbours in CR Ad Hoc Network (CRAHN) is undoubtedly a challenging task; we propose a neighbour discovery mechanism to address this issue. Initially, it is assumed that individual nodes are tuned to different channels and have no prior knowledge of their neighbours and the network topology. Furthermore, each CR node maintains the local idle channel list based on the information received from the spectrum sensing.

At the beginning of constructing the network, every CR node has to beacon its information (node's id and its available channels) onto all the locally available channels, one by one. As a result, all single-hop neighbours that are tuned to any idle channels are able to receive a copy of this message. Each CR node receives this beacon message and records the transmitter's CR node information in its single-hop neighbours list *Ni* . After forming and configuring the network, the CR nodes do not have to beacon messages unless there is a change in their channel availability.
