System Model

We consider a distributed multi-hop CR-WMNs, consisting of N_p primary users and N_c secondary users. The primary users are stationary nodes that have li-cense to access a given spectrum portion. We suppose that there are total M numbers of licensed channel, for each channel we have one PU, i.e. N_p=M. The bandwidth of each channel is W^k(k = 1,2,3...M). In this paper, we assume the primary system is a small power CSMA type wireless network like wireless LAN, which has a conventional type carrier sense function. The secondary system is a system , which has high sensitive sensing device[68]. For spectrum sensing there are several methods have been studied [28][29]. Energy detection is one of the very popular methods because of its simplicity and adequate performance [30].

The sensing function of our proposed method is based on this energy detection

5.3 System Model

Rtx

Rcs

RI

i d ^j

Carrier sensing area of the CR with lower sensing level

(-92dBm) Carrier sensing area of the PU with higher sensing level

(-62dBm) Transmission range /radio

coverage range

Interference range

Figure 5.1: Transmission range, interference range and carrier sensing range

method. This method detects unknown signals embedded in the noise by com-paring the observed received signal power level with a threshold. The SU users undertake periodic sensing and can reliably detect the presence of a PU. Both the channel usage ratio and the sensing information (SOP) is broadcasted by the SU,through beacons, to the other SU within its transmission range. The com-munication parameters of the primary wireless LAN are fixed. However, the SU changes the communication parameters such as the operating frequency, modu-lation scheme, code rate, and transmission power in real time [34]. An important goal of performing such communication parameters adaptation is to allow the cognitive radio to coexist with the primary networks without giving interference toward the spectrum used by the primary wireless LAN.Here, it is assumed that, each SU is equipped with one control radio and two data radios. The control radio is used for transmitting and receiving control messages (routing, spectrum sensing) works on the Common Control Channel (CCC). The control radio is a high sensitive sensing device with lower sensing level has larger carrier sensing area. It can sense the PUs in their carrier sensing area to get the SOP. The data radios are used for data transmission. In addition, we assume that CCC is common to all the SUs. The data radios are works on the data channels which are licensed channels.

Let the multi-hop CR-WMN is modeled by the general topology graphG(V, S, E) that has a set of cognitive users,V = (n₁, n₂, n₃...n_Nc), a set of SOP(spectrum op-urtunities) for the SUs,S = (SOP₁, SOP₂, ....SOP_Nc), and a set of network links E. A link (e_ijϵE), will exist between SUa,iand j if the two nodes are within the transmission range of each other. There must be at least one available common channel among their respective SOPs, i.e. (SOPi ∩SOPj)̸= 0. We assume the sensing function of the primary system is a conventional carrier sensing device.

5. SPECTRUM AWARE ROUTING FOR COGNITIVE RADIO WIRELESS MESH NETWORKS

The primary system has high sensing level such as -62dBm. We also assume that each cognitive user is equipeed with a high sensitive sensing device with lower sensing level such as -92dBm. The transmission range Rtx, carrier sensing range, R_cs and interference range,R_I shown in Fig.5.1.

We assume the SU in the CRN can only transmit when the PUs are not active or idle . The availability of the primary band for transmissions by the SU is characterized by an ON-OFF process. An ON period T_on^k represents the time duration that the PU is active on channel k . An OFF period T_{of f}^k represents the time duration that the PU is not active or idle on channel k. We model the channel usage ratio of PU on channel k (channel unavailable for secondary transmission) is defined as,

α^k= T_on^k

T_{of f}^k +T_on^k . (5.1)

The probabilty of the PU not transmitting (channel available for the secondary user) is thus (1−α^k). Hence the channel usage ratio of the SU, can be defined as,

β^k= 1−α^k . (5.2)

The radio coverage area of a PU indicates the transmission area of the PU. If any SU is located within the radio coverage area of a PU, the transmission of that SU is affected by the PU. Let, θ_n = (a₁, a₂, a₃....a_M) denote the activation indicator of each SU,n = (1,2,3...N_c). It can be defined as,

a_k =

{ 1 If SU is located inside the PU coverage area on channel k

0 Otherwise (5.3)

Consider a transmission linke_ij between two SUs, iandj, that ared(i, j) distant from each other. Let, P_r,dBm^k (i, j) be the received signal power at SU, j over channelk. Assuming a simple path loss propagation model the received signal at the intended receiver j from the SU i is determined by[69]:

P_r,dBm^k (i, j) =P_t,dBm+ 20 log₁₀ λ

4πd_o + 10ηlog₁₀ do

d(i, j). (5.4) Where P_t,dBm is the transmission power, d_o is the reference distance, λ is the wavelength of the signal andηis the path loss exponent. The capacity of the link eij on channel k is given by Shannon’s Theorem:

C_ij^k =W^klog₂(1 +SN R^k_ij) . (5.5)

5.3 System Model

PU transmission range

CRa

Tx CRb

Rx

CR transmitter carrier Sensing range

CR receiver carrier Sensing range

CR transmission range

PEN

(a)

CR transmission range

CRa PHN Tx

CRb Rx

PU transmission range

CR receiver carrier Sensing range CR transmitter carrier

Sensing range (b)

Figure 5.2: (a)Primary Exposed Node.(b)Primary Hidden Node

where W^k is the bandwidth of channel k,and SN R^k_ij is the signal to noise ratio for linke_ij on channel k. It can be computed as:

SN R^k_ij = P_r,dBm^k (i, j)

N . (5.6)

where N is the bachground noise. We consider each link is characterized by a utility function. The purpose of using utility function is to find the best route with minimum end to end transmission delay or minimum end end to end interference.

The utility function,U_ij^k for the link e_ij on channel k is defined as,

U_ij^k =C_ij^k ·β^k . (5.7)

5. SPECTRUM AWARE ROUTING FOR COGNITIVE RADIO WIRELESS MESH NETWORKS

5.4 Primary Exposed Node and Primary Hid-den Node

In the multi-hop CR-WMNs the effectiveness of the secondary networks is de-pends on the activity of the primary user. Due to the inefficiency of the carrier sensing in the multi hop CR-WMNs, the Primary Exposed Node (PEN) and Pri-mary Hidden Node (PHN) may occur. The sensing device of the priPri-mary user is a conventional carrier sensing device used in the wireless LAN. That is why, the sensing sensitivity is low and the sensing level becomes high. On the other hand, the sensing device of the secondary user is a high sensitive sensing device which can detect the signals under the noise level[68]. In this paper, such high sensitive sensing devices have been used by the cognitive radio for the spectrum sensing.

The carrier sensing range is an area which can detect the signals from the spec-trum shared different transmitter. The carrier sensing range of the conventional primary network like the wireless LAN is short due to the low sensitive sensing device. On the other hand, the carrier sensing range of the secondary user is longer due to the high sensitive sensing device. The carrier sensing range of the primary user and the secondary user are different. The carrier sensing range of the primary user and the secondary user is shown in the Fig.5.1.

The primary exposed node (PEN) may occur in the multi-hop CR-WMNs, when a CR is preventing from its data transmission due to its neighbor PU activation. The CR transmitter can detect the presence of the PU, because the CR carrier sensing range includes the positions of the PU. When the CR transmitter detects the presence of the PU, it has to refrain from transmitting.

On the other hand, the CR receiver can not detect the situation of the PU. The reason behind this, the primary user is out of the CR receiver carrier sensing range. Therefore, if the transmitting side of the CR has to halt its transmission due the sudden appearance of the primary user, even though the primary user transmission does not interfere the CR receiver, that primary user is known as the primary exposed node (PEN). Figure.5.2a shows the example of the primary exposed node (PEN). The transmitting side of the secondary userCRais affected by the sudden appearance of the primary user. On detecting the presence of the primary user, CR_a defers its transmission to CR_b, while the receiver CR_b can not detect the presence of the primary user. In this case we can say that the link between the CR nodesCR_aand CR_b will break by the presence of the PEN.

Though the receiving sideCR_bis not influenced by the PEN, the transmitterCR_a stops its transmission and keeps waiting until the primary user vacant the channel.

Due to the PEN problem, the link utilization can be significantly impaired, which leads to low end to end throughput and high end to end delay.