Link Level Capacity Analysis in CR MIMO Networks

Size: px

Start display at page:

Download "Link Level Capacity Analysis in CR MIMO Networks"

Marian Griffin
5 years ago
Views:

Volume 114 No. 8 2017, 13-21 ISSN: 1311-8080 (printed version); ISSN: 1314-3395 (on-line version) url: http://www.ijpam.eu Link Level Capacity Analysis in CR MIMO Networks 1M.keerthi, 2 Y.

1 Volume 114 No , ISSN: (printed version); ISSN: (on-line version) url: Link Level Capacity Analysis in CR MIMO Networks 1M.keerthi, 2 Y.Prathima Devi, 3 A.Ravi Kumar reddy. marrikeerthi8@gmail.com 1 ijpam.eu Apr 14, Abstract In this paper we are going to investigate capacity analysis of a CR network with fading and without fading. A CR network is proposed that can operate well in a primary and secondary band. AVAA is considered to allow multiple transmissions between the CR and secondary users. A single VAA case is considered to estimate the link level capacity. The capacity of a real MIMO and virtual MIMO is also compared. Different environment and scenarios are considered by taking different number of users and different fading environment. Keywords: CR networks, Link level capacity, Virtual MIMO, VAA 1 INTRODUCTION Radio spectrum (30 KHz-300GHz) is used in many applications. There is increase in number of user s every day and in future also the dramatic growth continues due to this the demand for spectrum is also increasing. The competition for spectrum below 3GHz is more (3-4GHz 0.5% spectrum is used and 4-5GHz 0.3% spectrum is utilized) and the spectrum band allocated for certain users are not always utilizing and is idle sometimes [1]. All these factors led to spectrum congestion. To overcome these drawbacks Mitalo proposed cognitive radio (CR) [2]. It is split into full CR and spectrum sensing CR. The important task of a CR is to improve spectrum usage and this can be achieved by DMA (Dynamic memory access) where the spectrum is accessed dynamically according to the user s need [3]. The unused frequencies are sensed through DMA technique. CR main objectives are sensing, understanding, adapting and deciding. In the CR the secondary user or unlicensed user utilize the 13

2 spectrum when it is idle. The secondary user has to continuously sense the spectrum and evacuate when the primary user or licensed user activity is detected [4]. CR s are grouped together into a network with this data can be retransmitted which improves coverage and energy efficiency. Basically CR networks are classified as underlay, overlay and interweave networks [5]. Overlay networks are considered in this paper. With a slight change in transmission behaviour, interference and density the capacity of CR networks changes. In order to expand capacity a hybrid CR network is proposed, which combines both CR network and licensed network. The architecture of this is classified as cooperative and non-cooperative hybrid networks. Non cooperative: - Has independent physical layers operating separately in licensed and unlicensed bands. Capacity of this network is the sum of licensed and unlicensed physical layers capacity [6]. The communication is better for only short to medium ranges. Cooperative CR:-Single physical layer. Licensed band transmits the data to destination and unlicensed band helps in the transmission. Cooperative is better than non-cooperative and communication is possible from medium to long ranges. Cooperative network has centralized and distributed approach. MIMO is an multiple antenna communication which is used at both the transmitter and receiver With this the errors are minimized and the data speed is optimized.it is one of the forms of many smart antenna technology such as MISO, SISO, SIMO and with use of single transmitting antenna and the single receiving antenna then these many problems such as multipath fading and reaction of data speed to counter attack these problems.multiple antennas such as two or more antennas is used. The major advantages of MIMO include array gain, spatial diversity and suppression of interference. In an MIMO CR network two antennas are placed at CR transmitter and receiver. The data rate between the CR networks can be improved by placing MIMO antennas in between primary users and secondary users. With this multiple transmissions can be possible and capacity is increased. Virtual MIMO can be created by an cooperative communication. If VAA is present on the both sides of the communication it is a virtual MIMO. VAA is an process f combining or grouping multiple terminals which are used for communication into one with this virtual MIMO we can use all neighbouring terminals. By this the size of channel sources which would result in 14

3 achieving spatial diversity gain. With use of maximum like hood and these types the interference cancellations is possible and high capacitance and BER performance can be achieved. Capacity of a channel can be estimated by the Shannon theorem. In this paper to achieve high channel capacity VAA is considered. In this a single VAA case is considered in which the in a single cell multiple users will be present. The users receive the information from the base station and the user takes the information it needs and transmits the remaining information to its neighboring users. In [7,8] 2 SYSTEM MODEL Multi antenna base station communicates with a cell which consists of multiple single antenna cognitive radio users distributed in it. For communication between a BS (base station) and cognitive radio users licensed band is utilized. For selforganizing of neighboring cognitive radio users into multiple virtual antenna array groups unlicensed band are utilized. Fig. 2.1 System model for link level capacity. In the above figure two stages are considered.the main CR user transmits information to other users in the virtual antenna array in the unlicensed band. This would protect the co-channel licensed network from the interference caused by the CR. Here the first stage ends and in second stage the VAA is considered as one and communicates with the base station in a licensed band. 3. Working Scenario There are four assumptions that are considered: A maximum power P is considered to be transmitted in the unlicensed band to protect licensed network[5]. The maximum power is ρ max = I 0(d min ) 4 (h c h p ) 4 (1) I 0 is the average interference constraint. The CR transmitter and receiver distance is taken as d i andd min = min {d i }. 15

4 For relying process an Amplify and forward scheme is used. All transmissions in unlicensed bandare done by using TDMA schme. CRusers can operate independently or simultaneously in the licensed and unlicensed band. In the central access type CR network the base station and the target CR user transmits atρ max. They are transmitted based on TDMA scheme. The gain of channel power is h A = g A p g A A s g m (2) g A A p is the path loss of the power gains, g m is the multipath fading A and the g s is the shadowing. We consider the log normal shadowing in this paper. In the virtual MIMO based network we need to establish a VAA. In order to establish VAA a protocol is proposed to achieve this. CR user transmits the data with power P in broadcast channel [6].The secondary users which are idle continuously monitors the broadcast channel and valuates the SNR.If the SNR is above the threshold, the CR users which are idle sends the feedback information to CR.The feedback information obtained from the idle cognitive radio user is used by the main CR to select a channel with good qualities to form a VAA only high quality signals and signals with SNR is forwarded in the second stage which prevents propagation of harmful noise[6]. The selected idle users are tuned to allocate TDMA channels to receive the data. M t is represented as the number of single antenna users in virtual antenna array. 4. Mathematical Analysis The virtual MIMO link capacity if the channel knowledge is not known at the transmitter is derived as C = E log2[det (I P bs P CR Mr R M t 1+P nn HH)] (3) cr Where expectation operator is E(.),(H)is the Hermitian Transpose, I Mr is the identity matrix, R nn is the covariance matrix of the noise vector n, P bs is the average SNR at the base station, P CR is the average SNR at the cognitive radio users. R nn = I Mr + (I Mr + P bs 1 M t 1+P cr R n^n HH (4) WhereR n^n = diag[1,1,1..0] H=HH (5) where H=diag(exp(iθ1), exp(iθ2),.. exp(iθmt-1),1) is the cooperation phase matrix and H is the primary band MIMO. The three parameters that determine the link capacity p cr are average SNR at cognitive radio users, at base station and the 16

5 no.of transmit or receive antennas. We assume the different number of antennas at the receiver. For comparison purposes we consider the normal MIMO channel. 5. Fading C(real) = E log2[det (I Mr + P bs M t HH)] In communication theory to model the scattered signals which reach the receiver by multipath Rayleigh, Rician, Nakagami, Weibull Distributions are used. The fading characteristics of the signals are dependent on the density of scatter. Dense scatter are modelled by Rayleigh and Nakagami whereas the signals with LOS Rician fading is used. In Rayleigh, there are so many objects in the nature which would scatter the signal. This fading is the reasonable model for them such as the urban environment. The PDF of amplitude d= g = α then the g(t) components are independent. This has Rayleigh PDF of f d = d d 2 σ 2 e( 2 σ 2) Where E{d 2 } = 2σ 2 and d 0. As we don t consider because Rayleigh fading don t have the LOS (line of sight) this represent the worst fading case. This is used in wireless communication. The phase is distributed uniformly and the power is distributed in exponential form. The phase and amplitude are independent to each other.in rician case there is LOS with complex gaussian channel with mean of non-zero andthe envelope d= g is Rician distributed. We denote g = α e (jϕ) +υ e (jθ). Where α still follows Rayleigh distribution, υ > 0 is constant such that υ 2 is the LOS signal component power. The angles θ and are mutually independent and distributed uniformly on [-Π, Π). Rician PDF f d = d 2 +υ 2 σ 2 e( d 2σ 2 ) I. ( dυ σ 2) Where I is the zero order modified Bessel function and 2σ 2 = E {d 2 }. The Rician factor k1 = υ2 2σ2.The rician factor is defined as the relation between the power of LOS and Rayleigh components.when k1-> there is no LOS. Then Rayleigh and rician are equal.the Nakagami pdf f d = 2 2k1 1 k1 rd Γ(k1) 2σ 2 exp ( k1d2 2σ 2 ) and d 0 Where 2σ 2 = E{d 2 }, Γ. is the gamma function and k1>=1/2, it 17

is originally developed based on measurements. When k1=1 Nakagami and Rayleigh are equal. In weibull distribution, the envelope of d = (A 2 + B 2 ) 1 2 is Rayleigh distributed.

6 is originally developed based on measurements. When k1=1 Nakagami and Rayleigh are equal. In weibull distribution, the envelope of d = (A 2 + B 2 ) 1 2 is Rayleigh distributed. When the envelope is d = (A 2 + B 2 ) y the PDF is Weibull distributed. f d = ydy 1 dy exp ( 2σ2 2σ 2) 6. Numerical Results and Discussions In the figure 6.1, we considered avaa link level capacity with MIMO system. Channel capacity is estimated by using different cases where different number of transmitting and receiving antennas. We considered mt as transmitting users and mr as the receiving users. Different cases are considered mt=1,mr=2;mt=2,mr=2,mt=4,mr=3,mt=8,mr=8. We concluded that by increasing in the number of transmission antennas and receiving antennas the capacity increases. In the fig6.2, we then included some fading channels such as Rayleigh, Ricean, weibull andnakagami.the weibull fading has high capacity compared to other fading channels even in the case of similar SNR s. In this for weibull the lambda value is taken as 3 and beta value as 2 and in the nakagami mu value is taken as 3 and omega=1. After weibull, Ricean has the high capacity compared to nakagami and Rayleigh. Nakagami and Rayleigh have similar capacity but nakagami fading has more capacity. 1 18

Fig. 6.1 Link level capacity for different Fig 6.

3 Capacity with and without fading Fig 6.4 Real MIMO vs virtual MIMO In fig 6.

fading has high capacity with similar SNR values. In fig 6.

7 Fig. 6.1 Link level capacity for different Fig 6.2 The link level capacity for different number of user fading channels. Fig 6.3 Capacity with and without fading Fig 6.4 Real MIMO vs virtual MIMO In fig 6.3, we included channels without fading and with different fading then the channel without fading has high capacity with similar SNR values. In fig 6.4, We compared the real MIMO with the virtual MIMO and the virtual MIMO have high capacity compared to the real MIMO when taken for different number of users such as mt=1,mr=1;mt=2,mr=2,mt=3, mr=3. Individually there 19

8 is an increase in the capacity but when compared with each other the proposed system i.e.., the virtual MIMO has high capacity. ρ bs is taken as 15, ρ cr is taken as 35. The users are considered to be distributed in a uniform direction. 6. Conclusion In this paper we discussed about the how to improve the capacity through link level analysis. We have taken different transmitting and receiving users to estimate the capacity and how the capacity is improves with increase in the more transmitting users also. We introduced some fading channels into this and then finally concluded that the weibull fading have high capacity. We have considered a VAA and introduced fading into it. By increase in the number of users and multiple transmissions between different cells in a VAA also there is an increase in the capacity. 7. References [1] Federal Communications Commission, Unlicensed operation in the TV broadcast bands, ET Docket No , [2] J. Mitola and G. Maguire, Cognitive radio: making software radios morepersonal, IEEE Personal Commun. Mag., vol. 6, no. 6, pp , Aug [3] Ian F.Akyildiz, Won-Yeol Lee, MehmetC.Vuran, Shantidev Mohanty NeXt generation/dynamic spectrumaccess/cognitive radio wireless networks: A survey Elsevier Computer Networks (2006) [4] Cognitive Radio: Brain-Empowered Wireless Communications Simon Haykin, Life Fellow, IEEE journal on selected areas vol. 23, no. 2, FEB [5] Xuemin Hong, Cheng-Xiang Wang, Hsiao-Hwa Chen, and Yan Zhang Secondary spectrum access networks Recent Developments on the Spatial Models IEEE vehicular technology magazine 2007 [6] Cheng-Xiang Wang, Senior Member, IEEE, Xuemin Hong, Member, IEEE, Hsiao-Hwa Chen, Senior Member, IEEE, and John Thompson, Member, IEEE. On Capacity of Cognitive Radio Networks with Average Interference Power Constraints IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 8, NO. 4, APRIL

9 21

10 22

Abstract. Keywords - Cognitive Radio, Bit Error Rate, Rician Fading, Reed Solomon encoding, Convolution encoding.

Abstract. Keywords - Cognitive Radio, Bit Error Rate, Rician Fading, Reed Solomon encoding, Convolution encoding. Analysing Cognitive Radio Physical Layer on BER Performance over Rician Fading Amandeep Kaur Virk, Ajay K Sharma Computer Science and Engineering Department, Dr. B.R Ambedkar National Institute of Technology,