Self-Scheduled MAC-Layer Protocol for Spectrum Sharing in Cognitive Radio Communication Shweta Pandit and G. Singh Department of Electronics and Communication Engineering, Jaypee University of Information Technology, Solan - 173234, India. Email:
[email protected],
[email protected] optimal frequency band usage [2]. It operates in the multiple frequency bands and maximize the utilization of the limited radio resources. This flexible use of the spectrum is provided by two new spectrum allocation method namely: dynamic spectrum access (DSA) [3] and opportunistic spectrum access (OSA) [4], which defines a set of techniques and models to support the dynamic management of the frequency bands for wireless communications systems. These new spectrum licensing methods improve the spectrum efficiency and enhance the performance of communication systems. Therefore, the dynamic spectrum access or opportunistic spectrum access is the key approach in the cognitive radio communication system, which is adopted by a cognitive radio user to access the radio spectrum opportunistically [2]. Cognitive radio systems are quickly reshaping the future of the wireless communications, sensing, and data sharing. The fundamental concept of the cognitive radio has been adopted from the software defined radio (SDR) [5-6], which can operate on multiple frequency bands without any hardware modifications, however the selection of frequency band and operating parameters is manually controlled by the user through software. The artificial intelligence part for learning and decision making is not available in SDR in contrast to cognitive radio, which is software defined radio along with the capability of sensing their environment and making decision such as about modulation scheme, transmission power etc. without the human intervention [7].
Abstract — Recently, the spectrum scarcity has become the bottleneck for the development of wireless communication. Therefore, cognitive radio is a promising technology geared to solve the spectrum scarcity problem by opportunistically identifying the unused portions of the licensed user’s spectrum and establish the communication in the unutilised regime of the spectrum, while ensuring that the licensed or primary users of the spectrum are not affected. However, one of the major aspect and potential challenge is in the licensed channels, the sensing and access decision. In this paper, we have explored the concept of the multiple access control protocol for the distributed cognitive radio network. In this control channel protocol, the secondary users share the sensing results to each other and each channel is divided into four intervals such as idle, sensing– sharing, contention, and transmission. The sensing-sharing and contention interval are further divided into number of slots and throughput of the communication system has been computed. Keywords—Cognitive radio, medium access control, contention, sensing and sharing, average throughput, average real throughput.
I.
INTRODUCTION
Fixed spectrum allocation policy, governed by Federal Communications Commission (FCC) [1] or any other governmental agency, results the spectrum scarcity problem because the allocated spectrum to the different services like cellular and TV broadcast services are not fully utilized. Thus, it is the wastage of available spectrum resources and some new services which want to enter the world might not get enough spectrums for their functioning due to the spectrum scarcity created by fixed spectrum allocation policy. However, the limitations of fixed spectrum allocation scheme have been discussed in detail in [2]. The spectrum utilization efficiency can be improved if the unlicensed users are allowed opportunistically to access the spectrum as long as their transmissions do not interfere with or interrupt the transmissions of the licensed users. Cognitive radio is a promising technology geared to solve the spectrum scarcity problem by opportunistically identifying the unused portions of the licensed user’s spectrum and establish communication in the unutilised regime of the spectrum, while ensuring that the licensed or primary users of the spectrum are not affected. Cognitive radio is a wireless communication device, which observe, learn, optimize and intelligently adapt to achieve
978-1-4799-0192-0/13/$31.00 ©2013 IEEE
Firstly, the cognitive radio should identify the spectrum holes [8] in the licensed user’s spectrum band and secondly, utilize them in flexible manner, according to a medium-access control (MAC) scheme. To identify the available spectrum resource through the spectrum sensing, decision on the optimal sensing, transmission time and proper coordination with other users for the spectrum access are the important characteristics of the medium access control protocol. However the, throughput maximization of a cognitive radio user with the help of frame structure is demonstrated in [9, 10]. Further, the fading channel consideration and its effect on cognitive user’s performance is presented in [11, 12]. Also, licensed user’s unutilized spectrum has been shared using the game theory [13]. Moreover, the MAC protocol is responsible for the spectrum sensing and spectrum access decisions [14]. The major objectives of cognitive MAC protocol design as follows.
250
the secondary users share the sensing results of each other. Each channel is divided into four intervals: Idle, Sensing – Sharing, Contention, and Transmission. The sensing-sharing and contention interval are further divided into number of slots [18]. In the FCS, each secondary user senses only fixed number of licensed channels regardless of the traffic load of licensed channels. A cognitive user is able to send a frame successfully in the transmission interval only if that cognitive user is not met with a collision in contention interval. This is possible only if the each transmitting cognitive user has chosen different contention slot in contention interval. This probability of collision is high if the number of contention slots is limited and secondary users who want to send data are large. Also, in SMC-MAC, it has been proposed that each secondary user randomly chooses a contention slot which makes it more vulnerable to collision among the cognitive users.
1) To optimize the spectrum sensing and spectrum access decision, 2) To control the multiuser access in the multichannel network, 3) To allocate the radio frequency spectrum and schedule traffic transmission, and 4) To support the spectrum trading function. For DSA-based cognitive radio networks, MAC protocols followed by traditional wireless networks need to be modified to add the sensing and adaptation functionalities. The design of MAC protocols of cognitive radios are very challenging task due to the requirement of the coexistence of unlicensed users with licensed users. However, such a protocol needs to achieve the highest spectrum utilization by detecting all the spectrum opportunities and access the spectrum so that the collision with the other unlicensed users has to be minimized. However, some of the cognitive radio MAC protocols have been discussed in [15-19]. The hardware constrained-MAC (HC-MAC) [15] protocol considers the hardware constraints and proposed efficient spectrum sensing and access decision. A distributed spectrum-agile MAC protocol has been proposed in [16], which is a multichannel carrier sense multiple access (CSMA) -based protocol equipped with a dynamic channel selection algorithm. This model can be used for single or multiple channels and single or multiple users. In [17], the cognitive MAC (C-MAC) protocol for the distributed cognitive radio network is proposed. However, other technical issues of the C-MAC are also discussed in detail in [17]. Further, a self-scheduling multi channel cognitive radio-MAC (SMC-MAC) protocol has been proposed in [18], in which cooperation among secondary users are incorporated to enhance the throughput of cognitive users by utilizing more idle channels for data transmission than the ones sensed by single cognitive user. In [18], the fixed channel sensing (FCS) and adaptive channel sensing (ACS) schemes are considered. In this paper, we have explored the self scheduled multichannel-MAC (SMC-MAC) protocol for the distributed cognitive radio network to improve the throughput of the communication system. In this communication network, the sensing results have been shared among the cognitive users. However, in these communication systems, each cognitive user can sense the limited number of licensed channels. The remainder of the paper is organized as follows. In Section II, the system model has been discussed in which the sensing of primary licensed channels and sharing of these results among all the cognitive users is explored. Also, the idle sensed licensed channels reservation to transmit the data and the throughput analysis has been performed. Further, in Section III simulation results of the analysis have been discussed. Finally, Section IV concludes the work and recommends the future directions. II.
A. Sensing Sharing Analysis In [20], the behavior of cellular communication system subscribers, which follows the Poisson distribution and the arrival time between two calls is exponentially distributed. Poisson process is a Markov process with state transitions limited to the next higher state or to the same state and having a constant transition rate. In [20], the licensed user network has been utilized by the cellular network and it is also assumed that the all the licensed channels have same utilization probability α. Also, the total number of secondary users is denoted by . Therefore, the probability , that number of idle licensed channels is , follows the binomial distribution is [18]: 1
, 0
(1)
is the total number of licensed channels and the where average number of licensed idle channels is: ∑
(2)
where is from (1). Let us assume that cognitive user can sense only channels randomly among total licensed channels. Here, the is fixed, therefore it is also called FCS. Then the probability distribution of the number of sensed idle channels among channels by a single cognitive user is [18]: 1
,0
(3)
Thus the average number of sensed idle channels by a cognitive user is: ∑
(4)
where is from (3). Then, the probability that an idle channel is sensed among the actual idle channels by a cognitive user is defined as: μ
SYSTEM MODEL
We have considered the distributed cognitive radio network and its MAC protocol as in SMC-MAC [18]. The working of SMC-MAC for FCS has been shown with the help of flowchart in Fig. 1. This protocol consists of a control channel on which
251
(5)
B. Contention Analysis After sensing the licensed channels and sharing these cognitive users during sensing sharing results among interval, cognitive users compete with each other for reserving the idle licensed channels during contention interval. Each secondary user, which has data to send to its intended receiver, randomly selects a contention slot among total number of contention slots in contention interval. Since the contention slot selection by each cognitive user is random, therefore it is possible that two or more cognitive users have been selected the same contention slot, which results in collision and collided secondary users cannot reserve idle licensed channels for data transmission during data transmission interval. However, the case in which a contention slot is selected by a single cognitive user, it results in successful contention slot and the data is transmitted over reserved idle licensed channel/channels during transmission interval by successful cognitive user. Since we have number of contention slots, therefore probability of selecting each . The number of cognitive users which contention slot is:
Fig. 1 Flow diagram of the self- scheduled MAC protocol for cognitive user network.
From (5), we can obtain the probability that an idle channel is not sensed by any number of cognitive users, which is given by: 1
μ
(6)
From (6), the probability that an idle channel is sensed by at least one cognitive user is: 1
(7)
The probability distribution of the number of sensed idle channels by cognitive users are determined using (2), and (7) as: 1
,0 (8)
From (8), the average number of sensed idle channels by cognitive users is calculated as: ∑ where
select a contention slot is denoted by random variable, which follows a binomial distribution:
(9)
1
is from (8).
,0
,
(10)
The probability of a contention slot being successful is determined from (10), when s = 1 that is when single cognitive user has selected a contention slot. Therefore, we get from (10), the probability of success given as:
252
1
A. Sensing- Sharing In this sub-section, the simulation results of contention analysis, which is discussed in Section II (A) have been presented. The total number of licensed channels are assumed to be =20. Since a cognitive user is able to sense only a fixed number of channels given by , therefore Fig. 2 shows that as the number of channels sensed by each cognitive user increases, the average number of sensed idle channels by number of cognitive users also increases. The results are presented from (9) and for the case when total number of cognitive users are 5 and 10. Also, the traffic load is taken 0.5. Further, Fig. 3 demonstrated the actual number of idle channels and number of idle channels sensed by 10 the cognitive users and its dependence on . It is clear from Fig. 3 that there is gap between the actual number of idle channels and number of sensed idle channel when Ch 2, which is due to the small number of channels sensed by individual cognitive user. It has been shown in [18] that as the cognitive user’s ability of sensing licensed channels increases, the total number of channels sensed by all cognitive users almost approaches to total number of idle channels.
1
1 1
(11)
Let be the random variable, which denotes the number of successful cognitive users and the probability of secondary users being successful is: 1
,0
(12)
The average number of successful secondary users is calculated from (12) and is defined as: ∑
(13)
From (12), the average number of collided secondary users is: ∑ where
(14)
is from (12).
C. Data Transmission and Throughput analysis After contention interval, the successful secondary users transmit their data during data transmission interval. However, is defined by subtracting the the data transmission interval , the sensing sharing time , and the from idle time [18]. This transmission interval is utilized the cycle time for computation of the throughput of cognitive users [18]. However, two types of throughputs has been discussed in [18], maximum throughput and real throughput. The maximum throughput is the throughput for the case when all the sensed secondary users. idle channels are utilized by total However, real throughput is the throughput of only successful secondary users of contention interval. Therefore maximum throughput is defined as the product of the amount of time / , available for data transmission per cycle interval average number of sensed idle channels and data rate per sensed idle channels . Hence, the maximum throughput is given as [18]: (15)
Fig. 2 Average number of the sensed idle channels by 10 cognitive users for varying number of
5 and .
where is from (9). However, in case of the real throughput the number of idle channels used for data transmission is the minimum of the × and the average number of sensed idle channels from (9). Therefore, the real throughput of cognitive users is given as [18]: × ,
(16)
where is the number of idle channels that a cognitive user is allowed to use and is the number of successful cognitive users during contention interval. Therefore, × defines the total number of idle channels that all successful cognitive users transmit data frame in. III.
SIMULATION RESULTS Fig. 3 Variation of the average number of sensed idle channels with traffic load for 10 .
The parameters for cognitive user network are employed from IEEE 802.11a [20]. However, the simulation results have been classified as follows:
253
cognitive user during the contention interval. The average throughput is less for 5 in comparison to 10, because the total number of sensed idle channels in sensing sharing interval are less for 5 than 10. Therefore, all the idle channels among 20 are not utilized for transmission and hence less is the average throughput for 5 compare to 10.
B. Contention Results There are some limitations of the SMC-MAC [18], which is discussed through the simulation results in this part. Fig. 4 has presented the results from [18] and (13) when total number of contention slots are 1000. Fig. 5 has plotted the average collided cognitive users from (14) and it reveals that the number of collisions is very high even when the number of contention slots is comparable to the total number of competing cognitive users. For example, at 30 and 30, the value of from (14) comes out to be 14.2426, that is the average number of collided secondary users are 14 and rest of only 16 cognitive users are successful in contention interval. This signifies that only half of the contention slots are successful. Also, for low values of contention slots number of collisions among cognitive users is very high. However, the large number of contention slots increases the successful cognitive user but decreases the data transmission interval, since the total cycle time is fixed. This is one of the major drawbacks of SMC-MAC protocol as discussed in [18]. Also, in SMC-MAC it has not been possible that collided cognitive users in contention interval can once again select a contention slot in that cycle time so that there transmission can be successful. This is because all secondary users have selected the contention slot randomly and once collision is detected in chosen contention slot, the collided cognitive users wait for next cycle time for transmission. However, in case binary exponential back-off mechanism is applied to resolve contention among collided cognitive users, more number of the users can be successful.
Fig. 5 Average number of the collided cognitive users for different number of contention slots
Further, Fig. 7 shows the average real throughput, which takes into account the cognitive user’s collision during the contention interval and then evaluate the throughput, called real throughput. In this, 10 and 1 . From Fig. 7, it is depicted that average real throughput is maximum when the number of contention slots is equal to the number of cognitive users that is at 10 because after 10 cognitive users, however, the number of sensed idle channels increases but the competition among cognitive users to occupy the contention slots and thus collision increases, which has reduced the average real throughput.
C. Data Transmission and Throughput However, the data transmission and throughput analysis presented in Section II(C) has been simulated and discussed Fig. 6 demonstrated the average throughput variation with traffic load for 5 and 10 numbers of cognitive users with 2, 1 and 10.
Fig. 4 Average number of the successful cognitive users for different number of contention slots.
Fig. 6 Average throughput variation with traffic load, 2, 1 and 10.
The average throughput is the maximum achievable throughput when all sensed idle channels are successfully utilized and no cognitive user is having collision with other
10,
Therefore, it is necessary that when the number of cognitive users is large, contention slots must also be high. Since in the
254
wireless communication system number of transmitting cognitive users is randomly changing, therefore to have fixed number of contention slots is not practical. However, the number of contention slots must vary according to the number of cognitive users to enhance the performance of SMC-MAC.
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
Fig. 7 Average real throughput variation with 2, 1 and
IV.
at chosen value of 10.
[12]
CONCLUSION AND FUTURE DIRECTION
In this paper, the self-scheduled multi-channel - MAC protocol for the distributed cognitive radio communication system has been explored. However, the mathematical analysis of MAC for FCS with numerical results is presented. As in the proposed MAC protocol, the contention slots are fixed and this value should be chosen appropriately because the number of slots is large in number this will reduces the data transmission interval and if it is too small, it will results multiple collisions among cognitive users. As in the selfscheduled multi-channel - MAC protocol the modification can be made by incorporating the flexible contention interval whose time interval and contention slots will depend on the number of cognitive users in the cognitive network. Also, by using the binary exponential back-off mechanism instead of randomly selecting contention slots during contention interval will improve the throughput of cognitive users, which will be reported in future communication.
[13]
[14]
[15]
[16]
[17]
[18]
[19]
REFERENCES [1] [2]
[3]
NTIA, “U.S. frequency allocations,” [online] http://www.ntia.doc.gov/files/ntia/publications/2003-allochrt.pdf. E. Hoossain, D. Niyato, and Z. Han, “Dynamic Spectrum Access and Management in Cognitive Radio Networks,” Cambridge University Press: New York, 2009. Qing Zhao and B. M. Sadler “A survey of dynamic spectrum access: signal processing, networking, and regulatory policy”, IEEE Signal Processing Magazine, vol. 24, no. 3, pp. 79-89, May 2007.
[20]
[21]
255
C. Santivanez, R. Ramanathan, C. Partridge, R. Krishnan, M. Condell and S. Polit , “Opportunistic spectrum access: challenges, architecture, protocols,” Proc. of 2nd Annual International Workshop on Wireless Internet (WICON), New York, NY, USA, 2006. Joseph Mitola III “Cognitive radio: an integrated agent architecture for software defined radio,” Ph.D. Thesis, Royal Institute of Technology (KTH) Sweden, May, 2000. F. K. Jondral, “Software-defined radio: basics and evolution to cognitive radio,” EURASIP Journal on Wireless Communications and Networking, vol. 5, no. 3, pp. 275–283, 2005. C.-I. Badoi , N. Prasad, V. Croitoru and R. Prasad, “5G based on cognitive radio,” Int. J. of Wireless Pers. Commun., vol. 57, pp. 441464, 2011. R. Tandra, S. Mubaraq Mishra and A. Sahai, “What is a spectrum hole and what does it take to recognize one?,”Invited paper, Proc. of IEEE, vol. 97, no. 5, pp. 824-848, May 2009. S. Pandit and G. Singh, “Throughput maximization with reduced data loss rate in cognitive radio network,” Telecomm. Systems, (in press), Oct. 2012. Shweta Pandit, “Spectrum sharing in cognitive radio system for throughput optimization and interference reduction,” M. Tech. thesis, Jaypee University of Information Technology, May 2012. S. Pandit and G. Singh, “Adaptive transmit power and data rate variation channel state information and sensing information,” Proc. of UGC Sponsored National Seminar on Wireless Communication and Networks, India, 24-25 March 2012, pp. 53-56. S. Pandit and G. Singh, “Channel capacity comparison in fading environment based on sensing and channel state information for cognitive radio,” Journal of Signal Processing Systems, (under review), Feb. 2013. S. Pandit and G. Singh, “Spectrum sharing in cognitive radio using game theory,” Proc. of IEEE International Advance Computing Conference (IACC-2013), India, Feb. 22-23, 2013, pp. 1505-1508. Shweta Pandit and G. Singh, “Spectrum sharing in cognitive radio- a technical review,” Int. J. of Wireless Pers. Commun. (under review), Feb. 2013. J. Jia, Q. Zhang and X. Shen, “HC-MAC: a hardware-constrained cognitive MAC for efficient spectrum management,” IEEE Journal on Selected Areas in Communications, vol. 26, no. 1, pp. 106–117, 2008. A. Motamedi and A. Bahai, “MAC protocol design for spectrum-agile wireless networks: stochastic control approach,” Proc. of 2nd IEEE International Symposium on New Frontiers in Dynamic Spectrum Access Networks (DySPAN), Dublin, 17-20 April 2007, pp. 448–451. C. Cordeiro and K. Challapali, “C-MAC: a cognitive MAC protocol for multi-channel wireless networks,” Proc. of 2nd IEEE International Symposium on New Frontiers in Dynamic Spectrum Access Networks (DySPAN), Dublin, 17-20 April 2007, pp. 147–157. S. Lim and T.- J. Lee, “A self-scheduling multi-channel cognitive radio MAC protocol based on cooperative communications,” IEICE Trans. on Communications, vol. E94-B, no.6, pp.1657-1668, June 2011. H. Su and X. Zhang, “CREAM-MAC: an efficient cognitive radioenabled multi-channel MAC protocol for wireless networks,” IEEE Proc. of International Symposium on World of Wireless, Mobile and Multimedia Networks (WoWMoM), June 2008, pp. 1–8. D. Willkomm, S. Machiraju, J. Bolot, and A. Wolisz, “Primary user behaviour in cellular networks and implications for dynamic spectrum access,” IEEE Commun. Mag., vol. 47, no. 3, pp. 88-95, March 2009. IEEE 802.11, Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications, IEEE, June 2007.
