Abstract—Cognitive Femtocell has been envisioned as a highly promising solution for spectrum scarcity problems and local-convergence demands for indoor network applications. However, the achievable gains from Cognitive Femtocell deployment is limited by the hidden node problem leading to increased interference and consequently reduced spectrum efficiency. In this paper we compare the degree of performance enhancement when a cooperative mode is used to share information about the channel versus when standalone decisions are taken about the state of the channel. Simulation results indicate superior performance of the cooperative mode in terms of the number of observed hidden nodes and probabilities of detection and false alarm.
Keywords—Femtocell; Cognitive Radio; Dynamic Spectrum Access; Energy Detection; Hidden Node Problem; Spectrum Sensing; IEEE 802.22.
I. INTRODUCTION
Recent surveys have shown that indoor traffic accounts for more than 50% of the total data and voice traffic carried in wireless networks with most of the traffic originating from homes, airports, and schools [1] .This skyrocketing growth of indoor traffic coupled with the increased user demand for high broad band services necessitates new solutions that will provide the required indoor coverage and bandwidth. The existing macrocells are not efficient at delivering indoor coverage due to the high penetration losses in walls during signal propagation which leads to low coverage regions (black spots) [2].The promising solution towards satisfying the user requirements lies in the deployment of the low-cost low-power Femtocell Access points (FAP) which will enhance indoor coverage, deliver the required high bandwidth and also offload traffic from the macrocell networks [3]. Due to the proximity of indoor mobile users and the femtocell Access Point (FAP), the indoor signal to interference and noise ratio (SINR) can be kept sufficiently high [3]. However, indoor measurements have shown that significant power leakage from the femtocell access point (FAP) exists outdoor, depending on the nature of the building and location of the FAP within the building [4, 5]. This can cause significant interference to outdoor macrocell users with weak reception considering the fact that the FAPs use same frequency band as the existing macrocells. Cognitive Radios (CR) which are able to make smart decisions based on channel status have been proposed as a promising technology to mitigate the interference problem and also improve the spectrum efficiency in Femtocell deployment scenario. However, the performance gains achievable from the use of CR is still limited by the hidden node problem in which even nodes in close proximity will not hear each other [6]. Many standards such as WiFi (IEE 802.11), Zigbee (IEE 802.15.4) and WiMAX (IEEE 802.16) already include some degree of CR technology today but IEEE 802.22 WRAN is the first standard developed using CR technology that operates on TV white spaces and focuses on constructing fixed point-to-multipoint WRAN that will utilize VHF/UHF TV bands between 54 MHz and 862 MHz [7]. IEEE 802.22 WRAN systems share the geographically unused TV spectrum on non-interfering basis in rural environment where it is difficult to provide broadband access. IEEE 802.22 is developed to utilize unused TV bands without providing harmful interference to incumbent users [8]. Two major challenges that are faced by IEEE 802.22 are the issue of self co-existence and the hidden incumbent problem.
In a system like IEEE 802.22 where unlicensed devices are sharing the spectrum under the presence of licensed incumbents, the issue of self co-existence among multiple IEEE 802.22 operators in an overlapping region is very significant. When multiple unlicensed operators are operating using a small available band of frequency, there is a chance that the operators will try to act greedy and occupy the available bandwidth [7]. As all the operators will act in the same way, this may result in interference among IEEE 802.22 networks themselves. Thus an efficient channel allocation method needs to be invoked in order to use the channels with least interference.
The hidden node scenario is illustrated in the Figure 1 comprising of three communicating nodes. In this case, Node A can hear Node B, but not Node C, as it is not in the same transmission range. Therefore, as can be seen from the Figure, Node B can receive packets from both Nodes A and C. However, there will be a collision at Node B if these two nodes send their packets at the same time, and Node B cannot successfully receive any packets. [9] Figure 1: Hidden node Problem [10] In this situation, blind nodes cannot receive any control packets, so packets are sent to the visible node regardless of any other nodes sending packets, which leads to collisions and packet loss [9]. Let us assume that a BS and a Consumer Premise Equipment (CPE) are communicating using a specific frequency channel and an incumbent returns to the same frequency channel near the CPE but outside the BS sensing region. The CPE can detect the incumbent transmission in-band, but the BS cannot. The BS will continue transmission and might interfere with the incumbent. The CPE cannot report this licensed incumbent as its transmission will cause interference to the incumbent. On the other hand, due to the centralized nature of the IEEE 802.22 network (on-air activities of CPE is controlled by BS), the CPE cannot choose any other channel to connect to the BS as it is not permitted to use any other channel unless BS provides the permission. The problem gets worse as the CPEs do not have any reporting period. Instead what they do have is a channel move time (2 seconds) which means that if they sense any incumbent present in the same frequency band they have to move within the stipulated channel move time [7].
Cooperation between the FAP and femtocell users has been suggested as one way of improving spectrum sensing performance [11]. Cooperation solves the hidden node problem by improving performance in multipath fading and shadowing channels by providing spatiotemporal diversity [6,11]. Spectrum sensing is an important component of cognitive radio and the IEEE 802.22 standard [12].
In this paper, we compare the performance of the cooperative scheme and a non-cooperative scheme in which there is no coordination between the different cognitive femtocell access points.
In the literature, several approaches have been adopted to mitigate the hidden node problem in wireless networks for example use of the floor acquisition multiple access (FAMA) protocol in [13]. Other approaches adopt busy tone control packets sent by sender/receiver to the neighbors to reserve the channel during transmission [11, 14]. In [15-17], the different variants of handshaking mechanisms involving exchange of request to send and clear to send signals to eliminate the hidden node problem are proposed. It should be noted that none of these works addresses the hidden node challenge in cognitive femtocell networks. The rest of this paper is organized as follows: Section II presents the system model adopted by the paper, while in section III, our proposed approach is discussed in detail. The performance of our proposed algorithm is evaluated in section IV through computer simulations and the paper is concluded in section V.

II. SYTEM MODEL
We consider a sample of 50 cognitive radio users uniformly distributed in an area of about one square kilometer in an obstacle free environment. As such, the only loss that the devices face will be the path loss. Among the 50 CR devices, eight of them are the femtocell access points (FAPs) to control the operation of the femto user’s equipment within their range. We consider the licensed users as the primary users and the other users intending to use the poorly utilized spectrum as secondary users.
We model the received signal at the energy detector of each Cognitive user as [18]: (1)
Where s(n) is the signal to be detected, w(n)is the additive white Gaussian noise (AWGN) sample, and n is the sample index. Note that s(n) =0 when there is no transmission by primary user. The decision metric for the energy detector can be written as (2) where N is the size of the observation vector. The decision on the occupancy of a band can be obtained by comparing the decision metric M against a fixed threshold λE. This is equivalent to distinguishing between the following two hypotheses: (3) where ℋ1 is the alternate hypothesis, i.e. the hypothesis that the observed band is occupied by a primary user. The performance of the detection algorithm can be summarized with two probabilities: probability of detection PD and probability of false alarm PF. PD is the probability of detecting a signal on the considered frequency when it is truly present. Thus, a large detection probability is desired. It can be formulated as [18]:

(4)

PF is the probability that the test incorrectly decides that the considered frequency is occupied when it is actually not, and it can be written as

(5)

The white noise is modeled as a zero-mean Gaussian random variable with variance , i.e. w(n) =N(0, ). For a simplified analysis, we model the signal term as a zero-mean Gaussian variable as well, i.e. s(n)= . The model for s(n) is more complicated as fading should also be considered. Because of these assumptions, the decision metric (2) follows chi-square distribution with 2N degrees of freedom and hence, it can be modeled as (6)
For energy detector, the probabilities PF and PD can be calculated as (7) (8)
Where λE is the decision threshold, and Γ(a, x)is the incomplete gamma function .

III. ALGORITHM TO REDUCE THE HIDDEN NODE
Let the licensed users be considered as primary users and the other users intending to use the spectrum as secondary users. We require that at each instant, the user performs spectrum sensing using energy detection technique to avoid interference to the licensed users. In our cooperative approach, there is communication between the different CR users and the Femtocell Access Point (FAP). Each CR user locally performs spectrum sensing and forwards the results to the FAP. The sensing results received at the FAP are then forwarded to the Fusion Centre (FC) for a final decision. The final decision as to whether the channel is occupied is determined using the OR fusion rule at the Fusion center. Each CR user makes its own decision regarding the presence of the PU, and forwards the binary decision (1 or 0) to fusion center (FC) for data fusion. If we let N to denote the number of users sensing the spectrum, and assuming independent decisions, the fusion problem where k out of N CR users are needed for decision can be described by binomial distribution based on Bernoulli trials where each trial represents the decision process of each CR user. The generalized formula for overall probability of detection, Qd for the k out of N rule is adopted from [19]: (9) where Pd is the probability of detection for each individual CR user.
The OR fusion rule (i.e 1 out of N rule) can evaluated by setting k = 1 in the above Equation (10)
For the case of a non-cooperative approach, each individual user carries out a local spectrum sensing and makes individual binary decision regarding the occupancy of the channel depending on the average energy detected.
IV. SIMULATION RESULTS
We compared the performance of the cooperative spectrum sensing approach with the non-cooperative. Figure 2: Receiver operating characteristics for non-cooperative simulations Figure 3: Receiver operating characteristics curves for cooperative sensing From Figure 2 and 3, it is observed that the probability of false alarm reduces significantly with cooperative sensing and thus the number of hidden nods reduces. This is attributed to the fact that when spectrum sensing is performed using a single node (non-cooperative), that sensor may be in a deep fade and because of this possibility, a secondary unit basing its decisions on single node sensing may not engage in a secondary transmission unless it is highly confident in its detection of a spectrum opportunity, i.e., it must be able to detect a transmitter even as it experiences deep fading. To this end, the sensing node must use conservative detection thresholds and/or highly sensitive receivers, which cause high false alarm probability (the probability of reaching a “detect” decision when there was nothing there) and high cost devices, respectively [20]. The status of the cognitive femto nodes after non cooperative and cooperative spectrum sensing are shown in Figure 4 and Figure 5 respectively from which hidden nodes can easily be identified. Figure 4: Nodes hidden in non-cooperative spectrum sensing Figure 5: Nodes hidden in Cooperative spectrum sensing
As can be seen in Figure 4 for the non cooperative case, we can see seven nodes (nodes 20, 24, 25, 36, 42, 49 and 50) that are visible but not available for communication which are the hidden nodes .On the other hand, only two nodes (nodes 36 and 42) are hidden in the case of cooperative spectrum sensing as seen from Figure 5.
A. Discussion of results
Cooperation is the solution to problems that arise in spectrum sensing due to noise uncertainty, fading, and shadowing. Cooperative sensing decreases the probabilities of misdetection and false alarm considerably. In addition, cooperation solve hidden primary user problem and it can decrease sensing time.
Collaborative spectrum sensing is most effective when collaborating cognitive radios observe independent fading or shadowing. It is found that it is more advantageous to have the same amount of users collaborating over a large area than over a small area. It is shown that cooperating with all users in the network does not necessarily achieve the optimum performance and cognitive users with highest primary user’s signal to noise ratio are chosen for collaboration. Constant detection rate and constant false alarm rate are used for optimally selecting the users for collaborative sensing [21].

B. Varying the system parameters
In doubling the number of nodes in non-cooperative spectrum sensing, an increase in the number of hidden nodes was registered and we were able to observe up to about 29 nodes which are hidden. Figure 6: Hidden nodes in Non-Cooperative spectrum sensing on the increase of the number of nodes Figure 7: Hidden nodes in Cooperative spectrum sensing on increase of the number of nodes When we increased the number of the nodes, an increase in the number of recovered hidden nodes was observed in cooperative sensing. The two hidden nodes that were observed when the number of cognitive femto users was 50 have been recovered as shown in Figure 7. We notice an increase in the number of hidden nodes in the non-cooperative spectrum sensing as the number of nodes is increased. This means that increasing the number of nodes improves spectrum sensing performance in the case of cooperative sensing thus leading to reduced number of hidden nodes as the number of cooperating nodes increases. On the contrary, the number of hidden nodes increases with increase in the number of communicating nodes in the case of non-cooperative spectrum sensing. This can be attributed to the reduction in energy used in spectrum sensing process. Also when we halved the sensing radius, about 28 nodes out of 50 nodes were hidden nodes in non-cooperative spectrum method whereas for cooperative about 8 nodes out of 50 nodes were hidden. This means that reducing the sensing radius reduces the efficiency of spectrum sensing in both cooperative and non-cooperative sensing leading to an increase in the number of hidden nodes.
C. Performance Analysis
The cognitive radios can solve the hidden node problem using cooperative spectrum sensing but the percentage of reduction varies according to the parameters used [22]. For all transmission ranges, cooperative spectrum sensing indeed guarantees that after T time units the percentage of hidden nodes will decrease. Reducing the transmission range means reducing the energy used in the spectrum sensing process. This reduces the sensing rate and reporting speed leading to a reduction in the sensing efficiency. This causes an increase in the number of hidden nodes [23].
D. Cooperative Spectrum sensing features
In cognitive radio networks, multiple cognitive radio nodes perform spectrum sensing cooperatively in order to detect the primary user more accurately. Previous works on cooperative spectrum sensing have shown that the detection performance can be improved through increasing either the observation interval or the number of the sensing nodes. However, increasing the observation interval will result in the reduction of transmit efficiency and the agility of cognitive users, and at the same time increasing the number of sensing nodes may lead to the overhead increase of control channel and computational complexity [24]. For 10,000 Monte Carlo simulations, the following results were obtained from the sensing time measurements;
TABLE I. SENSING TIME
CR users Observed Sensing Time in seconds Non-cooperative sensing cooperative Reduction
20 2275.00 2160.15 5%
50 2803.83 2708.57 3.4%
10 578.71 520.77 10%

This means that the higher the number of cognitive femto users, the lower the percentage reduction in sensing time but in all cases, cooperative spectrum sensing time is less than that for non-cooperative for the same number of nodes.
E. Comparing our sensing time previous work
From the simulation results that we obtained and comparison with the previous work, we conclude that cooperative sensing reduces sensing time on the cost of a reduction in the transmission efficiency. Due to the hardware limitation that a single RF transceiver equipped in each CR user cannot simultaneously perform sensing and transmissions, the more time is devoted to sensing, the less time is available for transmissions and thus reducing the CR user throughput. This is known as the sensing efficiency problem or the sensing-throughput trade off in spectrum sensing [25] .
CONCLUSION
In this paper, a performance comparison between cooperative and non-cooperative spectrum sensing has been made in terms of the observed number of hidden nodes and probabilities of detection and false alarm. Simulation results show superior performance of the cooperative approach over non–cooperative techniques in which each node takes local control of its channel state information. The main idea of cooperation is to improve the detection performance by taking advantage of the spatial diversity, in order to better protect a primary user, and reduce false alarm to utilize the idle spectrum more efficiently.
The most significant application of IEEE 802.22 is to provide rural and remote regions with wireless broadband access. The significance of such an application comes from the fact that about half of the population in many regions including Africa and Asia exists in rural and remote areas. It should be noted that the applicability of IEEE 802.22 is not restricted to such rural regions. As a matter of fact, other targets of IEEE 802.22 applications include single-family residential, multi-dwelling units, small businesses, multitenant buildings, public and private campuses, etc [26]. Data, voice, and video traffic with appropriate QoS are also examples of services that IEEE 802.22 supports. Cooperative spectrum sensing is an effective method to improve the detection performance by exploiting spatial diversity in the IEEE 802.22 networks (through) improvement of the performance in multipath fading and shadowing channels.

REFERENCES

[1] G. Mansfield, “Femtocells in the US Market –Business Drivers and Consumer Propositions,” FemtoCells Europe, ATT, London, U.K., June 2008. [2] D. Lopez-Perez, A. Valcarce, G. de la Roche, and J. Zhang, “OFDMAfemtocells : A roadmap on interference avoidance,” IEEE Communications Magazine, vol. 47, pp. 41–48, sep. 2009.
[3] J. Zhang, “Femtocells: Technologies and Deployment,” Wiley, 2010, pp. 2-130.
[4] D. Lopez-Perez, A. Ladanyi, A. Juttner and J. Zhang, “OFDMA femtocells: Intracell Handover for Interference and Handover Mitigation in Two-Tier Networks”, Wireless Communication and Networking Conference, 18-21 April 2010. [5] S. Yeoul Choi, T. Jin Lee, M. Young Chung, and H. Choo, “Adaptive Coverage Adjustment for Femtocell Management in a Residential Scenario”, 12th Asia-Pacific Network Operations and Management Symposium, APNOMS 2009, Jeju, South Korea, September 23-25, 2009.
[6] J. K. Sreedharan and V. Sharma, “Spectrum Sensing using Distributed Sequential Detection via Noisy MAC”, 23 Nov 2012. [7] S. Sengupta, S. Brahma and M. Chatterjee, “Enhancements to cognitive radio based IEEE 802.22 air-interface”, Communications, 2007.ICC ’07. IEEE international Conference, pp.5155-5160, 24-28 June 2007.
[8] M. Sherman, A. N. Mody, R. Martinez, and C. Rodriguez, “IEEE Standards Supporting Cognitive Radio and Networks, Dynamic Spectrum Access, and Coexistence.” Communications Magazine, IEEE Vol:46, pp. 72-79, July 2008.
[9] L. Boroumand, R. H. Khokhar, L. A. Bakhtiar and M. Pourvahab, “A Review of Techniques toResolve the Hidden NodeProblem in WirelessNetworks”, Smart Computing Review, vol.2, no. 2, April 2012.
[10] Hidden node problem, http:// www.soi.wide.ad.j p/class /20060035/ slides/03/28.html accessed on 19th June, 2013.
[11] T. S. Dhope, S. Patil,V. Rajeshwarkar, D. Simunic, “Performance Analysis of Hard Combing Schemes in Cooperative Spectrum Sensing for Cognitive Radio Networks”, International Journal of Engineering and Advanced Technology (IJEAT) ISSN: 2249 – 8958, Volume-2, Issue-3 February 2013.
[12] Dr.-Ing. M. Kalil, “Cognitive radio The IEEE 802.22 standard”, Integrated Communication Systems Group Ilmenau University of Technology, 15 Dec 2011.
[13] C. L. Fullmer and J.J. Garcia-Luna-Aceves, “Solutions to Hidden Terminal Problems in Wireless Networks”, In proceedings ACM SIGCOMM Conference, pp.39-49, 1997.
[14] R.Severino, A. Koubaa, M. Alves and E. Tovar, “Improving Quality-of-Service in wireless Sensor Networks by Mitgitaing Hidden-Node Collisons”. IEEE Transactions on Industrial Informatics.Vol.5,pp 299-313,August 2009.
[15] P. Karn. “Macaa new channel access method for packet radio,” in Proc. of Computer Networking Conference on ARRL/CRRL Amateur Radio, pages 134140, 1990.
[16] K. Liu, S. Leng, H. Fu, L. Li, “A novel dual busy tone aided mac protocol for multi-hop wireless networks,” in Proc. of IEEE International Conference on Dependable, Autonomic and Secure Computing (DASC ’09), pp. 373-378, Dec. 2009.
[17] C.-K. Toh, V. Vassiliou, G. Guichal,and C.-H. Shih, “March: a medium access control protocol for multihop wireless ad hoc networks,” in Proc. of 21st Century Military Communications Conference, pp. 512-516, 2000.
[18] L. Perera and H. Herath, “Review of spectrum sensing in cognitive radio," in Proc. IEEE 6th International Conference on Industrial and In-formation Systems, Kandy, Sri Lanka, pp. 7-12, 16-19 Aug. 2011.
[19] M. A. hagos and M. Mohamed, “Inference from primary user on the performance of cognitive radio networks”, Blekinge Institute of Technology September, 2012.
[20] IST-2007-216248 E3 Project, http://www.ict-e3.eu/ accessed on 12th April, 2013.
[21] K. Elleithy and V. Rao, “International Journal of Next-Generation Networks, Femto Cells: Current Status and Future Directions (IJNGN)” Vol.3, March 2011.
[22] Y. B. Reddy, "Spectrum Detection in Cognitive Networks by Minimizing Hidden Terminal Problem", Ninth International Conference on Information Technology , pp.77-82, 2012.
[23] Z. Tan, Y. Liu, Z. Zhang, “Performance Requirement on Energy Efficiency in WSNs”,3rd International IEEE Conference on Computer Research and Development (ICCRD), Vol. 3, pp. 159-162, March 2011.
[24] Z. Yong-hui ,Z. Qiu, D. Mu and K. Xiong, “Performance Tradeoff for Cooperative Spectrum Sensing in Cognitive Radio Networks”, Wireless Communications, Networking and Mobile Computing, 2009, WiCom '09 5th International Conference on, 24-26 Sept. 2009.
[25] I. F. Akyildiz, B. F. Lo, R. Balakrishnan, “Cooperative spectrum sensing in cognitive radio networks: A survey”, Broadband Wireless Networking Laboratory, School of Electrical and Computer Engineering, Georgia Institute of Technology, Physical Communication, pp.40-62, April 2011.
[26] R. Al-Zubi, M.Z. Siam, M. Krunz, “Coexistence Problem in IEEE 802.22 Wireless Regional AreanNetworks”, Global Telecommunications Conference, pp. 1-6, Nov. 30 – Dec. 4 2009.