Maximum Channel Throughput via Cooperative Spectrum Sensing in Cognitive Radio Networks

Maximum Channel Throughput via Cooperative Spectrum Sensing in Cognitive Radio Networks


Chintan I. Patel

1. Introduction

Cognitive Radio is a form of wireless communications which consists of transceivers that can detect the presence of communication channels, either busy or idle. The vacant channels are hence occupied without causing any or minimum interference to the occupied or busy channels.

Since the radio spectrum is considered to be scarce, techniques are required in order to improve the capacity of the spectrum by maximizing channel allocation. Spectrum sensing is one such technique, by which the secondary (unlicensed) users need to find the idle channels for transmission in the cognitive radio networks. This can be done in two ways, i.e., the secondary users can sense the channels individually, one at a time, and/or by sensing multiple channels simultaneously.

Spectrum sensing is considered a fundamental problem in cognitive radio networks but the secondary users can make efficient and effective use of this technology to detect primary signals from the primary users. Other effects such as low signal to noise (SNR) ratio and fading add to the problems involved in spectral sensing.

2. Issues related to Cooperative spectrum sensing (CSS) technique

  • Improvement in the channel capacity using optimal CSS in cognitive radio networks by increasing the channel throughput considering the major constraint, i.e., interference by the secondary users on the primary users.
  • Significant improvement in channel capacity under the following two scenarios:-

a)      By sensing individual channels at a time using algorithms developed for maximizing capacity.

b)      By sensing multiple channels simultaneously using convex-optimization problems that can be solved with efficiency and reliability.


  • Network issues of spectrum sensing in cognitive radio networks: - Network issues are considered by including the number of secondary users, detection delay and signal interference.

a)      Number of secondary users: - To have a high chance of finding the idle channels, higher number of secondary users are required and hence high secondary channel capacity is obtained due to the increase of the number of secondary users.

b)      Detection delay: - If a primary channel is found to be occupied in a spectrum channel, then the secondary signals sensing the primary channel will interfere with the primary system and cause interference till the primary channel is found to be occupied in the next sensing by secondary signals. To avoid interference, secondary users must sense the channels periodically for a fixed time span. According to researches, with the same sensing performance, more secondary users involved leads to less sensing time/delay. Therefore, CSS will decrease the sensing delay, and subsequently reduce the interference to the primary system [1].

c)      Signal interference: - The transmitted signal and the interference signal of a primary user are considered as primary signals. For example, if primary user A and primary user B transmitted signals with power X1and X2 respectively, the signal of primary user B is considered as interference to primary user A. However, the secondary system regards the primary signal power in this band as X1+X2 [1]. This takes place because energy detection is done on the basis of signal power not the characteristics of the primary signal.


  • Analysis of optimal CSS settings for single-band channel sensing and wide-band sensing.
  • Another issue related to the sensing in cognitive radio networks is the sensing objective [1]. The main objective is to improve the performance and maximize throughput. Two parameters are involved to prove the objective, namely, probability of detection and probability of false alarm.


3. Core ideas to maximize channel throughput


3.1 Major types of CSS

  • The counting rule: - This includes AND-rule-based CSS, OR-rule-based CSS and MAJORITY rule [1]. In AND-rule, primary channel is sensed to be occupied only if the secondary users claim so [1]. In OR-rule, primary channel is sensed to be occupied if one or more secondary users claim so [1]. In MAJORITY rule, primary channel is sensed to be occupied if more than half of secondary users claim so [1].
  • Linear combination methods: - These are optimal methods for tractable mathematical analysis [1].
  • Linear-Quadratic (LQ) strategy: - This strategy is proposed to combat the correlation between secondary users [1].
  • Spatial diversity in multiuser networks: - Relay-based CSS are analyzed using this method [1]


3.2 Contributions of the counting rule

Counting rule as a fusion rule (AND, OR and MAJORITY) is used at the network center as the counting rule only sends the decisions of the secondary users to the network center and the network center takes the final decision.

  • There are less data transmissions between the center and the secondary users as the counting rule only sends the decision. Relay based CSS is considered to be unnecessary as in this the secondary users present near the primary users relay the primary signals to other secondary users far from the primary user. This increases the network range and is hence not considered as the cognitive radio networks are generally small compared with its distance from primary users [1].
  • Since in wideband sensing, multiple channels are sensed simultaneously, the overall system optimization level is considered thereby giving credit to the time efficiency of the system. Counting rule is applied to wideband sensing also.
  • Considering the system architecture of the counting rule, comparison with other sensing schemes can be done.





Let   be number of secondary users claiming the existence of primary users and K be the decision threshold at the network center where K= 0, 1, M-1 and ‘M’ are the total number of secondary users.

                        H1 are the number of primary users present and H0 are the number of primary users absent. For AND rule, K=0; for OR-rule K= M-1.

b)      Counting rule is considered compared to the weighted sum as it is assumed that secondary users receiving primary signals are independent and identically distributed (i.i.d) random variables due to which every secondary user receiving these primary signals are given equal weight and hence weighted sum is not considered.

c)       In linear combination, the measurements are sent by the secondary users to the network center. By sending different measurements at the network center would increase the number of bits (b) due to which number of secondary users which send sensing information to the center (U) would reduce. This is because, U=W/b where W is the maximum throughput between secondary users and the center. In counting rule, only decisions are sent, hence just one bit is sent, i.e., b=1, due to which maximum throughput is achieved. Due to these reasons, the counting rule is considered compared to linear combination.


3.3 Single channel sensing

Maximum channel throughput J is given by,


Let  denote probability of channel occupied, Tp and T 'p are throughput of the primary system without and with the existence of the secondary user respectively,  Qd is final detection probability. Hence, the product  indicates throughput generated when the probability of channel occupied by primary users without the presence of secondary users are taken into consideration when detected.

Let T 's and Ts be the throughput of the secondary system with and without the existence of the primary user respectively. Hence, the product

indicates throughput generated when probability of the channel occupied simultaneously by both primary and secondary users.

The product  indicates the throughput generated when the probability of the channel is occupied by only secondary users without the existence of the primary user.

Hence, the sum of all the products mentioned above gives the overall maximum channel throughput of the single band channel.


3.4 Wide-band sensing

In wide-band sensing, multiple channels are sensed simultaneously as seen earlier.



Assume that ‘L’ channels are sensed simultaneously and (1 <= j <= L), where j is channel between 1 and maximum number of channels ‘L’ (inclusive of 1 and L).

Let Qf be the false-alarm probability in the jth channel, Qd is the detection probability in the jth channel, P'f denote false-alarm probability at local detector in the jth channel. The level of interference at the primary user due to secondary devices is denoted by CT.

The total interference at the primary user is equal to CT(1 - Qd).

Let Jj be the throughput at jth  channel. The overall throughput of all primary channels is given by,.

Taking the above mentioned terms into consideration, the sensing problem can be formulated as [1],



The optimization of the above equation is a convex problem and can be solved efficiently and reliably by using the existing methods [1].


4. Simulation Results


4.1 Single channel sensing simulation discussion




The graph (Fig. 4) above illustrates results of channel throughput J versus P 'f with different K channels under different channels, i.e., AWGN, Rayleigh and Nakagami-m (m = 3) fading channels, in which sensing parameters are given as SNR = 8dB (average SNR), N= 8 and M= 4 [1].

For a given value of K, particular value exists. In the area before, J increases as P 'f  increases and in the area after, J decreases as the value for P 'f increases. For a smaller value of K, the channel fading effects are very harmful in the sensing performance. In other words, there are bigger differences between J in an AWGN channel and J in fading channels for larger K [1].


4.2 Wide-band sensing simulation discussion



The graph illustrated above indicates the overall channel throughput R versus under AWGN. Here, is the total interference of the system. The sensing parameters are taken as SNR = 8dB (average SNR), N= 8, M= 3 and K= 1 [1]. From the graph (Fig.5) it is evident that optimal settings can achieve a much higher throughput as compared to the uniform thresholds.


5. Conclusion

The optimal CSS settings for both the single channel sensing and wide-band sensing are analyzed. Theoretical analysis and simulation results show that the optimal settings can improve the channel capacity substantially [1]. Effects such as interference and fading can be controlled to a great extent using the counting rule, a major type of CSS.


6. Potential application of CSS

Cooperative spectrum sensing technique can find its useful use for military applications, as maximization of the spectrum can give more room for a number of devices to work together.


7. Future opportunities in deploying CSS

Due to the scarcity of radio spectrum, the need for CSS technology can reach to great heights. Since, the technology is increasing at a high rate, the number of devices using the wireless technology are also increasing at an equally fast pace. Deployment of CSS would enable a number of wireless devices to use the radio spectrum as the capacity of the channels would increase.


8. Summary

  • Various issues related to the Cooperative spectrum sensing techniques such as improvement in channel capacity in presence of interference, in single channel band sensing and wide-band sensing were discussed. Network issues of spectrum sensing in cognitive radio networks such as the number of secondary users, detection delay and signal interference.
  • Core ideas to improve the channel throughput consisted of major types of CSS, namely, the counting rule as fusion rule (AND, OR and MAJORITY rules), Linear combination methods, Linear-Quadratic (LQ) strategy and Spatial diversity in multiuser networks. The counting rule was discussed in detail, i.e., its contributions and throughput efficiency provided by using it.
  • The overall throughput efficiency of the two scenarios, namely, single band sensing and wide-band sensing were studied.
  • Simulation results and graph discussions for each of the scenarios were discussed.



9. Reference

[1]Maximum Channel Throughput via Cooperative Spectrum Sensing in Cognitive Radio Networks”, IEEE Transactions on Wireless Communications, Vol. 8, No. 10, OCTOBER 2009, Junyang Shen, Tao Jiang, Siyang Liu, and Zhongshan Zhang