Evaluation of the Failed Handoff Rate in Collaborative Wireless Networks

Cognitive radio is an essential technology to overcome fixed spectrum allocation problems and improve its use through dynamic spectrum access techniques. Spectral decision-making is a function of the cognitive cycle that aims to choose the best spectral alternative among a set of finite possibilities. An inadequate process of decision-making can significantly affect the network parameters. Therefore, it is important to determine the best strategy and if this decision making should be collaborative. The present research analyzes the performance of failed handoff rate according to the amount of information shared between secondary users. The aforementioned information corresponds to the spectral occupancy for the GSM frequency band, that is characterized by a set of users and five levels of collaboration, 10%, 20%, 50%, 80% and 100%, where each level represents the percentage of users are part of the training model implemented. The decision-making process is carried out by using two multi-criteria techniques: Feedback Fuzzy Analytical Hierarchical Process (FFAHP) and Simple Additive Weighting (SAW). Based on the obtained percentage ratios and the information shared, it is determined that for the given number of failed handoffs, the level of collaboration that leads to efficient results is between 20% and 50%.


Introduction
Cognitive radio (CR) is a smart technology that can efficiently solve limited access issues in wireless networks. The concept was created by Joseph Mitola III in 1999 and several entities such as the IEEE, the International Telecommunications Unit (ITU) and the National Telecommunications and Information Administration (NTIA) have come to terms with a general definition [1][2][3][4][5][6].
The main principle of CR consists on granting access to the spectrum through a dynamic strategy, through the opportunistic exploration in the space-time dimensions of the network.
In contrast with traditional networks, there are two types of users in CR. The licensed or primary user (PU) has a license to use frequency bands and the secondary user (SU) that makes opportunistic use of the spectrum while it is still available [6][7][8][9][10].
Cognitive radio networks (CRN) operate with a management model that performs smart adaptations, based on progressive learning and information exchange [11] to implement a dynamic and opportunistic access method. Spectrum decision-making is a key function of CRN. However, it has not been as researched in comparison to other functions of the cognitive cycle [7,[12][13][14]. The decision-making process aims to choose the best spectral opportunity among a finite set of possibilities, allowing the SU to generate a sequence of actions that will lead to the completion of its objectives [13][14][15][16]. An inadequate decisionmaking process can significantly affect the parameters of quality, latency, throughput, reliability, signalization, PU interference, energy efficiency, bandwidth, SINR and error rate [2,[17][18][19][20]. Therefore, a proper decision making process it is required.
Collaborative strategies have delivered new models for the efficient use of radio resources, as well as the decision-making process for CRN. In collaborative decision-making, users can communicate between them to exchange local measurements regarding availability and interference, as well as other variables [14]. The purpose is to take advantage of spatial diversity which is achieved by the collaborative of unlicensed users who share information with neighbors [21]. The collaborative approach offers more advantages than the non-collaborative approach because the diversification of the observation of the spectrum avoids possible interference. A challenge in spectrum selection is the way in which the information from collaborative users is combined while the transmission is carried out [22].
The decision making processes based on collaborative models allow the of use of different techniques, under probabilistic models. In [23], secondary users are classified into two levels: users in real time (RT) and users in non-real time (NRT). This enables to study the conflict and cooperation between both levels that make use of game theory. An auction game-based model is proposed that analyzes the decision-making process. The results show that a high number of secondary users is catered along with the reduction of user blockage probability, user handoff and channel saturation, as well as an increase in the probability of user acceptance.
In order to overcome the hidden terminal matter that a SU could face [24], an incentivebased framework is proposed that adopts a reputation strategy, where the problem of collaboration is modeled as an indirect reciprocity game. The results obtained with the strategy show that the proposed scheme achieves proper performance.
Another type of strategy uses techniques based on logical statements. In [25] a collaborative decision-making process is proposed that uses fuzzy logic. The decisionmaking criterion works with relative link quality (RLQ) and signal plus interference noise ratio (SINR). The model uses fuzzy logic rules based on the fuzzy interference system (FIS). Finally, due to low complexity and response times, the proposed strategy is characterized by being easily implemented in real systems.
Algorithms based on multi-criteria decision-making (MCDM) are widely used for the decision making processes based on collaborative models. They operate using weight allocation, which are adjusted according to certain requirements. [26] compares the performance of four algorithms for spectral occupancy of the PU during the communication with a SU. The performance of the algorithms is assessed using five metrics: handoffs, failed handoffs, bandwidth, delay and throughput.
The decision-making process requires the work of spectral occupation data that can be generated randomly or obtained through measurements. The collected data is used to train and validate the models. However, using few data can lead to inefficient decisions and gathering many data can incur in significant computational costs and delays.
To obtain an efficient decision-making process, with no computational costs. The research work presented analyze the decision making under collaborative scenarios in order to obtain on where users share different amounts of spectral occupancy information.
According to the results obtained in [26], metrics such as number of failed handoffs are used in order to measure performance collaborative decision-making process Information is characterized by five levels of collaboration (10%, 20%, 50%, 80% and 100%) where each level represents the percentage of information that is shared for training and subsequent validation of the model. The levels are chosen according to the limits of the data where 10% and 100% correspond to the criteria of few data and excessive data respectively. 50% corresponds to the intermediate value of the limits, and, the levels of 20% and 80%, comprises the interval to analyze the collaboration within an information range.
For the analysis of the decision making proposed, a collaborative algorithm is designed to segment the matrix according to five levels of collaboration. The decision-making process is based on two multi-criteria (MCDM) techniques: Feedback Fuzzy Analytical Hierarchical Process (FFAHP) and Simple Additive Weighting (SAW) [2,15,16,[27][28][29].
As part of the decision-making process, the connection between the collaborative module and the multi-criteria techniques is carried out with a search algorithm. The spectral occupancy data studied correspond to a GSM frequency band, obtained from previous measurements organized into power matrixes.
The research work is organized and presented into three section. The section 2 describes the methodology used to establish the input variables, the description and the functions of the collaborative model, the decision-making models and the performance metrics. The section 3 presents the results obtained and their respective comparative analysis, the multicriteria techniques and the collaborative structure. Finally, the last section presents a set of conclusions on the overall work.

Methods
In order to perform a decision-making analysis in a collaborative environment, the methodology is structured in four stages. The first stage consists on establishing the information that feeds the general model which includes the input variables based on real measurements for the GSM frequency band (uplink). The second stage defines the collaborative model. The block diagrams presented in this section identifies the inputoutput relationship. A general example is discussed on how the proposed implementation works. The third stage describes the mathematical foundation behind multi-criteria models.
In the fourth stage, the performance metrics for model quantification are analyzed.

Input variables
Power measurements in the GSM frequency band are used as input variables to carry out the analysis. The corresponding amount of information is shown in Table 1. The rows represent time in seconds and the columns represent the frequency channels. The sampling time for data capture was 290 ms.
The measurements require pre-processing to characterize the traffic level as High (few spectral opportunities) and Low (many spectral opportunities). Additionally, the type of model to be implemented makes it necessary to identify a data group that can train the models and another data group for validation. The matrix chosen for training allows setting the initial parameters of the algorithms and is then used for collaborative analysis. The assessment matrix is used to obtain the results of the assessment metrics of the implemented algorithms.
Based on the previous statements, two databases are used. One is used for training while the other one is used for assessment. Each database is classified into two traffic levels: High and Low. In total, there are 551 frequency channels, for 1 hour of training and 10 minutes for assessment. The size and classification are described in Table 2.

Collaborative model
The structure of the collaborative model consists on sectioning the training matrix according to the established number of users. The following sections include a detailed description of the strategy.

Operation of the collaborative model
The input and output diagram of the collaborative model is presented in Fig. 1. It has three input parameters (the specific description of each module is shown in Fig. 8

Fig. 3 Availability matrix
The availability matrix must be adjusted according to the user characteristics. The adjustments carried out for each scenario are presented in Table 3 and Table 4. For both cases, the availability matrix is sectioned into six users which is equivalent to six submatrices. Like the number of users, the User percentage variable is a randomly selected value, 50% for case 1 and 67% for case 2, this parameter indicates the percentage of users that will be part of the training, equivalent to a level of 50% collaboration (case 1) and 67% (case 2), therefore, if six is the total number of users (100%), 50% corresponds to training based on the information of three users, and the 67% corresponds to a training based on the information of four users. This percentage is an adjustable parameter that varies between 10% and 100% depending on the case of study. The collaborative model is programmed to adapt itself to any input parameter. For instance, if the collaboration level is adjusted during a training process based on the information of a non-integer number of users, then the model takes the closest integer. The variations in the model for the variables Division and Segmentation are described in the individual analysis of each study case. For more information on all the input variables, see Table 5.  Table 3, the case 1 sections the availability matrix into six users with row division. Since the number of users is less than ten and the segmentation is per row, the model takes the availability matrix and divides it into two columns (for users greater than ten see Table 5), then, it establishes in how many rows it must divide the availability matrix to be able to obtain a segmentation of six submatrices equivalent to six usurious, it must make a division into three rows. The Fig. 4 presents the availability matrix according to the number of users and division.
Only 50% of them are part of the training (variable that characterizes the collaboration levels), which corresponds to three users (User Simulation = 3) chosen randomly. The Fig.   5 illustrates the training matrix according to the adjustments described in Table 3.  Table 4, case 2 sections the availability matrix into 6 users with column division. Since the number of users is less than ten and the segmentation is per column, the model takes the availability matrix and divides it into two rows (for users greater than ten see Table 5), then, it establishes in how many columns it must divide the availability matrix to be able to obtain a segmentation of six submatrices equivalent to six users, it must make a division into three columns. The Fig. 6 shows the availability matrix according to the number of users and division.
Out of the 6 users ( Fig. 6) only 67% of them will be included in the training phase (variable that characterizes the collaboration levels), which corresponds to 4 users (User Simulation = 4) selected continuously. Fig. 7

Fig. 7
Training for four secondary users, with 67% collabo-ration and continuous selection

Functions of the collaborative model
In the previous section the operation of the collaborative module was described, in general terms the input variables were mentioned, and through two case studies, the methodology implemented to segment the availability matrix. In this section of the article we present the specific description of the collaborative algorithm, the programming structure, the implemented functions, and the input and output variables.
The collaborative algorithm segments the matrix through three functions, the

Input variable
The Fig. 8 shows the three sets of input variables: User Relation, Number User and Input Data. The description for each variable is presented in tables 5, 6 and 7 respectively.

Output variable
The Fig. 8 the algortihm has three output variables: Power Segmentation Training, User Full and User Simulation. The description of each variable show in Table 8.

Random Zone
The percentage of users selected for simulation is taken in an orderly manner, row-wise or columnwise.

Division
Column If the number of users is greater than 10, the rows of the power matrix are divided into 10 equal parts and the columns are split into n parts until the number of users is completed. If the number of users is less than 10, the rows of the power matrix are divided into 2 equal parts and the columns are split into n parts until the number of users is completed.

Row
If the number of users is greater than 10, the columns of the power matrix are divided into 10 equal parts and the rows are divided into n parts until the number of users is completed. If the number of users is less than 10, the colums of the power matrix are divided into 2 equal parts and the columns are split into n parts until the number of users is completed. User percentage 10-100 Percentage of users participating for the training.

Decision-making models
The decision-making process is done separately from the collaborative model, however, although its design is external to the collaborative block, it requires the Power Segmentation Training output matrix to work. Next, the results of the collaborative model and the decision-making will be connected through a search algorithm (section 2.4).
To analyze the decision-making process, two MCDM techniques are implemented. The chosen strategies are: Feedback Fuzzy Analytical Hierarchical Process (FFAHP) [17,28] and Simple Additive Weighting (SAW) [17,27]. Furthermore, the analysis is carried out according to the class service: real-time (RT) and best effort (BE).
The decision-making analysis uses a module called "MCDM" in charge of assigning a score to various channels. The purpose is to establish channels with a higher probability of spectral opportunity. The Fig. 9 shows the input variables required to parameterize the module, as well as the output vectors of the block. The first output vector one contains the score allocated to each channel, while the second output vector contains the ranking of the channels in ascending order according to the score obtained (Ranking). The Table 9 presents the description of each of the input vector.

FFAHP
The FFAHP algorithm (Feedback FAHP) proposes the feedback of the information from previous assessments based on the FAHP method [28]. The FAHP algorithm uses fuzzy logic as a tool which is particularly adequate to make decisions in scenarios where the inputs are often uncertain or inaccurate. In essence, the FAHP algorithm uses the same methodology of the AHP algorithm [17]. However, fuzzy logic supports in handling the subjectivity and uncertainty generated during the assessment process [28]. After normalization, the FAHP algorithm states that the weight vector is given by (1).
In methods based in the AHP algorithm, judgment matrixes must be built, which performs comparative assessments that define the level of relative importance between each possible combination of the criteria, sub criteria and alternatives, independently. For instance, the RT and BE applications are criteria based on different approaches. In RT, the sub-criteria with the highest priorities are those which reduce the delay. For BE, the subcriteria with the highest priority are those which increase the data rate, such as BW and SINR.
( ) T If the algorithm is using the RT application, the score for each channel is computed using (2), while (3) is used for BE applications. The coefficients of each of the equations are taken from [28]. The Table 9

SAW
This algorithm develops a decision making matrix comprised of criteria and alternatives.
The spectral opportunity with the highest score is chosen, where ri,j belongs to the matrix and the sum of weights is equal to 1 (4) [27,30,31]. For each intersection of the matrix, the algorithm assigns a weight ( (5) and (6)) his enables to establish a score for each assessed spectral opportunity and deliver the ranking of the alternatives. The spectral opportunity with the highest score is chosen. The coefficients of each of the equations are taken [28].

Performance metrics
The connection between the collaborative module and multicriteria techniques is carried out through a search algorithm, as shown in Fig. 10. The output variables of the "Collaborative" and "MCDM" modules correspond to the inputs of the "Search Algorithm". Additionally, this module requires two additional parameters for adjustment, the Time and Time Range variables. The output of the search algorithm corresponds the assessment metrics of failed handoff.

Search algorithm
The

Failed handoff
A spectral handoff (SH) can be defined as the process where a SU changes its operation frequency, when the conditions of a channel are degraded or when a PU asks for service, given that that the secondary user is using a licensed band [26]. The SH concept of CR differs from traditional wireless network mechanisms in the sense that there are two types of users with different priorities [2]. A failed handoff corresponds to the number of unsuccessful handoffs, meaning that the SU cannot materialize handoff since it finds that the respective spectral opportunity is occupied [2].

User Full
Power Segmentation Training User simulation Collaborative

Results and Discussion
The results are presented through two analyses. The first analyzes the number of failed handoffs, for each of the five collaboration levels (10%, 20%, 50%, 80% and 100%), according to the decision-making model, the class of service (RT, BE) and the traffic level (HT, LT). The second quantifies the collaboration level for the number failed handoffs according to the 100% of collaboration. Finally, the comparative analysis determine how much information is required to be shared.

Number of failed handoffs
The The results reveal that both multicriteria techniques show an equivalent tendency for all five levels of collaboration. The trend of the number of handoffs is equivalent for all five levels of collaboration. 100% collaboration leads to the lowest levels while 10% collaboration leads to the highest levels. The other collaboration levels can be found between these two limits.

Percentage analysis of the collaboration
To establish a collaboration criterion in terms of the number of failed handoffs, a percentage-based analysis is carried. The objective is to use 100% of the shared information as the base criterion, this level has the lowest indicators. Based on the percentage relations obtained, the level of information shared and the average increase rates, it is determined that, for a given the number of failed handoffs, the collaboration range that enables efficient results varies between 10% and 20%, where 20% of the shared information is considered the representative limit of the range. For better effort (BE), the range from 20% to 50% delivers the highest efficiency according to the average increase rates.

Availability of data and materials
Please contact the authors for data requests

Competing Interest
The authors declare that they have no competing interests

Authors Contribution
DG, CH, and ER declare that they contributed to the manuscript. All authors prepared, read and approved the final manuscript.         Table 3 Example of a collaborative model adjustment -Case study 1 Table 4 Example of a collaborative model adjustment -Case study 2 Table 5 User relation Table 6 Number of users Table 7 Power matrix Table 8 Output variables Table 9 Number of users