HANDOVER NECESSITY ESTIMATION FOR 4G HETEROGENEOUS NETWORKS

Transcription

1 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January 0 HANDOVER NECESSITY ESTIMATION FOR 4G HETEROGENEOUS NETWORKS Issaka Hassane Abdoulaziz, Li Renfa and Zeng Fanzi 3,,3 College of Information Science and Engineering, Hunan Uniersity, Changsha City, China fabismer@hotmail.com, scc_lrf@hnu.cn, zengfanzi@6.com ABSTRACT One of the most challenges of 4G network is to hae a unified network of heterogeneous wireless networks. To achiee seamless mobility in such a dierse enironment, ertical hand off is still a challenging problem. In many situations handoer failures and unnecessary handoffs are triggered causing degradation of serices, reduction in throughput and increase the blocking probability and packet loss. In this paper a new ertical handoff decision algorithm handoer necessity estimation (HNE), is proposed to minimize the number of handoer failure and unnecessary handoer in heterogeneous wireless networks. we hae proposed a multi criteria ertical handoff decision algorithm based on two parts: traeling time estimation and time threshold calculation. Our proposed methods are compared against two other methods: (a) the fixed RSS threshold based method, in which handoers between the cellular network and the WLAN are initiated when the RSS from the WLAN reaches a fixed threshold, and (b) the hysteresis based method, in which a hysteresis is introduced to preent the ping-pong effect. Simulation results show that, this method reduced the number of handoer failures and unnecessary handoers up to 80% and 70%, respectiely. KEYWORDS Handoer failure, Handoer necessity estimation HNE, Mobile Network, Time threshold, unnecessary handoer, WLAN IEEE 80., 4G network,. INTRODUCTION In the 4G wireless enironment, a mobile user is able to continue using the mobile deice while moing from one point of attachment to another. Such process is called a handoer, by which a mobile terminal keeps its connection actie when it migrates from the coerage of one network access point to another []. Depending on the access network that each point of attachment belongs to, the handoer can be either horizontal or ertical [], []. A horizontal handoer or intra-system handoer takes place between PoA supporting the same network technology, e.g., two geographically neighboring BSs of a 3G cellular network. On the other side, a ertical handoer or intersystem handoer occurs between PoA supporting different network technologies, [3] e.g., an IEEE 80. AP and a 3G BS. An example of horizontal and ertical handoers is illustrated in Figure, where a horizontal handoer happens between two WLANs and a ertical handoer takes place between an AP of a WLAN and a BS of a UMTS cellular. Vertical handoers are implemented across heterogeneous cells of access systems, which differ in seeral aspects such as bandwidth, data rate, frequency of operation, etc. The different DOI : 0.5/ijist.0.0

2 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January 0 characteristics of the networks inoled make the implementation of ertical handoers more challenging as compared to horizontal handoers [4]. Figure : illustration of horizontal and ertical handoff. Three stages are inoled in a ertical handoer process: network discoery, handoer decision, and handoer execution [5]. During the system discoery, mobile terminal equipped with multiple interfaces hae to determine which networks can be used and what serices are aailable in each network. During the handoff decision phase, the mobile deice determines which network it should connect to. During the handoff execution phase, connections are needed to be re-routed from the existing network to the new network in a seamless manner.this requirement refers to the Always Best connected (ABC) concept, which includes the authentication,authorization, as well as the transfer of user s context information [6]. This paper presents the ertical handoff management and focuses mainly on the handoff decision problem. Handoer decision is the ability to decide when to perform the handoer and to which access network to handoer. A decision for ertical handoer may depend on seeral issues relating to the network to which the mobile node is already connected and to the one that it is going to handoer. For example, the decision to perform mobile-controlled handoers may be made by a ertical handoer agent, sitting in the mobile deice based on policies such as network bandwidth, load, coerage, cost, security, QoS, or een user preferences [7]. Seeral parameters hae been proposed in the research literature for use in the ertical handoer decision (VHD) algorithms. We briefly explain each of them below. Handoer delay: Handoer delay is the duration between the initiation and completion of the handoer process, and is related to the complexity of the VHD process. Reduction of the handoer delay is especially important for delay sensitie oice or multimedia applications. Number of handoers: Reducing the number of handoers is usually preferred as frequent handoers would cause wastage of network resources [8]. A handoer is considered to be superfluous when a handoer back to the original point of attachment is needed within certain time duration [9], and the number of such handoers should be minimized. Handoer failure probability: A handoer failure occurs when the handoer is initiated but the target network does not hae sufficient resources to complete it, or when the mobile terminal moes out of the coerage of the target network before the process is finalized. In the former case, the handoer failure probability is related to the channel aailability of the target network [0], while in the latter case it is related to the mobility of the user [].

3 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January 0 Throughput: The throughput refers to the data rate deliered to the mobile terminals on the network. Handoer to a network candidate with higher throughput is usually desirable. Our proposed method (HNE) determines the necessity of making a handoer to an aailable network. HNE takes arious network parameters as its inputs and generates a binary alue as its output. The inputs include: the AP power leel, RSS samples, the radius of the network, the elocity of the MT, the handoer latency, and the handoer failure and unnecessary handoer probability requirements. An output describes of means a handoer is necessary, and an output of 0 means the handoer is not necessary. The block diagram of HNE is shown in Figure. Figure : block diagram of handoer necessity estimation (HNE) The paper is structured as follows. Section defines related work. The proposed algorithm for handoer necessity estimation (HNE) is described in section 3. Section 4 present theoretical analyzes of HNE apply to preious works. Section 5 tackles the simulations to show the system performance with the proposed technique; to end with the paper conclusion in section 6.. RELATED WORK A number of studies published earlier hae sureyed VHD algorithms. Zahran et al. [] proposed an algorithm for handoers between 3G networks and WLANs by combining the RSS measurements either with an estimated life-time metric (expected duration after which the mobile terminal will not be able to maintain its connection with the WLAN) or the aailable bandwidth of the WLAN candidate. Benefits of Zahran et al. s algorithm can be summarized as follows. First, by introducing the lifetime metric, the algorithm adapts to the application requirements and the user mobility, reducing the number of superfluous handoers significantly. Second, there is an improement on the aerage throughput for the user because of the mobile terminal s ability to remain connected to the WLAN cell as long as possible. Howeer, packet delays grow with an increase in the lifetime, due to the deterioration of the channel condition as the mobile terminal approaches the edge of the WLAN cell. This issue can be critical for delay sensitie applications and degrade their performance. 3

4 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January 0 In [3] Li Jun Zhang et a. proposes a method to send probe requests to the APs one after the other and perform handoff immediately after any AP sends the response. This allows us to scan fewer channels. All these processes inole scanning of APs, it may be selectie or all APs may be scanned. These methods are therefore time consuming as well as hae a certain probability of handoff failure. Yang et al. [4] presented a bandwidth based VHD method between WLANs and a Wideband Code Diision Multiple Access (WCDMA) network using Signal to Interference and Noise Ratio (SINR). SINR based handoers can proide users with higher oerall throughput than RSS based handoers since the aailable throughput is directly dependent on the SINR, and this algorithm results in a balanced load between the WLAN and the WCDMA networks. But such an algorithm may also introduce excessie handoers with the ariation of the SINR causing the node to hand oer back and forth between two networks, commonly referred to as ping-pong effect. Mohanty [5] presented an algorithm for calculating a boundary area based on the speed of the MT and the WLAN cell size. In this algorithm, a handoer from a WLAN to a 3G network is triggered when the MT enters the boundary area of the WLAN and handoer procedures are completed before the MT leaes the WLAN. This algorithm operates efficiently for handoers from WLAN to 3G as it reduces the handoer failure probability. Howeer, in the mobility architecture using this algorithm, and also in most of the other handoer decision methods such as in [6], handoers from the cellular network to the WLAN are initiated once the MT enters the WLAN coerage area. This is not effectie enough in situations where the MT traels through an area close to the coerage boundary of the WLAN at speeds aboe a certain threshold, since handoers to the WLAN become unnecessary. It is always better to aoid these handoers as much as possible since they lead to network resource wastage [8]. Furthermore, if the handoer process has not been completed before the MT leaes the WLAN coerage area, connection breakdown ineitably occurs. In the method presented in [5] for handoers from the cellular network to the WLAN, the MT remains connected to both networks while staying in a boundary cell of the WLAN in order to aoid connection breakdown and also the ping-pong effect. Howeer, this approach does not take into consideration the network resource wastage caused by unnecessary handoers. As yet, few studies on handoer necessity estimation or on efficient methods for minimizing unnecessary handoers hae been presented. 3. PROPOSED WORK In this paper, a handoer necessity estimation HNE method is introduced. This method is deised for minimizing handoer failures and unnecessary handoers between cellular networks and WLANs, by estimating the necessity of a handoer. Figure. The estimation inoles two steps: time threshold calculation for minimizing handoer failures and traeling time prediction for minimizing unnecessary handoers. 4

5 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January 0 Figure 3: scheme diagram of WLAN to cellular network handoer necessity estimation mechanism Figure 3 shows the trajectory of a MT traeling oer an area oer which cellular network serice is aailable and is also partially coered with a WLAN cell. The MT enters and exits the WLAN cell at points Pi and Po, respectiely, following a straight line. M is the middle point of the section of the trajectory inside the WLAN cell. The donut shaped (dashed) area is called boundary area. The radii of the outer and inner circles enclosing the boundary area are R and r, and dr and dr represent the half length of the trajectory segments inside the outer and inner circles respectiely. B is the intersection point of the trajectory and the inner circle. 3.. Time Threshold Calculation for Minimizing Handoer Failures The purpose of the time threshold calculation presented in this section is to keep the number of handoer failures under a desirable threshold. That is, for example, if the system designer has a requirement of limiting the probability of handoer failures under %, then the time threshold is adjusted to make the ratio of the number of failed handoers to the total number of handoers below %. The time threshold is calculated using mathematical modeling and probability calculation as explained below. It is assumed that the entry and exit points Pi and Po can be any arbitrarily chosen points on the circle enclosing the WLAN coerage area, with equal probability (Figure 3). Then the angles i θ and θ o are both uniformly distributed in [0, ], andθ = θi θo. The first step is to calculate the probability density function (PDF) [7] ofθ. The PDFs of the locations of Pi and Po are gien, respectiely, f ( Θ, Θ ) =, 4 0 Θ i, Θ o, i o 0, o th erw ise, () Θ = (), 0 Θ o, fp o ( o ) 0, o th e r w is e, Since the locations of Pi and Po are independent from each other, their joint PDF is gien by 5

6 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January 0 f ( Θ, Θ ) =, 0 Θ i, Θ o, 4 i o 0, otherwise, (3) The probability that, which is also the cumulatie distribution function (CDF) of θ, can be deried using the following integral [8] F ( Θ ) = P ( θ Θ ) = f ( θ θ ) dθ dθ, (4) Ω i, o o i Where Ω is the space of locations of entry and exit points Pi and Po such that θ Θ and 0 Θ. P( θ Θ ) = 0 for Θ 0 and P( θ Θ ) = for Θ >. From the obseration of Figure Equation (4) can be rewritten as F ( Θ ) = P ( θ Θ ) = 4Θ 4Θ, 0 Θ. (5) The PDF of θ can be deried by taking the deriatie of Equation (5) and is gien by { ( Θ ), 0 Θ, ( ) 0, otherw ise. f Θ = (6) The next step is to use the PDF of θ and the expression of the traeling time t WLAN as a function of θ to obtain the PDF of t WLAN. From the geometric configuration in Figure 3 and by using the cosine formula, the following equation is obtained: ( twlan ) = R ( COSθ ). Thus, t W L A N = g ( θ ) = θ (8) R ( c o s ). Using the theorem stated in [7], the PDF of t WLAN is expressed as f ( T ) = n, (9) f ( θ n ) g ( θ n ) Where θ, θ n are the root of function g(θ), and g (.) is the deriatie of g(.). In equation (8), for g(θ) there are two roots, θ and θ, which are expressed as (7) t a r W L A N R θ = c o s ( ), (0) t W LA N R θ = arcco s( ), () From (8), g (θ) is expressed as g R ( θ ) =. () s i n θ ( c o s θ ) g ( θ ) = t R s i n ( a r c c o s ( W L A N ) ) R t c o s ( a r c c o s ( W L A N ) ) R t W L A N 4 R = R (3) 6

7 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January 0 g ( θ ) = t R s i n ( a r c c o s ( W L A N R t c o s ( a r c c o s ( W L A N R = R (4) t 4 R And a r c c o s ( t W L A N ( ) R f θ ) = (5) a r c c o s ( t W L A N ( ) R f θ ) = (6) Thus using equations (9), (3) and (5) the PDF of t WLAN is calculated by f ( T ) { { = f ( θ ) f ( θ ), 0 R g + ( ) g T θ ( θ ) 0, o t h e r w i s e., 0 T, R 4 R T = (7) 0, o t h e r w i s e. The third step is to use the PDF of t WLAN to obtain the CDF of t WLAN, which is deried from the integral of Equation (7) as: [ ] F ( T ) = P r t T T = f ( T ) d T = 0, R p T, { T T R R a r c c o s ( ), 0. (8) A time threshold parametertis introduced to make handoer decisions: wheneer the estimated traeling time t WLAN is greater thant, the MT will initiate the handoer procedure. A handoer failure occurs when the traeling time inside the WLAN cell is shorter than the handoer latency from the cellular network to the WLAN, Ƭ i. Thus, using Equation (8) the probability of a handoer failure for the method using the threshold T is gien by P f { τ a rc sin ( i ) arcsin ( Τ ), 0 Τ τ, R R i 0, τ Τ. = p (9) i By using (9), an equation which can be used by the MT to calculate the alue of T for a particular alue of P f when 0< P f <: Τ = (0) R τ i s in (a rc s in ( R ) P f ) To calculate Τ, the speed of the MT and the handoer latency τ i need to be obtained. In this research, the knowledge of and i τ is assumed, and they can be measured by using accelerometers [9] and the technique described in [0], respectiely. 7

8 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January Traeling Time Prediction for Minimizing Unnecessary Handoers To eliminate unnecessary handoers, Yan et al. [9] deeloped a VHD algorithm that takes into consideration the time the mobile terminal is expected to spend within a WLAN cell. The method relies on the estimation of WLAN traeling time (i.e. time that the mobile terminal is expected to spend within the WLAN cell) and the calculation of a time threshold (t WLAN ). A handoer to a WLAN is triggered if the WLAN coerage is aailable and the estimated traeling time inside the WLAN cell is larger than the time threshold. The estimated traeling time (t WLAN ) is O S + ( s p i ) ( s in ) t R l t t = W L A N t t () Where R is the radius of the WLAN cell, l OS is the distance between the access point and where the mobile terminal takes an RSS sample, is the elocity of the mobile terminal, ts and t p i are the times at which the RSS sample is taken and the mobile terminal enters the WLAN cell coerage, respectiely. l OS is estimated by using the RSS information and log-distance path loss model. The time threshold (T WLAN ) is calculated based on arious network parameters as R τ T = sin(sin ( ) P ) () W LA N R where τ i is the handoer delay from the cellular network to the WLAN, and P is the tolerable handoer failure, unnecessary handoer or connection breakdown probability. A handoer to the cellular network is initiated if the WLAN RSS is continuously fading and the mobile terminal reaches a handoer commencement boundary area which size is dynamic to the mobile terminal s speed. Similar to the arguments used in section 3., another parameter T (T< T) is introduced to minimize the probability of unnecessary handoers. By using (8) the probability of an unnecessary handoer is calculated as { ( τ ) arcsin( i + τ o ) arcsin( Τ ), 0 Τ ( τ ), R R i + τ o P u = 0,, ( τ + τ ) p Τ. i o (3) Thus Τ = R ( τ i + τ o ) Pu sin (a rc s in ( ) ) (4) R Equation (4) is deried from (3) for a particular alue of P u when 0 < P u <. Parameters T and T depend on alues of constants Pf and P u which are selected by system designers. They also depend on measurement of, R, τ i andτ o. The parameter T can be further adjusted dynamically to encourage or discourage handoers to WLAN by considering other performance criteria such as network load. 8

9 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January 0 4. THEORETICAL ANALYSIS OF HNE FOR FIXED RSS THRESHOLD [6] AND HYSTERESIS BASED METHODS [] 4.. Handoer Failure Probability Using equation (8), the handoer failure probability for the fixed RSS threshold based method is gien by P =, τ > R, i fixed f τ fixed sin ( i ), 0 τ R fixed i R fixed where R fixed is the distance between the MT location and the AP of the WLAN cell when a handoer into the WLAN occurs in the fixed RSS threshold based method. It is calculated by (5) R 0 f i x e d E t R S S f i x e d 0 β = (6) Using Equation (8), the handoer failure probability for the hysteresis based method is gien by P =, τ > R, i hyst f τ hyst sin ( i ), 0 τ R hyst i R hyst where R hyst is the distance between the MT location and the AP of the WLAN cell when a handoer into the WLAN occurs in the hysteresis based method. It is calculated by (7) R 0 h y s t E t R S S h y s t 0 β = (8) 4.. Unnecessary Handoer Probability Using equation (8), the unnecessary handoer probability for the fixed RSS threshold based method is gien by =, ( τ + τ ) > R, P ( τ i + τ o ) fixed R fixed i o fixed u sin ( ), 0 ( τ + τ ) R i o fixed The unnecessary handoer probability for the hysteresis based method is gien by =, ( τ + τ ) > R, P ( τ i + τ o ) hyst R hyst i o hyst u sin ( ), 0 ( τ + τ ) R i o hyst (9) (30) 5. SIMULATION Our proposed methods are compared against two other methods: (a) the fixed RSS threshold based method [6], in which handoers between the cellular network and the WLAN are initiated 9

10 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January 0 when the RSS from the WLAN reaches a fixed threshold, and (b) the hysteresis based method [], in which a hysteresis is introduced to preent the ping-pong effect. 5.. Ealuation Parameter The parameters used in theoretical analysis and simulations of HNE are listed in Table. Table : Parameters used in the HNE performance ealuation. Parameter Symbol Value WLAN radius R 50 m AP transmit power P TX 0 dbm Distance between the AP and the d ref m reference point Path loss at the reference point PL ref 40dB Path loss exponent β 3.5 Standard deiation of shadow σ 4.3dB fading Handoer delay from cellular network to WLAN Handoer delay from WLAN to cellular network Tolerable handoer failure probability Tolerable unnecessary handoer probability Γ i S Γ S o P 0.0 f P 0.04 u MATLAB was used for the experiments, which generated 0000 random MT trajectories across the WLAN cell coerage area for speeds from 3.6 km/h to 00 km/h in km/h increments. For each trajectory, a random WLAN cell entry point was chosen, and a uniformly distributed random angle between 0 and was generated representing the moement direction of the MT. 5.. Simulation Analysis Two critical parametric quantities are examined: number of handoer failure and unnecessary handoer. The number of handoer failures and unnecessary handoers of the RSS threshold based ( R fixed = 50 m), hysteresis based ( R hyst = 0 m) and HNE methods are shown in Figures 4 and 5. From the figures it can be seen that, with HNE, handoer failures and unnecessary handoers are kept under the numbers of 00 and 500, respectiely. The total number of handoers declines with the increasing elocity of the MT. In the RSS threshold and hysteresis based methods, the numbers of handoer failures and unnecessary handoers increase sharply as the elocity increases. HNE is able to reduce the number of handoer failures and unnecessary handoers up to 80%, when the elocity of the MT is up to 00 km/h. HNE yields much better performance than the other two methods. Otherwise, for elocities less than 0 km/h, the other two methods yield marginally better results. 0

11 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January 0 Fig 4: Number of handoer failures of [6], [] and HNE methods. Fig 5: unnecessary handoer of [6], [] and HNE 6. CONCLUSION In this paper, a method to estimate the handoer necessity into a WLAN cell is presented. This method is based on two parts: traeling time estimation and time threshold calculation. The traeling time estimation relies on the RSS measurements and the speed of the MT. The time thresholds are calculated based on arious network parameters such as tolerable handoer failure probability or un-necessary handoer probability, the radius of the WLAN cell and the handoer latency. This method is able to reduce the number of handoer failures and unnecessary handoers up to 80% and 70%, comparing with the conentional RSS threshold based [6] and hysteresis based []. ACKNOWLEDGEMENTS This work was supported in part by the national natural science foundation of china with grant number and the science and technology planning project of Hunan proince; China. No. 00GK305 REFERENCES [] N. nasser; A. Hasswa; and H. Hassanein, 006. Handoffs in fourth generation heterogeneous networks. IEEE Communications Magazine, ol. 44, issue 0, pp [] F. zhu and J. McNair. Multiserice ertical handoff decision algorithms EURASIP Journal on Wireless Communications and Networking, ol., no. 5, April 006, pp. -3 [3] X. Guan, R. Tang, S. Bai, and D.Yoon, 007. "Enhanced Application Drien Vertical Handoff Decision Scheme for 4G Wireless Networks", Wireless Communications, Networking and Mobile Computing, Wicom 007, International Conference, pp

12 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January 0 [4] F. siddiqui and S. Zeadally Mobility management across hybrid wireless networks: Trends and challenges. Computer Communications, ol. 9, no. 9, pp [5] I. F. Akyildiz and W. Wang. 00. A dynamic location management scheme for next-generation multitier PCS systems. IEEE Transactions on Wireless Communications, ol., no., pp [6] Chandralekha,Praffula Behera. 00. Minimization of Number of Handoff Using Genetic Algorithm in Heterogeneous Wireless Networks. International Journal of Latest Trends in Computing, Vol., Issue, pp. 4-8 [7] F. zhu and J. McNair Optimizations for ertical handoff decision algorithms. In Proceedings of the 004 IEEE Wireless Communications and Networking Conference (WCNC 04), pp [8] Zahran, A.H; Sreenan, C.J. 0. Extended Handoer Keying and Modified IEEE 80. Resource Query Approach for Improing Vertical Handoff Performance. IFIP international conference on new technologies, mobility and security (NTMS) pp. -7. [9] X. yan, N. Mani, and Y. A. S. Ekercio Glu A traeling distance prediction based method to minimize unnecessary handoers from cellular networks to WLANs. IEEE Communications Letters, ol., no., pp [0] A. xhafa and O. K. Tonguz Dynamic priority queuing of handoer calls in wireless networks: an analytical framework. IEEE Journal on Selected Areas in Communications, ol., no. 5, pp [] F. Barcelo Performance analysis of handoff resource allocation strategies through the statedependent rejection scheme. IEEE Transactions on Wireless Communications, ol. 3, no. 3, pp [] A. H. zahran, B. Liang, and A. Saleh Signal threshold adaptation for ertical handoff in heterogeneous wireless networks. Mobile Networks and Applications, ol., no. 4, pp [3] Debabrata, S; Utpal Biswas. 0. Minimization of Handoff Failure by Introducing a New Cell Coerage Area in the Handoff Region. IJCSI International Journal of Computer Science Issues, Vol. 8, Issue 3, No., pp [4] K. yang, I. Gondal, B. Qiu, and L. S. Dooley Combined SINR based ertical handoff algorithm for next generation heterogeneous wireless networks. In Proceedings of the (GLOBECOM 07) Conference, pp [5] S. Mohanty A new architecture for 3G and WLAN integration and inter-system handoer management. Wireless Networks, ol., no. 6, pp [6] V. K. arma, S. Ramesh, K. D. Wong, and J. A. Friedhoffer Mobility management in integrated UMTS/WLAN networks. In Proceeding of the 003 IEEE International Conference on Communications ICC 03, pp [7] Steen M. Kay., 006. Intuitie probability and random process using MATLAB. uniersity of Rhode, Island Kluwer academic publishers. [8] C. bettstetter, H. Hartenstein, and X. P erez-costa Stochastic properties of the random waypoint mobility model. Wireless Networks, ol. 0, no. 5, pp [9] W.-H. Zhu and T. Lamarche Velocity estimation by using position and acceleration sensors. IEEE Transactions on Industrial Electronics, ol. 54, no. 5, pp [0] S. mohanty and I. F. Akyildiz A cross-layer (layer + 3) handoer management protocol for next-generation wireless systems. IEEE Transactions on Mobile Computing, ol. 5, pp

13 International Journal of Information Sciences and Techniques (IJIST) Vol., No., January 0 [] M. Liu, Z. Li, X. Guo, and E. Dutkiewicz Performance analysis and optimization of handoff algorithms in heterogeneous wireless networks. IEEE Transactions on Mobile Computing, ol. 7, no. 7, pp Authors Issaka Hassane Abdoulaziz receied the B.S. degree in Networks and Telecommunications from EST-LOKO institute, Abidjan, Cote d Ioire, in June 004 and the M.S. in computer science and Technology from Hunan Uniersity, Hunan, China, in July 009. He is currently working toward his Ph.D. degree in the college of information science and engineering at Hunan Uniersity, China. From 004 to 005 he worked in the Ministry of National Education at the Human Resource Department, Niger. He worked for the High Council of Communication, Niger, from 005 to 006. His research interests include wireless network, mobile communications, cognitie radio, and orthogonal frequency-diision multiplexing. Li Renfa born in 956. Professor and Ph.D superisor in Hunan Uniersity. His main research interests are embedded system and network. Zeng Fanzi born in 97. Associate professor, Hunan Uniersity. His main research interests include UWB and cognitie radio hort Biography 3