fuzzy-based handover decision scheme for next-generation heterogeneous wireless networks ·...

Fuzzy-Based Handover Decision Scheme for Next-Generation Heterogeneous Wireless Networks Shih-Jung Wu

Journal of Convergence Information Technology, Volume 6, Number 4. April 2011

Fuzzy-Based Handover Decision Scheme for Next-Generation Heterogeneous Wireless Networks

Shih-Jung Wu

Department of Innovative Information and Technology, Tamkang University, 180 Linwei Road., Chiao-hsi Shiang, I-lan County, 26247, Taiwan R.O.C.

Phone: +886 -3-987-3088 ext.7121 E-mail: [email protected]

doi:10.4156/jcit.vol6. issue4.31

Abstract In the development of the wireless communication of the future, mobility management and existing

heterogeneous wireless networks integration in next generation or 4G wireless networks are developing trend. Such an environment poses a significant challenge: how can we choose an appropriate Network Access Point (NAP) to achieve load balance and offer high quality communication in heterogeneous wireless handover decisions? However, mobility management as well as the integration and interworking are a complex task due to their specific characteristics. This paper we apply fuzzy logic in heterogeneous wireless network handover decisions and introduce the fuzzy normalization concept to make parameters for heterogeneous wireless network normalization. Using a simulation, we determine under what conditions our handover decision algorithm can balance network system loading better than the traditional fuzzy algorithms applied in heterogeneous wireless network handover decisions.

Keywords: Wireless, Handover Decision, Fuzzy, NGWN/4G

1. Introduction

The main concern of the next-generation wireless networks (NGWN/4G) is the integration of heterogeneous wireless technologies. This integration invokes some challenges, such as handover decision making and mobility management. Provisioning of seamless handover is one of the key topics in NGWN/4G. Vertical handover decisions evaluate candidate NAPs based on different standards. Making a suitable handover decision according to different standard based parameters of NAPs in heterogeneous wireless networks is worth researching. Whether roaming or handover is occurring, situations where target NAPs are overloaded and cause failure of the handover procedure happen easily in homogeneous wireless networks. But the failure will not easily occur in heterogeneous works because some heterogeneous NAPs have the same coverage areas, which we can be implemented to balance the load among the NAPs. Obviously if we do not consider the load balance of target NAPs, it will cause some NAPs that offer good channel conditions to overload. We should take account of the load balance in the handover decision mechanism. Every NAP has different capabilities of MN velocity and RSS (received signal strength) in heterogeneous wireless networks. For example, IEEE 802.11 NAPs only can accept human walk speed, whereas 3G can accept cars moving on highways. Consequently, there is a need to consider the velocity requirements and RSS. To obtain system loading balance and to avoid failures caused by MN handovers to NAPs with lower velocity capabilities and weaker RSS requirements, we present the FUzzy Normalization - HandOver Decision Strategy algorithm (FUN-HODS). We apply the characteristics of fuzzy normalization to handover decisions to solve questions of parameter comparison and preprocessing. We make vertical handover decisions as simple as horizontal decisions.

Section 2 introduces related work, and acquaints readers with how fuzzy theory is applied in handovers. The application of FUN-HODS in heterogeneous wireless networks will be discussed in detail in Section 3. Section 4 presents a simulation of FUN-HODS and a traditional fuzzy algorithm. The simulation results verify FUN-HODS efficiency. Finally, conclusions are stated in Section 5.

- 285 -



2. Related work

2.1. Mobility management

In order to enhance the experience of mobile users by facilitating handovers between heterogeneous

wireless networks, the IEEE 802.21 Working Group does an effort to ratify the Media Independent Handover (MIH) [1] standard. The MIH framework offers the common interface and provides link layer and other related network information to upper layers. The MIH Function (MIHF) offers three types of services: MIES (Media-Independent Event Service), MICS (Media-Independent Command Service) and MIIS (Media-Independent Information Service). Moreover, mobility management is one of the most important issues in the mobile communication, and it affects efficiency of the whole mobility network apparently. The reaching of mobility can be implemented though handover and mobility management protocol, containing Session Initiation Protocol (SIP) at application layer and Mobile IP (MIP) at network layer. Meanwhile, the protocol can improve the efficiency that heterogeneous wireless network integrated.

2.2. Interworking architecture for NGWN

In order to satisfy seamless mobility in heterogeneous wireless network, we need a heterogeneous

mobile integrated architecture in Figure 1. Which utilizes MIH in mobility management protocol.[2] In the following further details on this architecture are provided. The Mobile Node (MN) has IEEE802.11/WiFi, WiMAX and E-UTRAN (contain 3G) network interfaces, which support MIHF and mobility. The Correspondent Node (CN) has IEEE 802.3 network interfaces without MIHF and mobility. Moreover, a number of various access networks, such as IEEE802.11/WiFi, WiMAX and E-UTRAN, are connected to a common core network (the EPC) based on IP technology through different interfaces. All 3GPP networks are connected through the serving gateway 1 (S-GW1), and all non-3GPP networks are connected through the S-GW2. Different paths also are used in the case of IEEE802.11/WiFi and WiMAX. A WiMAX network is considered trusted non-3GPP accesses and directly connected to the S-GW2. On the other hand, a IEEE802.11/WiFi network is considered as distrusted access connects to the evolved packet data gateway (ePDG) and then to the S-GW2. For E-UTRAN, the S-GW1 is directly connected to it. The Mobility Management Entity (MME) is incorporated in the architecture for handling control functions such as authentication, security, and mobility. For distrusted non-3GPP accesses, the ePDG secures the access of the MN to the EPC by means of an IPSec tunnel between itself and the MN. All data paths from the access networks are combined at the P-GW, which incorporates functionality such as packet filtering, QoS policing, interception, charging, and IP address allocation, and routes traffic control. Besides, EPC also contains network control entities for keeping user subscription information (home subscriber server [HSS]), determining the identity and privileges of a user and tracking his/her activities (authentication, authorization, and accounting [AAA] server), and enforcing charging and QoS policies through a policy and charging rules function (PCRF). In order to achieve seamless handover of the architecture, extra functionality is needed in the network elements, based on the IEEE 802.21 protocol. The MIH functionality placed at the mobile node (MN), the wireless accesses networks (MIH PoSs), and the operator’s IP network (MIIS server).

- 286 -



Figure 1.Interworking architecture for NGWN

2.3. Vertical Handover

There are descriptions about varied handovers presented in [3-8, 10, 30]. According to network type,

handovers can be classified as horizontal handovers and vertical handovers. In this paper we emphasize the importance of vertical handovers and handover management. Handover procedures consist of three phases: handover detection, handover decision and handover execution. In handover detection, we need to collect some handover parameters (e.g., RSS, system loading, velocity and latency, etc.) about the NAPs of candidate networks. Our methodology, FUN-HODS, considers three parameters, RSS, loading and velocity, in heterogeneous wireless networks.

2.3.1 Fuzzy System Applied to Handover Decisions

Fuzzy systems are one of the most popular methods for processing imprecise values. In [9]-[29]

various fuzzy systems applied in handover decisions are presented. In Figure 2, the role of the fuzzifier in a fuzzy system is to convert a crisp input variable into a fuzzy set that is ready to be processed by the inference engine. The inference engine, using the fuzzifier inputs and the if-then rules stored in the rule base, processes the incoming data and produces a fuzzy output. The defuzzifier converts fuzzy variables to crisp variables. In traditional fuzzy systems applied to heterogeneous wireless network handover decisions, we can understand the merit of traditional fuzzy algorithms. We will present some traditional fuzzy algorithms in handover decisions and present improvements to these algorithms in the following discussion.

Figure 2. Flow chart of the proposed classification of vertical handover decisions.

In Majlesi and Khalaj [11], the author makes use of the Doppler frequency and a fuzzy logic system

in the handover decision algorithm called the Adaptive Fuzzy Logic Based Handover Algorithm for Hybrid Networks. As you can see in Figure 3, the initial process proceeds from Measurement through Averaging to Start Handover. The author considers the Doppler frequency in Averaging. When

- 287 -



Averaging compares the MN speed, which it gets in Measurements, if the MN speed is higher, Start Handover will decrease trigger handover times to avoid handover delay, or else Start Handover would increase trigger handover times to get more suitable networks. MN speed and Traffic of WLAN (TRw) would transfer HYSw and HYSm by fuzzy logic. The algorithm does not consider MN speed compatibility to NAPs with different wireless technologies, so when the MN speed is higher than the MN speed compatibility of the WLAN NAP, the handover will fail. The algorithm also does not consider system loading and complex preprocessing. System loading overload will lead to a handover fail.

The author presents the multi-criteria decision-making (MCDM) concept in [12]. It uses a different threshold in every network technology of a heterogeneous network. We should give a handover decision criterion for every network technology of a heterogeneous network. In Figure 4 MCDM uses a fuzzy logic system and a muti-input single output (MISO) method to simplify the handover decision complexity. There are signal strength (SS), network load (NL), network cost (NC) and bandwidth requirements (BR) used to in the fuzzifier of the MCDM method. He uses fuzzy logic to transform different parameters into parameters that can be compared across different network technologies.

Figure 3. Adaptive Fuzzy Logic Based Handover Algorithm for Hybrid Networks

Figure 4. Block diagram of the access network selection base on MCDM fuzzy.

The algorithm in [11] did not use a normalization of the fuzzy logic system to normalize the

parameters of the heterogeneous wireless network. The algorithm in [12] does not have the preprocessing step and parameter normalization of heterogeneous wireless networks. These two algorithms can actually be improved through the use of fuzzy normalization.

3. FUN-HODS

There are several types of main wireless technologies, including 3G, IEEE 802.11 and WiMAX, in

the present wireless network trends. We use some important and significant parameters, including the

- 288 -



RSS, velocity of MN (velocity), the system loading of an NAP (loading). We aspire to balance the system load of every system for all internet service providers (ISPs). In heterogeneous wireless networks, there are different velocity capabilities of MN velocity; for example, the normal MN velocity for communications of IEEE 802.11 is slower than that of GSM. So, we can't only consider receive signal strength as the basis in the heterogeneous wireless network, we should consider the MN velocity compatibility of different networks as well. We present the FUN-HODS algorithm, which includes three main components as shown in Figure 5. They are Environment Information Obtain, Fuzzy Normalization, Handover Decision Strategy and Handover execution. The first part is Information Obtain, which involves the collection of the environmental information that is needed to obtain fuzzy normalization and make a handover decision. In this part, we collect the system loading of the candidate NAP, the MN velocity capability of the candidate NAP, and the RSS between the candidate NAP and the MN; then we put non-process information into a FUzzy Normalization (FUN). FUN gives loading, velocity, and RSS different member functions to fulfill parameter normalization and implements a defuzzifier according to fuzzy logic if-then rules. Then it generates the rank of the NAPs to choose the suitable NAP in the handover decision.

Fuzzy normalization of FUN-HODS is not constrained by a particular wireless network technology or by particular parameters, so we use 3G, IEEE 802.11, and WiMAX. Parameter preprocessing of fuzzy normalization has many advantages. It can deal with parameters of different network technologies and allow for the comparison of parameters of different network technologies. The normalization parameter is a relativity standard parameter and a handover decision does not need to consider the parameter diversity.

Figure 5. FUN-HODS

The second advantage is that the design can be applied to any mobile management of heterogeneous

wireless networks. The procedure of the normalization is not confined to some particular wireless network technology, even if the network is not 3G, IEEE 802.11 or WiMAX, as long as the required parameters can be established. FUN-HODS can be applied to other heterogeneous wireless network environments. The third advantage of this methodology is portability and capability. It can filter unsuitable candidate NAPs in the fuzzy normalization procedure, reduce handover decision complexity and increase efficiency. Section 3 will be divided into two parts: an introduction to fuzzy normalization of FUN-HODS in section 3.1, and an explanation of how HODS makes handover decisions in section 3.2.

3.1. FUzzy Normalization (FUN)

We introduced the concept of fuzzy system logic in Section 2.3.1, in this section we will show how

to carry out normalization of the fuzzy logic system. From Figure 2, we know the three parts of the fuzzy system. Those are the input parameter the fuzzifier, inference engine and fuzzy rule, and the defuzzifier output parameter.

- 289 -



3.1.1. Fuzzifier

We assume 3G, IEEE 802.11 and WiMAX are the available systems in our heterogeneous wireless

network integration environment. We use fuzzy normalization to normalize the handover decision parameter. In Section III we described how our handover decision would focus on system loading of NAPs, RSS, and velocity of the MN. Figure 6 shows all the fuzzifiers that include 3G, IEEE 802.11, and WiMAX input parameters. After confirming the parameters that the fuzzifier needs, we should, according to the character of the different wireless network technologies, give a suitable fuzzy set. One can determine the fuzzy set of different wireless network technologies with different ownership membership functions from Figure 7-9. As for the RSS parameter in Figure 7, the RSS threshold of each wireless network technology is different, so we must set up a fuzzy set for each wireless network technology. Figure 8 shows the fuzzifier for the velocity parameter. From it we can determine the MN velocity compatibility of a wireless network, and the suitable fuzzy set that we offer. Figure 9 shows the fuzzy logic for the loading parameter. Loading is measured as a percent (a relative value) for all of the network systems. So, the fuzzy set of different networks is the same, in fact, because system loading is a parameter that can be compared directly for each network.

Consider a MN that is moving out of the range of a WiMAX NAP, and it obtains a new WiMAX RSS value between -155 dBm and -125 dBm. At the same time, the MN gets a 3G RSS value that is higher than -100 dBm and an IEEE 802.11 RSS value that is lower than -105 dBm. The MN can’t direct compare to decide which should be the target NAP. If the handover goes to the 3G NAP after direct comparison of RSS, handover may not fail, but if the handover goes to the IEEE 802.11 NAP, the handover must fail because the RSS drop down. The fuzzifier in this process will give the WiMAX RSS a low grade, the IEEE 802.11 RSS gets a low grade, and the 3G RSS gets a high grade. The MN will choose the 3G NAP with the high grade, to avoid choosing the IEEE 802.11 NAP with the low grade.

Figure 6. Fuzzy system of fuzzy normalization.

Figure 7. Membership function of RSS for 3G、WiMAX、and IEEE 802.11.

- 290 -



Figure 8. Membership function of velocity for 3G、WiMAX、and IEEE 802.11.

Figure 9. Membership function of loading for 3G、WiMAX、and IEEE 802.11.

3.1.2. Fuzzy Rule Base, Inference Engine and Defuzzifier

After the fuzzifier procedure, all parameters are distinguished in the fuzzy sets. The same type

parameters of different wireless networks can be directly compared after the fuzzifier procedure, because the parameters will be normalized to the same comparison base. The fuzzifier just processes the parameters of different wireless networks to directly compare them, then the inference engine references the rules of the fuzzy rule base to infer a result. The inference engine is based on the if-then rules defined as listed in Table 1. We will estimate the RSS, velocity, and loading parameters in every NAP. Every parameter has high, medium and low subsets, so there are 27 rules in the fuzzy rule base. Due to experience requirements of the system subscriber and repeated simulations, we can simplify this to 14 rules. Finally, according to the fuzzy rules we divide the quality into five ranks. We do simple quantification in the defuzzifier.

Table 1. Fuzzy Rules

- 291 -



3.2. HandOver Decision Strategy (HODS)

Figure 10. Handover decision strategy.

After the fuzzy normalization, every estimated NAP would get a rank value, this value distributed in

subset of defuzzifier, and then we take the NAP with the best rank subset of the defuzzifier. That can filter out unsuitable NAPs that have low RSS, who could not satisfy the MN velocity requirement, or whose loading is too heavy. The above situations will be filtered out after fuzzy normalization.

In Figure 10, after the fuzzy normalization, all NAPs have a rank value. The handover decision first preserves the NAPs that have the best rank. Then, handover decision considers user preference. If user didn’t assign a wireless technology, the handover decision would judge whether the NAP number is 0, 1, or more. If number is 0, decision would go back to the environment information obtained. If the NAP number is 1, the decision would choose the NAP to execute handover procedure. If NAP number is neither 0 nor 1, the handover decision would randomly choose a NAP to execute the handover procedure.

If the user assigned a particular network technology, the decision would look for NAPs of the particular network in the best subset and judge the NAP number. If the number is 1, the decision would directly execute the handover procedure. If the NAP number neither 0 nor 1, the handover decision would randomly choose an NAP to execute the handover procedure. If the number is 0, the decision would go to the user none assigned procedure to decide on a NAP and avoid handover fail.

4. Simulation

Figure 11. Simulation environment.

- 292 -



4.1. Simulation Environment

We assume that 3G, WiMAX, and IEEE 802.11 are the heterogeneous wireless network

environments in our algorithm. As shown in Figure 11, a WiMAX NAP, two 3G NAPs, and eighteen IEEE 802.11 NAPs make up the simulation environment. Figure 11 also shows NAP positions and MN movement direction clearly. We allocate 50 MNs, and every MN would set its initial location, movement direction, movement velocity, and loading of NAPs at random. The FUN-HODS algorithm needs the parameters and values we used in Section 3. It applies real-world needs to optimize testing and to get the best value. We will carry out the simulation for 100 seconds in the environment described above, and we will take a sample once every 20 seconds. Five samples (we show three figures) are selected as to check the system loading of the NAPs, and after the simulation finishes we calculate the handover fail probability. The handover fail probability is the number of handover fails divided by the number of times the handover was executed. The handover fail in this simulated environment in our paper only considers overload, too high velocity and too weak RSS in NAPs; we do not care about the other factors that cause handover fail. The following charts are the results of 10 simulations according to the above-mentioned procedure.

The simulation results in Table 2 proves that the FUN-HODS algorithm can get better loading balance and that it is possible for MNs to satisfy their requests and make each NAP loading amount 20%. Figure 12-14 indicate data of T0, T40 and T100 (unit: second), respectively. We can see that the NAP system loading of the FUN-HODS algorithm are 20% in Figure 12-14; even the highest NAP system loading is only 36%. But, for the traditional fuzzy algorithm, some NAPs are loaded at 100% and some at 0%. So, we can prove that for a simulation executed in a certain time, the NAP system loading of the FUN-HODS algorithm is more balanced than the NAP system loading of the traditional fuzzy algorithm. The MN movement velocity is random in the simulation system in Table 3. We can see that the handover fail probability in the FUN-HODS algorithm is lower than handover fail probability in the traditional fuzzy algorithm when each MN has random velocity. In Table 4, we assigned every MN movement a velocity between 20 and 118 km/h and increase the velocity by 2 km/h for every next MN.

Table 2. System loading (%) for fuzzy normalization (FUN-HODS) and a traditional fuzzy algorithm

- 293 -



Figure 12. System Loading (T0)

4.2. Simulation Result

We can see that the handover fail probability in the FUN-HODS algorithm is also lower than the

handover fail probability for a traditional fuzzy algorithm; when the velocity is below 70 km/h, FUN-HODS almost avoids handover fail (handover fail probability is 0%), but the traditional algorithm has a 7%-40% handover fail probability. When the velocity is between 70 to 100 km/h, FUN-HODS only has a 15%-50% handover fail probability, but the traditional algorithm has a 79%-99% handover fail probability. In this velocity interval, FUN-HODS demonstrate a decreased handover fail probability than a traditional fuzzy algorithm. When velocity is over 100 km/h, the NAP cannot accept the MN velocity, and both algorithms cannot decrease the handover fail probability. We can clearly know that when the velocity is below 100 km/h FUN-HODS has a lower handover fail probability than a traditional algorithm.


- 294 -




5. Conclusions and future work

We propose the FUN-HODS algorithm to process the handover decision in heterogeneous wireless

networks. Fuzzy normalization not only includes the advantage of imprecise data dealing, but also enables the comparison of different parameters in heterogeneous networks. It can make choosing a NAP easier and raise performance. After several simulations, we modify the inference engine values and change the fuzzy rule base to find the most efficient decision. At last, it proves that the FUN-HODS algorithm can balance system loading and decrease handover failure due to velocity and weak RSS. In the future, we would like to construct robust a quantification for the defuzzifier of the FUN-HODS algorithm. If the system subscriber has special demand, we can adjust the weight of the parameters flexibly with quantification in the defuzzifier. We should also consider in more detail QoS issues in FUN-HODS and cooperate with appropriate handover execution to perfect the whole handover procedure.

Table 3. Handover fails for fuzzy normalization (FUN-HODS) and a traditional fuzzy algorithm

(random velocity).

- 295 -



Table 4. Handover fails for fuzzy normalization (FUN-HODS) and a traditional fuzzy algorithm

(assigned velocity).

6. References

[1] IEEE P802.21/D14, "Draft IEEE Standard for Local and Metropolitan Area Networks: Media Independent Handover Services", Sep. 2008.

[2] George Lampropoulos, "Media-Independent Handover for Seamless Service Provision in Heterogeneous Network", IEEE Commun. Mag., Vol. 46, Issue 1, pp. 64-71, Jan. 2008.

[3] M. Buddhikot, G. Chandranmenon, S. Han, Y.W. Lee, S. Miller, L. Salgarelli, "Integration of 802.11 and Third-generation Wireless Data Networks", in Proceedings of INFCOM, Vol. 1, pp. 503- 512, Apr. 2003.

[4] F. Siddiqui, S. Zeadally, S. Fowler, "Novel Architecture for Roaming between 3G and Wireless LANs", in Proceedings of MSAN, pp.101-105, June 2005.

[5] E. Stevens-Navarro, V.W.S. Wong, "Comparison between Vertical Handover Decision Algorithms for Heterogeneous Wireless Networks", in Proceedings of VTC, Vol. 2, pp. 947-951, May 2006.

[6] Nasser, N., Hasswa, A., Hassanein, H., "Handover in fourth generation heterogeneous networks", IEEE Commun. Mag., Vol. 44, Issue 10, pp. 96-103, Oct. 2006.

[7] Akyildiz, I.F., McNair, J., Ho, J.S.M., Uzunalioglu, H., Wenye Wang, "Mobility management in next-generation wireless systems", in Proceedings of the IEEE, Vol.87, Issue 8, pp. 1347-1384, Aug. 1999.

[8] Hasina Attaullah , Muhammad Younus Javed , "QoS based Vertical handover between UMTS, WiFi and WiMAX Networks", JCIT: Journal of Convergence Information Technology, Vol. 4, No. 3, pp. 59-64, 2009.

[9] Qian Zhang, Chuanxiong Guo, Zihua Guo, Wenwu Zhu, "Efficient mobility management for vertical handover between WWAN and WLAN", IEEE Commun. Mag., Vol. 41, Issue 11, pp. 102-108, Nov. 2003

[10] Pyung-Soo Kim, and Yong Jin Kim, "Hierarchical Mobile IPv6 Based Fast Vertical Handover using IEEE 802.21 Media Independent Handover Function", JCIT: Journal of Convergence Information Technology, Vol. 2, No. 4, pp.41-45, 2007.

[11] Majlesi A., Khalaj B.H., "An adaptive fuzzy logic based handover algorithm for interworking between WLANs and mobile networks", Personal, Indoor and Mobile Radio Communications, Vol. 5, pp. 2446-2451, Sept. 2002.

[12] Bing Hongyan, He Chen, Jiang Lingge, "Intelligent signal processing of mobility management for heterogeneous networks", in Proceedings of Neural Networks and Signal Processing, Vol. 2, pp.1578-1581, Dec. 2003.

[13] Onel Tolga, Ersoy Cem, Cayirci Erdal, Parr Gerard, "A multicriteria handover decision scheme for the next generation tactical communications systems", Computer Networks, Vol. 46, Issue 5, pp. 695-708, Dec. 2004.

- 296 -



[14] Ezzouhairi A., Quintero A., Pierre S., "A fuzzy decision making strategy for vertical handover", in

Proceedings of CCECE, pp. 583-588, May 2008. [15] Edwards George, Kandel Abraham, Sankar Ravi, "Fuzzy handover algorithms for wireless

communication", Fuzzy Sets and Systems Vol. 110, Issue 3, pp. 379-388, March 2000. [16] Cheng-Jian Lin, I-Ta Tsai, Chi-Yung Lee, "An adaptive fuzzy predictor based handover algorithm

for heterogeneous network", in Proceedings of NAFIPS, Vol. 2, pp. 944-947, IEEE Annual Meeting of the June 2004.

[17] Liu Xia, Ling-ge Jiang, Chen He, "A Novel Fuzzy Logic Vertical Handover Algorithm with Aid of Differential Prediction and Pre-Decision Method", in Proceedings of ICC, pp. 5665-5670, IEEE International Conference on June 2007.

[18] García-Gaytán Cyntia, Munoz-Rodríguez David, Soto Rogelio, "Uncertainty in cellular fuzzy handover", Expert Systems with Applications, Volume: 15, Issue: 3-4, pp. 391-397, Oct., 1998,

[19] Wu G, Havinga P.J.M., and Mizuno M., "MIRAI Architecture for Heterogeneous Network", IEEE Commun. Mag., Vol. 40, Issue 2, pp. 126-134, Feb. 2002.

[20] MStemm and RH Katz, "Vertical handover in wireless overlay networks", Mobile Networks and Applications, Vol. 3, Issue 4, pp.335-350, 1998.

[21] Christoph Niedermeier, Reiner Schmid, Eiman Mohyeldin, Markus Dillinger, "Handover Management and Strategies for Reconfigurable Terminals", Software Defined Radio Forum Contribution , Sept. 2002

[22] Kaveh Pahlavan, Prashant Krishnamurthy, Ahmad Hatami, Worcester Polytechnic Institute, Mika Ylianttila,etc. "Handover in hybrid mobile data networks", in Proceedings of IEEE Personal Communications, Vol. 7, Issue 2, pp. 34-47, Apr. 2000.

[23] Shen, Xuemin, Mark, Jon W., "Mobility information for resource management in wireless ATM networks Computer Networks", Computer Networks, Vol. 31, Issue 9-10, pp. 1049-1062, May 1999.

[24] Paolo Dadone, "Design Optimization of Fuzzy Logic Systems" ph. D Dissertation, Xrginia Polytechnic Institute and State University, May 18,2001.

[25] P.M.L.Chan, YF. Hu, R.E. Sheriff, "Implementation of Fuzzy Multiple Objective Decision Making Algorithm in a Heterogeneous Mobile Environment", in Proceedings of WCNC, Vol. 1, pp. 332-336, Mar. 2002.

[26] N. Tripathi, "Generic Adaptive Handover Algorithms using Fuzzy Logic and Neural Networks", Ph.D. Thesis, Virginia Polytechnic Institute and State University, August 1997.

[27] ChoongHo Cho, YouChang Ko, Hyung Lee-Kwang, "Fuzzy adaptive guard channel assignment strategy for handoff in PCS system", in Proceedings of the Sixth IEEE International Conference, Vol. 3, pp. 1511-1517, Jul. 1997.

[28] Ching-Hung Lee, Chun-Jie Yu, "An intelligent handoff algorithm for wireless communication systems using grey prediction and fuzzy decision system", in Proceedings of Networking, Sensing and Control, Vol. 1, pp. 541-546, March 2004 IEEE International Conference.

[29] Kwong Chiew Foong, Chuah Teong Chee, Lee Sze Wei, "Adaptive Network Fuzzy Inference System (ANFIS) Handoff Algorithm", in Proceedings of ICFCC, pp.195-198, Apr. 2009.

[30] Tarek Bchini , Nabil Tabbane , Emmanuel Chaput , Sami Tabbane & Andre-Luc Beylot , "Interconnection & Handover between IEEE 802.16e & UMTS", IJACT : International Journal of Advancements in Computing Technology, Vol. 1, No. 2, pp. 99-09, 2009.

- 297 -

fuzzy-based handover decision scheme for next-generation heterogeneous wireless networks ·...

Documents