Liu Weifeng Wu Meng
(College of Communications and Information Engineering, Nanjing University of Posts and Telecommunications, Nanjing 210003, China)
Abstrac t:Multi-criteria handoff algorithms have been playing a more important role than the traditional handoff algorithms.In order to balance the satisfaction of users and the efficiency of networks,it is necessary to develop new technologies to improve the validity of handoff algorithms.Intelligent and optimized handoff algorithms in hybrid networks that integrate Ad hoc and mobile cellular systems are well-adaptive and robust.They are able to implement handoffs adaptively,according to specific multi-factors such as different Quality of Service(QoS)requirements,network states and mobile node conditions in the future hybrid networks.Therefore,these intelligent and optimized algorithms can make more effective handover decision,and accordingly improve the system's performance.The future research will tackle intelligent or optimized vertical handoff algorithms for integrated Ad hoc and mobile cellular networks to improve their whole system performance.
C ellular networks provide pervasive coverage and support high mobility while Ad hoc networks enable high-rate data transmission in hot spots of communications.Therefore,the two kinds of networks are regarded as complementary,and their integration will improve their systems'performance separately.Ad-hoc relaying fulfills the expansion of the air interface of cellular networks,extending the coverage of cellular networks.Besides,multihop relaying for cellular networks decreases transmitting power of Mobile Terminals(MT).Ad-hoc relaying also alleviates inter-celland intra-cell interference to improve system capacity and solve the problem of traffic load balancing at hot spots.Figure 1 shows two scenarios in the integrated Ad hoc and mobile cellular network.All the mobile nodes in the figure can access the Base Station(BS)through a two-hop relay.By relaying,the mobile nodes in Scenario 1 decrease transmitting powers and improve transmission rates while those in the dead spot in Scenario 2 can access the BS,that is,the coverage of the BSis extended.Therefore,the stability and robustness of the entire system is improved.
Afew problems have to be solved in order to provide users with transparent and self-configurable service in the integrated Ad hoc and mobile cellular network environment.
▲Figure 1. Scenarios in the integrated Ad Hoc and mobile cellular network.
One basic problem is the selection of technology(or communication interface)for a specialapplication to establish a connection.Another problem is how to transfer a connection in operation from one interface to another(i.e.,the vertical handoff).
In an Ad hoc supported mobile cellular communication system,the handoff plays a crucial role in system implementation.However,the handoff in the heterogeneous Ad hoc and mobile cellular network is more complicated than that in traditional cellular networks,involving the time and the handoff types.Mobile users'profiles(such as the price,data rate,battery lifetime,service type and moving modes)and network situations(such as signal intensity,link quality,obtainable bandwidth and congestion)can be used to help select access networks and handoffs.
Therefore,effective handoff algorithms not only meet users'demands,but also improve performance of the entire system and fulfill traffic load balancing by the selection of a specific technology or by executing a vertical handoff decision.
The development of mobile communication technologies is transforming existing networks into next-generation mobile Internet,which integrates wireless networks such as mobile cellular,Ad hoc and wireless local area networks with the wired Internet to offer"everlasting online",highest possible data rates,and dynamic network access.
The current study of wireless network convergence focuses on the integration of random two networks,especially the integration of cellular networks with Wireless Local Area Networks(WLAN)and of cellular networks with Ad hoc networks as well.Cellular networks and WLAN are all mature networks,so integrating them is now common.On the other hand,integrated cellular and Ad hoc networks get wide attention because of self-organization and self-maintenance of Ad hoc networks.The integrated Ad hoc and cellular network has several performance strengths;so,it has become one of the most researched topics in the mobile communications field.The research organizations related worldwide have been engaged in the study on the integration of Mobile Ad-hoc Network(MANET)with mobile cellular communication system[1].
Opportunities Driven Multiple Access(ODMA)[2]is an attempt to introduce Ad hoc networks into the 3G mobile communications system.For cellular systems,the path loss of channels makes the cell boundary only support low transmission rates.ODMA can extend the high-rate coverage of Time Division Duplex(TDD)system to the cell boundary by multihop relaying for radio transmission.In this way,ODMA reduces the transmission power and shared channel interference,expands the system capacity,and improves effective radio transmission at the cell boundary.
Aggelou and Tafazolli[3]propose an A-GSM architecture that introduces the Ad hoc network into the GSM system.The added relaying function in the architecture is used to extend the system coverage.In this way,the A-GSM system efficiently solves the problem of dead spots,improves system capacity,and enhances its robustness against erroneous links.Basically,it is similar to the ODMAsystem,except that A-GSM supports communications of the mobile nodes out of the coverage of the base station,which is not supported by ODMA.
Wu Hongyiet al[4]address the architecture of Integrated Cellular and Ad hoc Relaying Systems(iCAR).The system integrates cellular and modern ad hoc relaying technologies to efficiently and dynamically balance traffic loads.The iCAR architecture designs a set of Ad hoc Relaying Stations(ARS)in a geographic area,signals from mobile nodes,and the BSare transferred by ARS.Each ARShas two air interfaces:one is Cellular Interface for communications with the BS;and the other is Relaying Interface for communications with mobile nodes and other ARSes.ARScan relay traffic from one over-loaded cell to a cell without congestion,and the iCARsystem uses ARSto balance inter-cell traffic load.
Luo Haiyun et alpropose the Unified Cellular and Ad-Hoc Network(UCAN)architecture.The basic goal of UCANis to enhance cell throughput by using multihop routing when the quality of downlink signals is getting worse between the BSand the mobile node.The UCAN system uses proxy clients with better downlink quality to relay packets to destination clients.Each user terminal in UCANhas two air interfaces to support both 3G cellular and IEEE 802.11-based peer-to-peer links.
Zhou and Yang[6]propose Pervasive Ad-hoc Relaying for Cellular System(PARCelS).This is a new architecture that introduces Ad-hoc relaying into the cellular system and makes use of the strengths of both Ad hoc and cellular networks.PARCelScan balance traffic load and avoid traffic congestion,and aims to solve the problem of hot spots in current cellular networks.
When an MTin the traditional cellular network moves away from a BS,the signal quality becomes worse,and it is necessary to make a handoff from the BSto another one.The handoff is a mechanism that transfers the connection between a mobile node and its corresponding terminal from one connection point to another[7].Connection points in cellular networks are called BSs while those in WLAN are Access Points(AP).
A handoff algorithm includes three steps:handoff initiation,handoff decision,and handoff execution.Handoff initiation refers to the process in which the system decides whether to make a handoff according to the measurement results of such parameters as signalstrength or not.In handoff decision,the system uses certain algorithms to select a network as the object of the handoff,based on the circumstances or users'selection.
Handoff execution refers to radio link transmission and channel allocation for handoff calling during the handoff.The handoff metrics tell signal quality,and indicate whether to make a handof or not.The metrics used for inter-cell handoff decision in both mobile voice and mobile data networks are as follows[8]:[5]
▲Figure 2. Traditional RSS-based handoff algorithms.
(1)Received Signal Strength(RSS),path loss,Carrier to Interference Ratio(CIR),Signal to Interference Ratio(SIR),Bit Error Rate(BER),Block Error Rate(BLER),Symbol Error Rate(SER),power budget,and cell ranking are individually or jointly used for specific mobile voice or mobile data networks.
(2)In order to avoid the Ping-Pong Effect,that is,frequent handoffs take place to and fro two BSs because of rapid fluctuation of the strength of received signals at the BSs,the handoff algorithms use extra parameters such as hysteresis margin,dwelltimers,and averaging windows.The parameters also include the distance between MTand AP,velocity of MS,and traffic characteristics in the serving cell.These parameters(if they can be obtained)help make a more intelligent handoff decision.
The performance of handoff algorithms[9]is determined by the certain performance testing results.Performance indexes for mobile voice networks and those for mobile data networks are different.
(1)According to the performance testing in mobile voice networks,the indexes related include call blocking probability,inter-cell handoff blocking probability,delay between inter-cell handoff request and execution,and call interruption probability.The inter-cell handoff rate(inter-cell handoff times per unit time)is related to the Ping Pong Effect.Therefore,the algorithm design usually minimizes unnecessary inter-cellhandoffs.
(2)The performance testing shows that the inter-cell handoff rate is also very important in mobile data networks.Besides,there are other important issues in it,including throughput maximization,and Quality of Service(QoS)guarantee during and after inter-cell handoffs.
The traditional RSS-based handoff algorithms is shown in Figure 2,in which vertical coordinates represent the strength of signals the mobile node receives from BS1 and BS2,horizontal coordinate represents the handoff time or the distance the mobile node moves from BS1 and BS2,and the intersections of two curves represent the same signals received from the two BSs.
(1)Based on RSS
The algorithm selects the BSwith strongest received signals.
(2)Based on RSSand Threshold Values
An inter-cell handoff takes place only when the RSSof a new BSsurpasses that of the current BS,and when signals at the current BSare lower than the threshold.The corresponding thresholds and handoff points in Figure 2 are(T1,A),(T2,B)and(T3,D).
(3)Based on RSS,Hysteresis margins and Thresholds
An inter-cellhandoff will be made if the RSSof a new BSis greater than the sum of the RSSof the current BSand the hysteresis margin(h).The corresponding handoff point in Figure 2 is C.
(4)Based on Algorithms and Dwell timers
The dwell timer is sometimes used in some other algorithms,and it will begin measuring time when it meets the requirements of the algorithms.If the timer keeps meeting those requirements for the entire time,an inter-cell handoff will be made.
In traditional handoff algorithms,the handoff makes the BS selection based on the RSS.This will cause a large number of unnecessary handoffs.If signals of the current BSare still strong enough,the threshold based RSShandoff algorithm allows a user having a weak signal(lower than a threshold)to make a handoff when there is a target BSthat has stronger signals.
However,the delay in this algorithm will decrease the quality of communication links,which may lead to call loss.Besides,the algorithm will cause interference between users on the same channel.The hysteresis margins and threshold based RSS algorithm allows a handoff when the signalstrength of a target BSis greater than that of the current BS.This algorithm avoids the Ping Pong Effect.However,similar to the threshold based RSSalgorithm,it may worsen QoSbecause of long handoff delay.Most wireless systems still use the traditional handoff algorithms because they are simple and easy to implement.
It is clear that 4G mobile communication systems are a kind of all-IPbased multi-network integrated system;however,there is no unified architecture for it yet.The systems enable flexible handoffs between different-type access networks that use different protocols such as MANET,IEEE 802.11 and HIPERLAN 2.Besides,terminals in the 4G systems can seamlessly roam and handoff between different types of access networks.A handoff between different networks is called vertical handoff.When services offered by different types of networks are complementary with each other,a verticalhandoff,as an option,is possibly triggered(even in the condition that the current link quality does not decrease).Therefore,more intelligent handoff algorithms are required for handoff decision in 4G systems.
In traditional handoffs,only signal strength and obtainable channels are measured.However,the following metrics are suggested to be used together with signal strength for handoffs in 4G systems[10-11]:
(1)Service Type
Different-type services require different combinations of reliability,delay and data rates.
(2)Monetary Cost
Monetary cost is an important factor that will influence a user's decision.Therefore,different charging strategies used by different types of networks will influence users'handoff selection.
(3)Network Conditions
Network related parameters such as communication traffic,obtainable bandwidth,delay and congestion,may also be taken into account for efficient network usage.The use of the network information for handoff selection will be helpful to balance traffic load between different networks and to mitigate congestion of a certain system.
(4)System Performance
Various parameters such as channel propagation characteristics,path loss,inter-channelinterference,Signalto Noise Ratio(SNR)and BER,are used for handoff decision in order to guarantee system performance.Moreover,battery energy is also an important factor for some users.
(5)Mobile Node Conditions
(6)Personal Preference
Personalpreference can be used to cater for special demands when two different types of systems are compared.
In integrated cellular and WLAN networks,WLAN is used to cover hot spots in the cellular network with high data rates in order to offer high-speed data services.The integration of cellular networks and WLAN creates both the integrated General Packet Radio Service(GPRS)and WLANand the Universal Mobile Telecommunications System and WLAN networks.At present,solutions to the integration focus on routing and handoffs.The handoff in the integrated cellular and WLAN networks mainly takes places in the coverage overlaps of cellular BSs and WLAN APs[12].
As for the integrated cellular and Ad hoc system,every mobile node in it can be considered as a virtual pico cell.The node is equivalent to the BSin the cell,and the radius of the pico cell is the signal propagation range of the node.Therefore,this system is regarded as a platform having micro cells and a large number of virtual"mobile"pico cells.The handoff of mobile nodes in the integrated cellular and Ad hoc network is more complicated than that in a traditional cellular network because each mobile node has two air interfaces.There are four types of handoffs in this kind of integrated systems,and they are listed as follows:
·The handoff between different cellular BSs
·The handoff from cellular links to Ad hoc relaying links
·The handoff from Ad hoc relaying links to cellular links
·The handoff between different Ad hoc relaying links
随着消费时代的到来,旅游行业的发展日新月异,人们对游玩地点的选择也逐渐从风景名胜转向原生态的自然乡村。这一转型促进了乡村经济结构的调整,也为乡村发展注入了新能量。安徽省地貌多样,生态条件良好,自然资源丰富,且68%的旅游资源在农村,发展乡村旅游具有极大优势。近年来,安徽省乡村旅游的内容逐渐拓展,模式不断创新,而同时又伴随着新问题的产生。于此本文以枞阳县周潭镇大山村为例,概览安徽省乡村旅游的发展历程与现状,分析其存在的问题,并进行对策研究。
The handoff between different BSs is the same as the hard or soft handoff in the traditional cellular system while the handoff between different Ad hoc relaying links is similar to routing algorithms in Ad hoc networks.The other two special vertical handoffs in the integrated cellular and Ad hoc system are reversible between cellular links and Ad hoc relaying links.
Theoretically,handoffs are either imperative or simply for optimization.Ahandoff is unavoidable if link quality is poor.In this case,handoff decision and execution must be very quick to maintain the established connection.Primarily,the RSSvalues measured at the APfor a connection and its neighboring APs are used for handoff decision.Besides,other metrics such as CIR,SIRand BERcan be used for the decision.The handoff used to provide users with better performance or to meet a certain preference is seen as the optimizing handoff.For example,a user may require more bandwidth to transfer data faster,or a cheaper network to reduce service cost.This kind of handoffs may tolerate longer delay,and can further be divided into the QoSrelated and the Authentication Authorization Accounting(AAA)related.
Vertical handoffs in the integrated cellular and Ad hoc networks can also be either imperative or simply for optimization.
(1)Imperative Vertical Handoff
When the single-hop or multihop path for data transmission between a mobile node and the serving BSis broken off,a handoff is indispensable.This kind of handoffs may be triggered by the failure of the serving BSor moving of the mobile node,and will take place after the current link has been interrupted and before a new link is established.Therefore,they belong to the hard handoff.
(2)Vertical Handoff for Optimization
This kind of handoffs take places when a mobile node discovers a better path during the data transmission,or when the system wants to balance traffic load.Vertical handoffs for optimization can improve communication quality of every mobile node link,and mitigate congestion of hot spots through load balancing algorithms.The handoffs can be made either before a link interrupts or after a new link is established;so,they belong to either the hard or the soft handoff.
Three hybrid vertical handoff algorithms will be introduced as follows.They are the signal strength-based,link quality-based,and load balancing-based handoff algorithms.
The first algorithm fulfills the imperative handoff while the other two are handoffs for optimization.The three handoff algorithms are applied into different architectures of integrated Ad hoc and cellular networks to improve performance of special systems.
There is possibly a Non-Line-of-Sight(NLOS)path between a mobile node and a BSin suburbs with high-density buildings.When a mobile terminal moves around a corner,there can be a sudden drop(of about 20-30 d B)in the received signal level because it loses its Line-of-Sight(LOS)component with the serving BS.This problem is the so-called Corner Effect[13].Now,if the handoff cannot be initiated fast enough,the call may be dropped.Yang et al[14]propose a position-assisted relaying and handoff algorithm for hybrid ad hoc cellular system.In such a system,the mobile node is assumed to be able to estimate its geo-location and has the ability to establish direct connections with nearby nodes to form a temporary wireless ad hoc network.The position information of the mobile terminal allows an estimate of the mobility profile of each individualmobile node.This estimate is then used to dynamically adjust the hysteresis margin to encourage or discourage handoff.The position information also allows the mobile terminal to select the best nearby node for relaying if the link quality between it and the BSfalls below a threshold level.Simulation results show significant improvement in terms of the call outage probability,active set update rate and mean active set number over the soft handover algorithm for UMTS Terrestrial Radio Access(UTRA).
In some special areas such as metro platforms,indoor environments and basements,no communication system can enable calls.Therefore,these areas are called dead spots.Aggelou and Tafazolli[3]offer a solution to this problem by using the signal threshold-based handoff algorithm.According to the comparison of the measured signal strength of a mobile node and the set threshold,the handoff for the node may be initiated between the cellular and Ad hoc networks to extend the coverage of the cellular network,and to alleviate the dead spot's impact on the system.Besides,reducing transmitting power can improve the system capacity and can mitigate the inter-cell interference.This kind of signal threshold-based handoff algorithm works in different modes as follows:
(1)The handoff between different cellular BSs will be initiated if the following conditions are met at the same time:
·The average signal strength of the serving BSgets below the threshold
·The average signal strength of a neighboring BSis greater than the threshold
(2)The handoff from a cellular link to an Ad hoc relaying link will be initiated if the following conditions are met at the same time:
·The average signal strength of the serving BSgets below the threshold
·The average signal strength of a neighboring BSdoes not surpass the threshold
·The average signal strength of a neighboring MTis greater than the threshold
▲Figure 3. Signal threshold-based handoff decision algorithms.
(3)The handoff from an Ad hoc relaying link to a cellular link will be initiated if the average signal strength of the BSnearby the MTis greater than the threshold.
(4)The handoff between different Ad hoc relaying links will be initiated if the following conditions are met at the same time:
·No BShas the average signal strength greater than the threshold
·The average signal strength of the serving relaying MT gets below the threshold
·The average signal strength of a neighboring MTis greater than the threshold
The handoff decision algorithms are shown in Figure 3,in which RSSBS_Cand RSSBS_nare the average signal strengths the MTreceives from the current serving BSand the neighboring BS respectively;RSSMT_cand RSSMT_nare the average signal
strengths the MTreceives from the current serving relay and the neighboring moving relay respectively;and T_Cellular and T_Ad hoc are the handoff thresholds in the cellular and Ad hoc networks respectively.
When a node moves within the coverage of the serving BS,multihop routing can be used to improve the system throughput.If the downlink channel quality between the BSand the node gets below a special threshold,the node will handoff into the Ad hoc system.The system discovers the agent host with the best downlink channel quality to relay the data packets to the destination node.The UCAN uses the High Data Rate(HDR)downlink channel as the handoff algorithm metric.Each mobile node in the UCAN system maintains a mobile average downlink channel rate.The state of this downlink channel can be estimated by measuring the pilot signal in the HDRdownlink.
The mobile average can filter the high-frequency changes of the channel caused by fast fading,and obtain the distance-based signal strength in the slow-fading environment.If the downlink rate of the target node gets slower than 38.6 kb/s,the node will handoff into the multihop mode to make the BStransfer the data frames to the node(agent gateway)with the best channel rate.These data frames will be further relayed to the middle node through the 802.11b link with broad bandwidth,and finally reach the target node.
Similarly,Weiand Gitlin[15]use the downlink channel quality of the mobile node to make handoff decision.A node having received bad downlink quality from the BScan get a high data rate by two-hop connectivity.Different from the UCAN,the agent gateway in this system periodically broadcasts the relaying advertising message through its WLAN interface.The message includes the Channel State Information(CSI)and bandwidth indication value,which are used to decide to establish a direct cellular connection or a two-hop-relay connection.Weiand Gitlin[15]also mention that the power consumption of the direct and two-hop-relay communications,moving mode of the terminal,and time restriction of the application can be used as decision factors for handoff decision algorithms.
Zhao and Todd[16]propose a solution to improve the capacity of the cellular Code Division Multiple Access(CDMA)by introducing out-of-band multihop relaying.Communication between the relaying node and the mobile node uses bandwidth which is available to the ad hoc radio and does not consume the CDMAcapacity.Zhao and Todd[16]introduce three relay station selection criteria,namely,Ad Hoc Relaying with Low Relative Interference(ARRI),Ad Hoc Relaying with Best Link Gain(ARLG),and Ad Hoc Relaying with Shortest Distance(ARSD).
ARRIand ARLG dynamically track the current CDMA link quality.When a mobile node finds that the current direct link has a low quality,ARLG will look for the relaying node that has best link gain with the serving BS,and then,it will initiate the handoff;ARRItakes both the link quality of the serving BSand the interference into neighboring cells into consideration;and ARSD selects the nearest relaying node to the serving BS.The emulation results show that Ad hoc relaying,which dynamically tracks CDMA link quality(ARRIand ARLG),can achieve greater capacity improvements than using a distance-based relay station selection(ARSD).Therefore,the link quality-based handoff algorithms,used for the handoff of mobile nodes from direct cellular communications into multihop relaying communications,can not only improve the system capacity and throughput,but also extend the coverage of BSs.
The BSor APis indispensable in communications in traditional cellular networks.However,due to the limited frequency resources,establishing more BSs cannot effectively solve the problem of traffic congestion in the networks.The simultaneous usage of mobile devices in one particular cell can cause the cell to be heavily congested.Such congested cells are called hot spots.There are two solutions to avoid hot spots.One is to increase the number of cells and make every cell smaller so that the number of users in every cell willbe smaller.However,this solution can be expensive because more BSs are needed to provide the coverage,and the utilization of these BSs can be low.The other solution is to introduce the Ad hoc network into the cellular network.The Ad hoc network has good performance in throughput,delay and power saving,but it is not suitable for wide area network.The integration of Ad hoc and cellular networks can take advantage of both systems to solve the problem of congestion in hot spots.
The PARCelSsystem uses the load-balancing algorithm.It relays the traffic from the current cell into the neighboring cells upon congestion but before severe congestion happens.A BS in the system periodically indicates its congestion state to all mobile nodes in its cell,and this congestion state reflects the fraction of occupied Data Channels(DCH).When a mobile node receives this congestion indication and finds that the current cell is getting congested,it should start to search for a relay route by broadcasting Route Discovery Messages(RDM).When a RDM reaches a mobile node in a cell with more free channels,a Route Trace-back Message(RTM)is generated and sent back to the searching mobile node.After finding several relay routes,the searching mobile node selects the best relay route by computing relay desirability values.The relay desirability value is a weighted function of the length of the relay route,power state of the node,and motion of the mobile node.The searching mobile node finally sends the desirable route to its BS,and the BSselects certain mobile nodes and handoffs them to neighboring cells by relaying.In this way,the BSachieves load balancing.
In iCAR,ARSes are introduced into the cellular system.When congestion happens in a cell,these ARSes can be used to transfer the traffic in the congested cell to a less congested cell.When a mobile node is entering a congested cell,a new call it initiates or its call in operation will not get blocked or dropped.Instead,iCARhandoffs the calltraffic from the cellular interface to the Ad hoc interface and uses the routing algorithm to search for an ARSto establish a route from this mobile node to the neighboring cell.The mobile node first broadcasts“route searching”message to all neighboring ARSes in order to find a relay route to BSs in less congested cells.The qualified ARSes willsend routing information back to the mobile node.The node then selects the best ARSas its agent,and sends"route establishment"message to this ARS.After that,the gateway ARSrequests the BSto allocate a Dedicated Channel(DCH),and then sends acknowledgement message back to the mobile node,establishing the relay route.Once receiving the acknowledgement message,the mobile node handoffs to the Ad hoc interface,and begins to transport data through the relay route.
Handoffs in integrated Ad hoc and cellular networks are different from those in integrated WLAN and cellular networks.The former is based on multihop while the latter is based on one-hop handoff between the BSand AP,when a user wants to access the Internet.It is necessary to consider the impact of routing algorithms on handoff in integrated Ad hoc and cellular networks.
Gateway Nodes(GN)are used as the interface between the cellular infrastructure and the MANETin most hybrid architectures.When imperative handoffs or handoffs for optimization take place,the system must give the mobile node certain methods to discover GNs to avoid service interruption.
Two basic gateway discovery approaches,the proactive and the reactive,have their own strengths and weaknesses.The proactive gateway discovery generally enables faster response,but costs more control traffic overhead.Conversely,the reactive gateway discovery can reduce control traffic overhead,but fails to obtain the same short response.The proactive approach is applied into the A-GSM system,where the GN periodically announces advertising messages.When a mobile node in the UCAN needs multihop access to a BS,it begins to search a gateway.Therefore,the UCAN uses the reactive approach.Besides,there are two different discovery protocols:one is that a mobile node traces downlink rates of its neighboring nodes to find a gateway;and the other is to send the gateway request untila candidate gateway finds the request.
▼Table 1. Classification of handoff approaches in the Ad hoc and cellular network
As shown in Table 1,allthe available handoff approaches for hybrid networks can be classified according to the above analysis of handoff algorithms and gateway discovery algorithms.Signalstrength based handoff algorithms are mainly applied to avoid dead spots and corner effect.The algorithms can extend the coverage of the cellular system and improve system throughput.However,since they are not adaptive to changeable parameters in the system such as the propagation environment,traffic density and users'moving speeds,they may cause unbalanced load between the networks,which will further lead to decreased QoS.Traffic transfer can alleviate call blocking and handoff interruption caused by traffic overload in hot spots;so,traffic transfer is one of the important reasons for introducing the Ad hoc network into cellular system.The dynamic relaying function of the Ad hoc network can effectively adjust the traffic in hot spots to implement dynamic traffic allocation,which helps improve the service ability and QoSof hot spots.The load balancing-based handoff algorithms can transfer the signals of a node in a congested cell to a less congested cell by multihop relaying so as to reach load balancing between cells.However,the algorithms consider only the whole system performance,but do not optimize the performance of individual mobile nodes.Neither do they take the impact of other handoff metrics on system performance into account.The link quality-based handoff algorithms are mainly used for high-speed downlink data services.If the data rate of a mobile node is going down,the node will handoff to the multihop transfer mode,and implement high-speed data connectivity by receiving data through a GNwith better channel quality and by using high-speed IEEE 802.11 protocols to send data to the client end.The algorithms can improve the data rates at cell edges.However,they do not consider the impact of other handoff metrics on system performance,which may worsen the system performance.Therefore,these three handoff algorithms are designed for special systems,and each algorithm only considers one handoff metric.In fact,in order to reach the best network performance,it is necessary to comprehensively consider the influences in handoff algorithms caused by all kinds of factors such as current utility of network resources,the state of network load balancing,QoSrequirements of different services and moving characteristics of users.
The special architecture of integrated Ad hoc and cellular networks requires special handoff solutions.
Traditional cellular and hybrid communications coexist in the integrated Ad hoc and cellular network.The hybrid mode means accessing to BSthrough multihop relaying.Therefore,the handoff between different communication modes is unavoidable.Few studies of handoff algorithms in integrated Ad hoc and cellular networks have been done.Besides,the available algorithms are all based on traditional handoff algorithms,which are not optimum because they cannot improve the weakness of the traditional algorithms of which system parameters are not adaptive to.The proposed hybrid architectures are generally aimed to enhance users'throughput and to improve the whole system performance,but they do not consider QoSrequirements of applications in terms of transmission mode selection,routing and handoff processing.With the development of new handoff metrics,the traditional handoff algorithms can no longer meet the requirement of multi-criteria handoffs.Implementations based on multi-criteria handoff algorithms are becoming more important,especially in hybrid networks where more interaction between networks and users increases handoff delay.Therefore,it is necessary to develop new technologies to improve the effectiveness of handoff algorithms,and balance users'satisfaction and network efficiency.Recently,several new technologies are emerging such as intelligent technologies separately based on neural networks,fuzzy logic systems and pattern recognition,and optimization technologies separately based on cost function and fuzzy multiple attributive decision.Intelligent and optimized handoff algorithms for hybrid networks that integrate Ad hoc and mobile cellular systems are adaptive and robust.They are able to implement handoffs adaptively,according to specific multi-factors such as different QoSrequirements,network states and mobile node conditions in the future hybrid networks.
Therefore,these intelligent and optimized algorithms can make more effective handover decision,and accordingly improve the system performance.The future research will tackle intelligent or optimized vertical handoff algorithms for integrated Ad hoc and mobile cellular networks to improve their whole system performance.