Mobility-dependent call admission control in hierarchical cellular networks

https://doi.org/10.1016/S0140-3664(01)00398-XGet rights and content

Abstract

Two key goals of call admission control in the next generation wireless networks are efficient use of scarce wireless resource and quality-of-service improvement. As is well known, blocking handoff calls is normally more annoying than blocking new calls, whereas blocking new calls inevitably reduces resource utilization. That is, blocking new calls impact resource efficiency resulting in trade-off with handoff priority. A novel Mobility-Dependent Call Admission Control (MDCAC) scheme is proposed to achieve the aforesaid goals in a hierarchical cellular network. With MDCAC, new calls are accepted according to some acceptance probability, taking account of mobility difference between slow and fast mobility calls and mobility change over time. An iterative algorithm is developed to calculate performance measures of interest, i.e. new call blocking probability and forced termination probability under stationary scenarios. First, simulation results are shown to verify analytical results. Then, numerical results are presented to show the robustness of MDCAC. Last but not least, different CAC schemes are compared under non-stationary scenarios via simulation, where each cell (both microcells and macrocells) alternates between busy and normal states independently of one another. It is also shown that MDCAC outperforms the other CAC schemes under study.

Introduction

With the attraction of convenience and ubiquity, the demand for wireless communications services continues to grow dramatically. In addition to conventional voice service, a wide range of applications, such as data, video and multimedia, are also expected to be provided by the next generation wireless (cellular) multiservice network [1]. As is well known, radio channels, either time slots in time division multiple access (TDMA) systems or codes in code division multiple access (CDMA) systems, are characterized as scarce resource. Thus, how to maximize the number of serving subscribers while keeping QoS at an acceptable level is a critical concern and challenge to the wireless network designer for future cellular networks. Generally, for wireless cellular systems, more system capacity can be achieved by means of a decrease in the cell size. However, such policy may result in increased number of handoff requests and the corresponding increase in signaling load. Among many possible approaches, cellular system employing hierarchical or multi-tier structure is often taken into account [2], [3], [4], [5]. In such a system, the service area can be divided into several microcells, which are overlaid by macrocells. Such an arrangement can provide higher capacity with microcells, while reducing signaling overhead due to handoff calls with macrocells. The teletraffic analysis of such hierarchical cellular networks is generally a complex problem due to the non-Poisson nature of handoff and overflow traffic. Some papers did adopt models, which are more sophisticated taking account of both the mean and the variance of the handoff and overflow traffic. It is noted that most relevant papers only presented analytical results without validating simulation results.

Some essential issues should be addressed for an efficient channel assignment in such hierarchical networks. Among other things, micro–macro cell selection problem, i.e. how to assign mobiles to microcells or macrocells has attracted a lot of interest [2], [3], [4], [5], [6], [7], [8]. All originated calls are first assigned to a microcell in Ref. [7]. Upon handoff, if the residual sojourn time of the mobile in the origination microcell is longer than a predefined threshold, the call will be handed off to the neighboring microcell. Otherwise, the call will be overflowed to the overlaid macrocell. In Ref. [8], another approach is adopted for GSM systems. A mobile enters a microcell with a specified negative offset to the received power level from microcell base station for a specified time-period. If there is a call origination prior to the expiration of the specified time-period, due to the negative offset, the mobile will originate in the macrocell. On the contrary, if the specified time-period expires, the negative power offset is removed so that calls can originate in the microcell. In [2], [4], [5], [6], a call is classified as fast or slow by the cellular system. Two strategies are compared in Ref. [6]. With strategy 1, the mobile users with different velocities are served at different layers. Contrastingly, with strategy 2, the mobile users are served at either layer, regardless of their velocities. It is revealed that the completion probability of mobile user with strategy 1 is higher than that with strategy 2. Moreover, it is also shown in Ref. [6] that the performance of strategy 1 is insensitive to imperfect speed estimation. Basically, there are two approaches to estimate user speed. One approach is to determine the user mobility from the characteristics of the received signals [9], [10], and the other one is to achieve this classification of mobile users through cell sojourn time estimation [11], [12]. Similarly, we assume that the mobile users can be divided into two groups with any appropriate speed estimator and then are served by different layers. The cell selection policy is that fast mobility calls are assigned to macrocells and slow mobility (including stationary) calls are firstly allocated to microcells, which are used in areas of high utilization.

Furthermore, handoff call handling is also an important topic in cellular networks. Generally, blocking handoff calls is more annoying for users. Thus, handoff calls are usually given priority over new calls. Basically, there are two ways to prioritize handoff calls for reducing forced termination probability: guard channel and queuing. Many variations or combinations of the above two schemes have been proposed [13], [14], [15], [16], [17]. By trunk reservation (TR) policy with fixed number of channels exclusively for handoff calls, this improvement of forced termination probability is accomplished at the expense of higher new call blocking probability. Handoff queuing is an alternative scheme by introducing handoff queue for storing temporarily unacceptable handoff attempts. It is noted that handoff calls still could be rejected due to the limited size of the handoff queue. Without any consideration of the user mobility characteristics is the common drawback of these aforesaid schemes. A dynamic channel reservation (DCR) scheme taking account of mobility parameter is proposed in Ref. [17]. The mobility parameter is defined as the ratio of handoff call arrival rate to new call arrival rate. One unrealistic aspect of DCR is that handoff call arrival rate is assumed to be independent of new call arrival rate. As is well known, handoff call arrival rate depends on new call arrival rate. Normally, the handoff arrival rate is unknown but can be derived with new call arrival rate, iteratively. Therefore, in this paper, the mobility parameter is chosen as the (more realistic) ratio of the sojourn time to the call holding time.

It is also well known that actual traffic and specifically the call arrival rate are seldom stationary because the new call arrival rate changes during the day, and the handoff arrival rates depend on both the number and movements of callers in adjacent cells. With the popularity of microcells, the actual traffic load of a given microcell may deviate significantly from its expected load more easily. As a result, performance under nonstationary traffic load is also important in its own right [18], [19], [20]. Thus, we also consider the scenarios where a cell could be either busy or normal as in Ref. [20]. A busy cell has a higher arrival rate and a longer cell sojourn time than a normal one. The robustness of our proposed scheme is also studied under such nonstationary traffic scenarios.

In this paper, a two-tier cellular network with users of two types of mobility is considered. With the mobility parameter defined as the ratio of the sojourn time to the call holding time, mobility-dependent call admission control (MDCAC) is proposed not only to maintain the forced termination probability at an acceptable level, but also to reduce the new call blocking probability and improve the channel utilization. Importantly, not only the teletraffic analysis is dealt with but also analytical results are verified with simulation results for the stationary scenarios. Further, in addition to stationary scenarios usually studied in other references, the cases with nonstationary traffic are also investigated in this paper.

The remainder of this paper is organized as follows. In Section 2, the associated system model and the proposed MDCAC are described. The analytical method is described in detail in Section 3 and numerical results for stationary and nonstationary scenarios are shown in Section 4. Some concluding remarks are given in Section 5.

Section snippets

System model

A two-tier cellular network for coverage of a particular area with mobiles moving at different velocities is considered. Every N microcells are overlaid by a macrocell. For simplicity, we consider a homogeneous cellular network, that is, all cells of the same hierarchical level are assumed statistically identical so that we can focus on one particular cell in each layer. Each macrocell (microcell) is allocated C2 (C1) channels.

Consider that there are two types of calls in this study — slow

Analytical method

In this section, an analytical method is developed to pursue the performance measures of interest in a two-tier cellular system with MDCAC. The main performance measures of interest are new slow mobility and fast mobility call blocking probabilities, Pbs and Pbf, and forced termination probability (call dropping probability) of handoff slow and fast calls, Pds and Pdf. Moreover, the weighted blocking probability Pwb is also evaluated, which a representation of system utilization. The analysis

Numerical results

A two-tier system with slow mobility and fast mobility calls is studied. Assume that every seven identical microcells are overlaid by a single macrocell. The number of channels allocated to each macrocell is 28, and so is each microcell. That is C1=C2=28. For comparison, fixed TR and MDCAC with fixed acceptance probabilities (MDCAC-FAP) are also considered. The number of reserved channels for handoff calls exclusively with TR is denoted as g. With TR, two cases are explored where g is 4 or 5.

Conclusion

A novel MDCAC is proposed to achieve both high system utilization and acceptable performance measures in hierarchical cellular networks with fast and slow mobiles, taking account of mobility difference between fast and slow mobiles. An iterative algorithm is developed to calculate performance measures of interest, i.e. new call blocking probability and forced termination probability under stationary scenarios. Analytical and simulation results are shown to be very close. Numerical results are

References (21)

  • H. Aghvami et al.

    A vision of UMTS/IMT-2000 evolution

    Electronics and Communication Engineering Journal

    (2000)
  • B. Jabbari et al.

    Teletraffic modeling and analysis of flexible hierarchical cellular networks with speed-sensitive handoff strategy

    IEEE Journal on Selected Areas in Communications

    (1997)
  • C. Mihailescu, X. Lagrange, D. Zeghlache, Analysis of a two-layer cellular mobile communication system, IEEE 47th...
  • Lon-Rong Hu, S.S. Rappaport, Personal communications systems using multiple hierarchical cellular overlays, Third...
  • Lon-Rong Hu et al.

    Personal communication systems using multiple hierarchical cellular overlays

    IEEE Journal on Selected Areas in Communications

    (1995)
  • W. Huang, V.K. Bhargava, Effects of user mobility on handoff performance in a hierarchical cellular system, Canadian...
  • W. Jolley et al.

    Modeling and analysis of layered cellular mobile networks

    (1991)
  • Unitel, Idle Mode Cell reselection for Microcells, ETSI GSM2,...
  • M.D. Austin et al.

    Velocity adaptive handoff algorithms for microcellular systems

    IEEE Transactions on Vehicular Technology

    (1994)
  • M. Hellebrandt et al.

    Estimating position and velocity of mobiles in a cellular radio network

    IEEE Transactions on Vehicular Technology

    (1997)
There are more references available in the full text version of this article.

Cited by (8)

  • A simple and scalable handoff prioritization scheme

    2005, Computer Communications
    Citation Excerpt :

    This way, disruptions of connectivity and packet losses during the handoff signaling can be avoided. Example approaches are currently discussed in the IETF [30] as an extension to the Mobile IP protocol [31] or the many micro-mobility approaches [32,33] which are intended to complement macro-mobility approaches such as Mobile IP. There are also approaches which combine such a mobility management with resource management [34].

  • Performance design and analysis of handoff in hierarchical wireless networks

    2011, Proc. 10th IEEE Int. Conf. on Trust, Security and Privacy in Computing and Communications, TrustCom 2011, 8th IEEE Int. Conf. on Embedded Software and Systems, ICESS 2011, 6th Int. Conf. on FCST 2011
  • MAC-layer scheduling based on service coefficients in heterogeneous wireless networks

    2009, Proceedings - 5th International Conference on Wireless Communications, Networking and Mobile Computing, WiCOM 2009
View all citing articles on Scopus
View full text