Quality of service estimation for soft handoff region ratio and call admission control in CDMA cellular systems☆
Introduction
Optimal handoff management is important to guarantee the quality of service (QoS) in mobile cellular networks. A soft handoff scheme is among the prominent features in code division multiple access (CDMA) cellular systems. This technique allows mobile stations (MSs) near a cell boundary to transmit and to receive from two or more base stations (BSs) simultaneously. An MS in a soft handoff is power controlled by the BS, which requires less power. Consequently, a soft handoff increases the system capacity by reducing interference. Several previous research studies have demonstrated that a soft handoff results in a larger cell coverage area and a subsequent increase in reverse link cell capacity (Gilhousen et al., 1991, Viterbi et al., 1994).
In general, the soft handoff effects are estimated on the basis of either the soft handoff region ratio (SHRR) or soft handoff percentage. The former is defined as the ratio of the area of the handoff region relative to the area of the entire cell, while the latter is defined as the average fraction of the call duration in a soft handoff (Hong, 2001, Su et al., 1996). The values of these two measures are said to be equivalent if the MSs are uniformly located in cell areas. The value of the SHRR is determined by soft handoff parameters such as T_ADD, T_DROP and T_TDROP in IS-95A and three additional parameters in IS95B/CDMA2000. If a pilot signal from a new BS is detected above T_ADD, it is added to the Active Set. Thereafter, a new link is established. If one of the pilot signals from the BSs in the Active Set is below T_DROP for T_TDROP, it is removed from the Active Set and the link is disconnected. Refer to Ahmed (2002) for the specification of soft handoff procedures. The relations of the SHRR and soft handoff parameters have been studied analytically or experimentally under various propagation environments (Avidor et al., 2006, Liu and Wang, 2003, Su et al., 1996, Zhang and Holtzman, 1998).
In terms of resource allocation, the SHRR is important in the optimization of the performance of CDMA systems (Wong & Lim, 1997). A low SHRR reduces the macrodiversity gain. In contrast, a high SHRR increases the macrodiversity gain to some extent at the expense of network resources. Therefore, the SHRR is typically set to be in a proper range of 34–45% to balance the tradeoff between the diversity gain and the network resource overhead during the network planning stage (Liu & Wang, 2003). However, the soft handoff parameters are generally set according to the condition of a full network traffic load. As such, the actual SHRR often falls outside the proper range for real traffic conditions. In the operational stage of CDMA systems, the cell capacity and the traffic load, such as a handoff call and new call arrivals, are directly influenced by the value of the SHRR. Therefore, the analysis of system performance affected by the SHRR is needed from both the standpoint of QoS management as well as resource allocation. The accurate estimation of the handoff arrival rate is important for guaranteeing the QoS because this rate becomes the load of the networks in the call level. The handoff call arrival rate is the function of various factors, such as the value of the SHRR, new call arrival rate, mobility of the MS and call admission policy.
Analytical system modeling for an accurate soft handoff is a difficult problem due to irregular cell boundaries, propagation conditions, traffic conditions and the various movement patterns of the MSs, among others. For this reason, analytical research has not been conducted to the extent of empirical or simulation-based research.
Su et al. (1996) analyzed network teletraffic performance in terms of blocking new and handoff calls by using a three-dimensional continuous time Markov chain (CTMC) under a given SHRR. However, they assumed a fixed CDMA capacity in their analysis. Narrainen and Takawira (2001) proposed a more accurate analytical model for soft handoff by taking into consideration the improvement on the CDMA capacity incorporating with the traffic load. Ma, Cao, Liu, and Trivedi (2006) constructed a CTMC model for the guard channel CAC and obtained closed-form solutions of the QoS measures, such as new call blocking probability (NBP) and the handoff dropping probability.
These previous research studies have defined the number of active MSs in the soft handoff region and the number of active MSs in the non soft handoff region as state variables of the CTMC for the estimation of handoff rate. However, this modeling fails to capture the additional soft handoff requests, which may occur in the soft handoff region. This can be explained by the model’s limited ability to count the handoff request only at the time when an MS enters the soft handoff region from a non soft handoff region. Although an MS roams within the soft handoff region, another handoff could occur if the pilot signal from another BS became stronger than one of the original pilot signals. Kim and Sung (1999) derived the number of handoff call attempts during a call holding time to obtain an accurate estimation of the handoff call arrival rate. However, they assumed the squared cell shape and did not consider the prioritized call admission control for handoff calls.
This paper assumes a hexagonal cell shape, which is more general than a squared cell shape, due to its close approximation of the circular shape. This is theoretically similar to the appearance of the BS’s coverage area. We propose a handoff arrival rate estimation model under the hexagonal cell assumption using a semi-Markov process and develop an analytical method which can investigate the effects of the SHRR and CAC policy by simultaneously taking into account the variation of the traffic load and the capacity. The numerical result illustrates the effects of various soft handoff system parameters, which are not revealed in the previous research efforts. The proposed modeling is expected to provide a practical guide to improve the QoS of CDMA soft handoff systems.
Section snippets
Cellular geometry and model assumptions
The main notations used in our model are given as follows:α soft handoff region ratio S0 (Sw) area of original (a whole extended) cell Sn (Sov) area of a normal (an overlap) region C cell capacity denoted by the number of channels C0 cell capacity in hard handoff η(α) capacity increasing factor by soft handoff TC call holding time mean call holding time RT0 (RTw) resident time in an original (whole extended) cell region RTn (RTov) resident time in a normal (overlap) region mean resident time in an
Handoff call arrival rate estimation
An MS can move through several different cells while being involved in a call. Each handoff call requires network resources to reroute the call through to a new base station. An accurate estimation of the soft handoff arrival rate is important in terms of QoS management, such as CAC, because it is directly related to the traffic load in the cell. If we define K as the number of handoff call attempts during a call holding time TC, the handoff call arrival rate per cell can be expressed as (see
Numerical results and discussion
This paper considered the operational stage of the CDMA system where the distance between BSs is already determined, and it assumed that the intersection of the cell coverage of neighboring BSs is large enough to cover the soft handoff region. The distance between the BSs is a critical factor for the QoS of MSs since the QoS is surely degraded when the BSs are located farther apart. However, the investigation of the cell coverage and capacity to determine the optimal distance between BSs is
Conclusions
A new analytical model is developed to analyze the performance of the CDMA soft handoff systems. In order to practically capture the new call and handoff call arrivals influenced by SHRR, this study employed a newly-proposed semi-Markov based traffic load estimation model under hexagonal cell shape and general cell resident time distribution. An iterative QoS calculation algorithm is presented under the prioritized call admission policy. By assuming an exponential cell resident time, the
References (17)
Simulation and performance evaluation of IS-95 handoff algorithms
Bell Labs Technical Journal
(2002)- et al.
On the impact of soft handoff threshold and maximum size of the active group on BS transmit power in the UMTS system
IEEE Transactions on Vehicular Technology
(2006) - Barcelo, F., & Jordan, J. (1997). Channel holding time distribution in cellular telephony. In Nineth international...
Modeling and performance analysis for wireless mobile networks: A new analytical approach
IEEE/ACM Transaction on Networking
(2005)- et al.
On the capacity of a cellular CDMA system
IEEE Transactions on Vehicular Technology
(1991) - Hong, S. J., & Lu, I.-T. (2001). Soft handoff algorithm parameter optimization in various propagation environments. In...
- Jedrzycki, C., & Leung, V. C. M. (1996). Probability distribution of channel holding time in cellular telephony...
- et al.
Characterization of soft handoff in CDMA systems
IEEE Transactions on Vehicular Technology
(1999)
Cited by (1)
A survey on classical teletraffic models and network planning issues for cellular telephony
2009, International Journal of Business Data Communications and Networking
- ☆
This research was financially supported by Hansung University, Rep. of Korea.