Downlink Packet Transmission Control in Soft Handoff Status on CDMA Wireless IP Networks

Transcription

1 Downlink Packet Transmission Control in Soft Handoff Status on CDMA Wireless IP Networks 125 Downlink Packet Transmission Control in Soft Handoff Status on CDMA Wireless IP Networks Abubaker Khumsi, Kazuo Mori, and Hideo Kobayashi, Non-members ABSTRACT A soft handoff scheme is very helpful to minimize the service disruption in CDMA wireless IP networks. In the conventional soft handoff scheme, the downlink throughput performance decreases due to fading fluctuations, this is because the base station transmits packets to mobile stations even in bad channel conditions. This paper presents a new transmission control scheme for the downlink IP packet transmissions where the system throughput could be efficiently improved. The proposed scheme focuses on the fact that the IP packets do not have stringent delay requirements, it aims to support the guaranteed delivery of IP packets along the network. Keywords: Soft handoff, IP packets, CDMA, cellular system 1. INTRODUCTION In cellular system, the handoff is a process that allows a mobile station s session in progress to continue without interruption when a mobile station moves from one cell to another. One of the factors affecting the Quality of Service (QoS) in the wireless IP networks is the service disruption during handoffs of the mobile station [1]. The packet loss or packet error over wireless links during handoff would cause significant throughput degradation. However, realizing an IP wireless network introduces many challenges including the soft handoff. The soft handoff allows a mobile station to communicate with multiple base stations simultaneously. Although the soft handoff is an effective way to increase channel capacity, reliability, and coverage range of the Code Division Multiple Access (CDMA) systems, but it has some disadvantages especially in downlink channels. One of the main problems is that the simultaneously multiple base stations transmission will cause an increase of the interference that affects other radio links, and consequently limits the downlink capacity [2]. To avoid this problem a soft 04PSI21: Manuscript received on January 15, 2005 ; revised on June 10, The authors are with the Department of Electrical and Electronic Engineering, Faculty of Engineering Mie University Kamihama-cho 1515, Tsu-shi, Mie, Japan. Tel: , Fax: {afakhomsi@com., kmori@, koba@}elec.mie-u.ac.jp handoff scheme which uses Site Selection Diversity Transmission Power Control (SSDT) is proposed in [3]. However, In the conventional soft handoff scheme, the base station keeps transmitting packets even in bad channel conditions which leads to the degradation of the system performance. The objective of this paper is to improve the system capacity with maintaining the minimum service disruption during handoffs of the mobile stations in downlink IP packet transmissions on CDMA wireless IP networks. This can be achieved by delaying the packet transmission from base stations to mobile stations till better channel conditions during the soft handoff mode. 2. CELLULAR SYSTEMS AND TRANS- MISSION POWER CONTROL In a DS/CDMA system, Transmission Power Control (TPC) is a vital necessity for system operation. The capacity of a DS/CDMA cellular system is interference limited since the channels are separated neither in frequency nor in time, and the cochannel interference is inherently strong. A single mobile station exceeding its required transmitted power could inhibit the communication of all other mobile stations. The TPC schemes have to compensate not only for signal strength variations due to the varying distance between base station and mobile but must also attempt to compensate for signal strength fluctuations typical for a wireless channel, these fluctuations are due to the changing propagation environment between the base station and the mobile station. In the uplink, the TPC serves to alleviate the nearfar effect. The use of the TPC is not limited to the uplink, but is also employed in the downlink. The controlled transmission maintaining the transmitted power level at the minimum acceptable level reduces the cochannel interference, which translated into an increased capacity of the system Transmission power control and handoff There are many studies regarding the downlink performance for systems applying TPC. Gejji [4] obtained a power control law that is radial distance dependent and provides uniform service to all mobile stations for the downlink transmission power control in CDMA cellular systems. This TPC method

2 126 ECTI TRANSACTIONS ON ELECTRICAL ENG., ELECTRONICS, AND COMMUNICATIONS VOL.3, NO.2 AUGUST 2005 is regarded as open-looped. Open-loop TPC mechanisms cannot overcome the fast fading effect which is arisen by multi-path fading. There are closed-loop TPC methods proposed to track and compensate the fast multi-path fading. Ariyavisitakul and Chang [5] studied a closed-loop signal to interference power ratio (SIR) based power control mechanism that the base station increases its transmitted power by an amount of unit if the received SIR is lower than a threshold value, otherwise it decreases its transmitted power by an amount of unit. Chang [6] proposes two closed-loop downlink TPC mechanisms. One is the centralized SIR balancing transmission power control method; this method is a base controlled scheme, the other is the distributed SIR based transmission power control method. In the all previous methods the effect of soft handoff has not been considered. The transmission power control together with soft handoff determines the feasibility of the DS/CDMA cellular system and is crucial to its performance. Accordingly, the multiple site transmission the straightforward realization of downlink TPC with fast power adjustment in soft handoff mode is that each active base station modifies its output power equally in accordance with TPC commands that are sent by the mobile station. The TPC commands transmitted by the mobile station requests a decrease or increase in the base station output power according to the reception quality. Though this scheme can enhance the system capacity, the required multiple site transmission for soft hand-off increases the downlink interference that will affects other radio links [7]. Moreover, the transmission power of active base stations will become imbalanced due to errors in receiving TPC commands. TPC commands reception errors cannot be avoided in practice. To avoid these problems Furukawa [3] proposed a closed-loop form of downlink transmission power control, called site selection diversity TPC (SSDT). We refer to this scheme as conventional soft handoff scheme throughout this paper. It is explained in Section SOFT HANDOFF IN DOWNLINK IP PACKET TRANSMISSION 3. 1 Conventional soft handoff scheme In the conventional soft handoff scheme SSDT, the mobile station selects three base stations from the service area which have a minimum average attenuation. These base stations are called active base stations and the base station which has the minimum average attenuation between the three selected base stations defined as the selected base station. The number of active base stations to be selected depends on the soft handoff margin (sho-m). In this selection criterion, the difference of the minimum average attenuation corresponds to each active base station should be less than the soft handoff margin. The soft handoff margin is a parameter that controls the area of the soft handoff region. The mobile station periodically chooses one of the active base stations which has the minimum instantaneous attenuation to the mobile station as a transmitting base station. The disadvantages of this method is that in bad channel conditions the larger transmitted power is required to compensate the fading. As a result, the downlink interference increases. Since the prevailing voice services in conventional cellular system, the base station tries to maintain the connection of the mobile station even in bad channel conditions, causing dramatic degradation of the system performance. 4. PROPOSED TRANSMISION CONTROL SCHEME In the proposed transmission control scheme, if the channel condition is bad due to fading fluctuations during soft handoff mode, the base station intentionally delays the packet transmission until better channel conditions are available. In other words, the base station sends the packets only when the instantaneous attenuation from the transmitting base station is less than a threshold value which is given by the following equation: Thresholdvalue = Ave atten + p [db] (1) Where p[db] is the controlled parameter in db, and Ave atten [db] is the average attenuation of the selected base station. This scheme is described in Fig.1. In the soft handoff mode, the mobile station first selects three base stations which have minimum average attenuations (active base stations). Afterwards, the mobile station periodically chooses one of the active base stations which has the minimum instantaneous attenuation to the mobile station as a transmitting base station (a) as shown in Fig.1. The transmitting base station changes as fast as TPC signal transmission period. In the case of the instantaneous attenuation greater than a threshold value (b), the base station delays the packet transmission {b}. In the case of the instantaneous attenuation less than a threshold value (c), the base station calculates the transmitted power {c}. The base station transmits the packets (d). In this scheme, during the soft handoff mode the concerned mobile station sends commands to the active base stations to choose the base station that has the minimum instantaneous attenuation for the mobile station. The corresponding base station controls its own power to keep the signal quality received in a connecting mobile station at a constant level. The output power of the other nonselected, active base stations is turned down to a minimum level (zero level in this paper). Accordingly, only one base sta-

3 Downlink Packet Transmission Control in Soft Handoff Status on CDMA Wireless IP Networks 127 (c) (b) {c} Active base stations (d) {b} Selected base station (Bs) Fig.1: (a) Mobile station (Ms) Proposed soft handoff scheme tion among the active base stations provides adequate power to the connected mobile station. By applying this proposed transmission control scheme the system throughput could be efficiently improved. The tolerable delay time would not affect IP packets transmission when these packets are not supporting real time applications. Although this scheme is intended to support applications, which do not have stringent delay requirements, it can be used for real time applications, because the proposed control scheme is limited to during soft handoff mode. The computer simulation results which compare our proposed scheme with the conventional soft handoff scheme are presented in section SYSTEM MODEL The service area consists of 27 hexagonal cells, as shown in Fig.2 with the assumption that the base stations are located at the centers of the cells and broadcast a pilot signal with constant transmission power. The mobile stations are uniformly distributed across the cells and straightly move with a constant speed in a random direction which follows a uniform distribution. The base stations transmit packets by DS/CDMA and the mobile stations receive the packets that satisfy the required SIR. The spreading sequences used in arriving packets do not collide. Radio channels suffer from propagation loss, shadowing fluctuation that has a log-normal distribution with a standard deviation of σsh [db], and fading fluctuation of which received power follows exponential distribution. In this model, the TPC technique is assumed to compensate the propagation loss, shadowing fluctuation, and fading fluctuation. TPC is assumed to be perfect without errors Signal to interference power ratio The average received power of a packet transmitted by the j th base station with transmission power of Ptx(j) and received at the i th mobile station can be expressed by the following equation: P rx (i, j) =P tx (j). 10S(i,j)/10 d(i, j) α (2) Where S(i, j) is the shadowing fluctuation in the path between the j th base station and the i th mobile station, d(i, j) is the distance between them, and α is the propagation loss coefficient. The Signal to interference power ratio SIR(i, j) can be calculated by the following equation: SIR(i, j) = G p.p rx f (i, j) (3) I intra f + I inter f Where Prx f (i, j) is the instantaneous received power from the j th base station, and G p is the processing gain. I intra f and I inter f are the instantaneous received power of intra-cell and intercell interference respectively. They are given by the following equations: I intra f = (1 Fo)(1 R A (i))p rxto f (i, j) (4) N I inter f = P rxto f (i, n) (5) n=1 n 1 Where P rxto f (i, j) is the instantaneous received power of total transmitted signal from j th base station, and N is the number of base stations within the service area. R A (i) is a ratio of transmission power for the i th mobile station to the total transmission power. Fo is an orthogonality factor defined as the fraction of total received power that will be experienced as intra-cell interference due to multi-path propagation. The Fo is 1.0 for perfect orthogonality and 0.0 for non-orthogonality. It depends on the radio channel model (i.e., number of multi-path rays) and has been evaluated in [8]. For example, Fois 0.6 for a 10 ray channel in a macrocell vehicular environment. When the received signal level is constant over the duration of a packet, the packet error rate P e (i, j) can be approximated by [9] { 0; SIR(i, j) SIRreq P e (i, j) = (6) 1; otherwise Where SIRreq is that required for a mobile station to receive packets correctly. The validity of this assumption is shown in [10]. Packets are correctly received at the mobile station when their P e (i, j) = Traffic model Base stations periodically generate packets for each mobile station at intervals which follow an exponential distribution with an average of Tint,which

4 128 ECTI TRANSACTIONS ON ELECTRICAL ENG., ELECTRONICS, AND COMMUNICATIONS VOL.3, NO.2 AUGUST 2005 Service area Table 1: Simulation parameters Fig.2: Service area is the average packet generation interval. The number of mobile stations per each cell is given by T int G/T slot. Here T slot is the slot duration and G is the offered load which is the average number of generated packets per slot duration per cell [packets/slot/cell]. After the packets are transmitted, they disappear regardless of whether they were successfully received. 5.3 Performance measures The downlink throughput is evaluated by computer simulation. The throughput is defined as Nsuc/Nslot, where Nsuc is the total number of successfully received packets at the destinations and Nslot is the total number of observed slots. The delay performance is also evaluated by computer simulation. The transmission delay is defined as the selection delay plus retransmission delay. Where the selection delay equals to the packet transmission time minus the packet generation time, and the retransmission delay is given statistically by (Ntx/Nsuc 1) Trtx,whereNtx is the number of the transmitted packets, Nsuc is the number of the successfully received packets and Trtx is the average interval of retransmission. The movement interval is defined as the interval in which we renew the locations of the mobile stations in the simulation. Other simulation parameters are shown in Table PERFORMANCE ASSESSMENT This section presents the computer simulation results to verify the effectiveness of the proposed method Throughput performance Figure 3 shows the downlink throughput performance for the proposed transmission control scheme Number of cells (N) 27 Cell radius 500 [m] Propagation loss coefficient (α) 3.5 Standard deviation of shadowing (σsh) 7.0 [db] Spreading factor (SF) 16.0 Standard deviation of TPC errors 0.0 [db] perfect Downlink orthogonality factor (F) 0.6 Required SIR (SIRreq) Slot duration (Tslot) Average of generation interval (Tint) Average interval of retransmission (Trtx) Mobile speed Movement interval 5[dB] 1.0 [ms] 25 [slots] 10 [slots] 20 [m/s] 10 [slots] Control parameter ( p) -3, -1, 0, 1 and 3 [db] Soft handoff margin (sho-m) 3, 6, 10 and 20 [db] where p [db] is -3, -1, 0, 1, and 3 db respectively, and for the conventional soft handoff scheme where the base station sends the packets regardless of the channel conditions. The soft handoff margin (shom) is set to 3 db. In this figure we observe a better throughput performance for the proposed transmission control scheme over the conventional soft handoff scheme. Figure 4 shows the downlink throughput performance with the same conditions as in Figure 3 but the soft handoff margin (sho-m) is set to 6 db. From these two figures it can be observed that the throughput performance for the proposed transmission control scheme is better than that for the conventional soft handoff scheme. Besides the best performanceisachievedwhen p [db] is equals to 1 db for both values of soft handoff margins. Figure 5 shows the maximum downlink throughput performance versus the soft handoff margins for the pro-posed transmission control scheme when p [db] equals to -3, -1, 0, 1, and 3 db. It is noticed that when the soft handoff margin (sho-m) increases the throughput performance increases because the handoff region becomes bigger and accordingly, the number of the transmitted packets in bad channel conditions is reduced Transmission delay Figure 6 illustrates the transmission delay performance for the proposed transmission control scheme when p [db] is set to -3, -1, 0, 1, and 3 db respectively and for the conventional soft handoff scheme. The soft handoff margin (sho-m) is set to 3 db. Figure 7 shows the transmission delay performance in same conditions as in Figure 6, but the soft handoff margin (sho-m) is set to 6 db. From these two figures we note that at low traffic load, the transmission delay

5 Downlink Packet Transmission Control in Soft Handoff Status on CDMA Wireless IP Networks 129 Downlink throughput [packets] Transmission delay [slots] Fig.3: Throughput performance (sho-m = 3dB) Fig.6: Transmission delay (sho-m =3dB) Downlink throughput [packets] Maximum downlink throughput Fig.4: Throughput performance (sho-m = 6dB) Soft handoff margin [sho-m] db Fig.5: Maximum throughput verses soft handoff margin (sho-m) db is a little larger in the case of the proposed transmission control scheme and smaller at higher traffic load and depending on the value of p [db]. Figure 8 shows the transmission delay performance at maximum throughput versus the soft handoff margins for the proposed transmission control scheme at p [db] equals to 0, and 1 db. It is noticed that the transmission delay is quite small, and the transmission delay is nearly constant regardless of the soft handoff margin Overall performance From the previous considerations on the basis of computer simulation results, we note that when p [db] is 1 db the throughput performance is maximum while the corresponding transmission delay is almost low for all traffic loads. The previous observations reveal that we can increase the throughput performance by delaying the IP packets transmission till better channel conditions are achieved. Moreover, at high traffic load this proposed scheme can also be efficient for IP packets which are sensitive to time delay because the transmission delay is smaller at most traffic load. 7. CONCLUSION This paper presents a new transmission control scheme to support IP packets transmission which does not have stringent delay requirements. The downlink throughput and delay performance were evaluated by computer simulations, and we observed a better throughput performance for the proposed transmission control scheme over the conventional soft handoff scheme. We noted that at low traffic load the transmission delay was a little larger in the case of the proposed transmission con-trol scheme and smaller at higher traffic load. These ob-servations reveal that the throughput performance would be increased by delaying the IP packets transmission till better channel conditions are achieved. Moreover,

6 130 ECTI TRANSACTIONS ON ELECTRICAL ENG., ELECTRONICS, AND COMMUNICATIONS VOL.3, NO.2 AUGUST 2005 Transmission delay [slots] Transmission delay at maximum throughput [slots] Fig.7: Transmission delay (sho-m = 6dB) Soft handoff margin [sho-m] db Fig.8: Transmission delay at maximum throughput versus soft handoff margin (sho-m) this new proposed transmission control scheme could also be efficient for IP packets which are sensitive to time delay because the transmission delay is smaller at most of traf-fic load. References [1] Eunsoo Shim, Hung-yu Wei, Yusun Chang, and Richard D. Gitlin, Low Latency Handoff for Wireless IP QoS with Neighbor Casting, Proc. Of ICC 2002, CD-ROM, May [2] A.J.Viterbi,A.M.Viterbi,K.S.Gilhousen,and E. Zehavi. Soft handoff extends CDMA cell coverage And increase reverse link capacity, IEEE J. Select. Areas Commun. Vol. 12, pp Oct [3] Hiroshi Furukawa, Kojiro Hamabe, and Akihisa Ushirokawa. SSDT Site Selection Diversity Transmission Power Control for CDMA Forward Link, IEEE J. Select. Areas Commun. VOL.18, NO.8, pp Agust [4] R.R.Gejji, ForwardlinkpowercontrolinCDMA cellular systems, IEEE Trans. Veh. Technol., vol.41, pp , Nov [5] S. Ariyavisitakul and L.F Chang, Signal and interference statistics of a CDMA system with feedback power control, IEEE Trans. Commun., VOL. COM-41, no.11, pp , Nov [6] Chung-Ju Chang and Fang Ching Ren, Centralized and Distributed Downlink Power Control Methods for a DS/CDMA Cellular Mobile Radio System, IEICE Trans. Commun., vol. E80-B, No.2, pp , Feb [7] M. Soleimanipour and G.H. Freeman, A realistic approach to the capacity of cellular CDMA systems, in Proc. VTC96, Apr pp [8] ETSI/SMG2, The ETSI UMTS Terrestrial Radio Access (UTRA) ITU-R RTT candidate submission, submitted to ITU-R TG 8/1 as a candidate for IMT-2000, Oct [9] K. Toshimitsu, T. Yamazato, M. Katayama, and A. Ogawa, A novel spread slotted Aloha system with channel load sensing protocol, IEEE J. Sel. Areas Commun., vol.12, no.4, pp , May [10] K. Mori, T. Kobayashi, Yamazato, and A. Ogawa, Service fairness in CDMA cellular packet systems with site diversity reception, IEICE Trans. Commun., vol.e82-b, no.12, pp , Dec Abubaker Khumsi received his B.E. degree in Electrical and Electronic engineering from Alfateh University, Libya, in 1991 and received his M.E. degree in Electrical and Electronic Engineering from Mie University, Japan, in From 1993 to 1998, he was a project engineer at Waha oil company in Libya. In 1998 he joined the higher institute of Electroncs in Libya as an assistant lecturer till the year He is currently working toward the Ph.D. degree in Systems Engineering at Mie University, Japan. His research interests include wireless IP networks. Kazuo Mori received the B.E. Degree in computer engineering from Nagoya Institute of Technology, Japan, in 1986 and received the Ph.D. degree in information electronics engineering from Nagoya university, Japan in In 1986, he joined the Hypermedia Research Center, SANYO Electric Co., Ltd. And was engaged in research and development on Telecommunication systems. From 1995 to 2000, he was a research engineer at YRP Mobile Telecomunications Key Technology Research Laboratories Co., Ltd., where he was engaged in research on mobile communication systems. Since 2000, he has been an Associate Professor of the Department of Electrical and Electronic Engineering at Mie University, Japan.

7 Downlink Packet Transmission Control in Soft Handoff Status on CDMA Wireless IP Networks 131 His research interests include mobile communication systems, CDMA schemes, radio packet communications, and teletraffic evaluation. Dr. Mori received the excellent Paper Award from IEICE, Japan in Hideo Kobayashi received the B.E., M.E., and Dr.E. degrees in 1975, 1977 and 1989, respectively from Tohoku University. He joined KDD in 1977, and engaged in research on digital fixed satellite and mobile satellite communication systems. From 1998 to 1990, he was with INMARSAT as a Technical Staff and involved in the development of future INMARSAT systems. Since 1998 he has been a Professor of Mie University. His current research interests include mobile communications and wireless LAN systems. Dr. Kobayashi is a member of IEEE.