MIMO Based Radio Resource Management for UMTS Multicast Broadcast Multimedia Services

Wireless Personal Communications (2007) 42:225–246 DOI 10.1007/s11277-006-9175-x

c Springer 2006


MIMO Based Radio Resource Management for UMTS Multicast Broadcast Multimedia Services ARMANDO SOARES2 , JOÃO C. SILVA1 , NUNO SOUTO1,2 , FILIPE LEITÃO2 and AMÉRICO CORREIA1,2 1

Instituto de Telecomunicações, Av. Rovisco Pais 1, 1049-001 Lisboa, Portugal Email: [email protected] 2 Associação para o Desenvolvimento das Telecomunicações e Técnicas de Informática, Av. das Forças Armadas, Edifício ISCTE, 1600-082 Lisboa, Portugal Email: [email protected]

Abstract. MIMO systems promise high performance gains over conventional single antenna systems. To accomplish high data rates over wireless links the use of multiple transmit and receive antennas is an alternative that does not require any extra bandwidth. It has the potential to address the unprecedented demand for wireless services, particularly for the Multimedia Broadcast Multicast Service (MBMS). This service supports downlink streaming and download-and-play type services to large groups of users. From the radio perspective, MBMS includes pointto-point (PtP) and point-to-multipoint (PtM) modes. New proposed MBMS related enhancements in the physical layer specifications and their effects on resource requirements are presented in this paper, such as the use of High Speed Downlink Packet Access (HSDPA) to broadcast/multicast video streaming using the Deficit Round Robin (DRR) scheduler algorithm as a way to maintain an acceptable QoS for all the users. Other improvements, based on the use of MIMO including macro-diversity, are independent of the mode used and are dependent on deployment scenarios, but can yield substantial reduction in resource demand. Keywords: MIMO, MBMS, Macro-diversity combining, QoS

1. Introduction There is still a lot of investigation in ways to improve the delivery of multimedia information. The multimedia paradigm has put pressure in resources optimization, and sharing channels is one of the most important aspects in network optimization. Efficient network resources usage should be the leverage for near-coming multimedia applications. Besides that, in order to guarantee scalability, enhancement schemes have to be considered in UMTS environments. Multimedia Broadcast and Multicast Services (MBMS), introduced by 3GPP in Release 6 is intended to efficiently use network/radio resources (by transmitting data over a common radio channel), both in the core network and, most importantly, in the air interface of UTRAN (UMTS Terrestrial Radio Access Network), where the bottleneck is placed to a large group of users. MBMS is targeting high (variable) bit rate services over a common channel. One of the most important properties of MBMS is resource sharing among many User Equipments (UEs), meaning that many users should be able to listen to the same MBMS channel at the same time. So, much power should be allocated to this MBMS channels for arbitrary UEs in the cell to receive the MBMS service. A flexible common channel, suitable for Point-to-Multipoint (PtM) transmissions is already available, namely, the Forward Access Channel (FACH), which is mapped onto the Secondary Common Control Physical Channel (S-CCPCH).

226

A. Soares et al.

In [1] it was shown that about 40% of the sector total power has to be allocated to a single 64 Kbps MBMS if full cell coverage is required. This makes MBMS too expensive since the overall system capacity is limited by the power resource. The HSDPA mode [2, 3] has been recently standardized for UMTS (based on W-CDMA), which provides more than 10 Mbps on a 5 MHz carrier, but only for best effort packet data services in the downlink. QPSK and 16 QAM modulation schemes are used in an adaptive manner depending on the channel state. The HSDPA as a mean to deliver MBMS video streaming will be studied with a suitable packet scheduler algorithm that tries to guarantee the same bit rate to the users in order to offer a good fairness and capacity. MIMO systems promise high performance gains over conventional single antenna systems. To accomplish high data rates over wireless links the use of multiple transmit and receive antennas is an alternative that does not require any extra bandwidth at all. Also in this paper we will analyse several MIMO based radio resource management techniques to provide MBMS and guarantee the optimal distribution of QoS depending on the location of mobiles. In Section 2 the HSDPA mode and the Deficit Round Robin packet scheduling model are presented, and MIMO Blast receivers are introduced in Section 3. Section 4 details macrodiversity combining techniques and simulation results are presented in Section 5. Finally some conclusions are drawn in Section 6. 2. High Speed Downlink Packet Access (HSDPA) Up to today no special transport channel for the purpose of multicast has been specified, but some proposal and preliminary studies have been provided. Therefore the driving concept to support multicast on the UMTS Terrestrial Radio Access Network (UTRAN) is to use the existing transport channels, with minor modifications. A flexible shared channel, suitable for Point-to-Point (PtP) transmissions is already available, namely the High Speed Downlink Shared Channel (HS-DSCH), which is mapped onto the Physical HS-DSCH. HSDPA supports new features that rely on, and are tightly coupled to, the rapid adaptation of transmission parameters to instantaneous radio conditions. These features are: • Fast Link Adaptation: Instead of compensating the variations of downlink radio conditions by means of power control, the transmitted power is kept constant and the modulation and coding of the transport block is chosen every Transmission Time Interval (HSDPA TTI2 ms) for each user, this is called Adaptive Modulation and Coding (AMC). To users in good conditions, 16-QAM can be allocated to maximize throughput, while users in bad conditions are penalized on throughput, reaching a point to which the service can be denied. • Fast Channel-Dependent Scheduling: The scheduler determines to which terminal the shared channel transmission should be directed at any given moment. The term channel-dependent scheduling signifies that the scheduler considers instantaneous radio-channel conditions. This greatly increases capacity and makes better use of resources. The basic idea is to exploit short-term variations in radio conditions by transmitting to terminals with favourable instantaneous channel conditions. • Fast Hybrid-ARQ with soft-combining: The terminal (user equipment, UE), can rapidly request retransmission of erroneously received data, substantially reducing delay and increasing capacity (compared to 3GPP Release 99). Prior to decoding, the terminal combines information from the original transmission with that of later retransmissions. This practice called soft-combining, increases capacity and robustness.

MIMO Based Radio Resource Management for UMTS

227

Streaming is one of the most expected MBMS service. The QoS constraint, for such service, is often defined by the maximum tolerable delay, which directly translates into the play-out buffer size at the mobile receiver. In our study, a packet is fragmented into frames of varying sizes (due to adaptive coding). If a complete packet cannot be successfully transmitted within the maximum delay, the transmitter discards the remaining frames and advances to the start of next packet in the buffer. Note that the maximum delay can be translated into a maximum buffer size for the given MBMS session at both the transmitter and the receiver. A scheduling algorithm, in general, is designed to reduce the system resource utilization required to satisfy the QoS constraints of simultaneous data sessions. When multiple MBMS sessions are active or in the presence of unicast services the appropriate metrics to consider are packet loss rate, packet delay and resource utilization. A typical feature of streaming applications is that they do not require as strict and small delay bounds as conversational applications do. The use of a receiver buffer makes a streaming application resistant against latencies and jitter to an extent that depends on the initial buffering delay which may be several seconds in many cases. On the receiver side, the client application stores the video frames in a buffer until it is time to show them. Typically, there is an initial buffering delay before the first frame is shown. Subsequent frames are taken from the buffer at a given codec frame rate. Re-buffering might be initiated if the receiver buffer runs empty due to extensive transport delays. In order to avoid underflows and overflows, which will result in an increased delay and increased packet loss rate, respectively, the scheduler must have the information of the state of the receiver’s buffers. 2.1. H S D PA D e ficit Round Robin Scheduler For streaming connections, the role of the packet scheduler is to provide the users with the amount of transmissions necessary to guarantee their bit rate requirements. The concept used in this paper is based on the Deficit Round Robin (DRR) scheduler used in wired networks [4] and adapted to EGPRS systems [5]. The HS-DSCH characteristics, like the fast variation of the capacity due to bit errors, fast fading, slow fading and interference, makes it necessary to modify the DRR algorithm. The main objective of this algorithm is to provide the same bit rate to all allocated users. This is done using bit counters. These counters track the different data flows, and serve those flow’s that need data transfer most urgently. Once all the throughput guarantees are fulfilled, the remaining capacity is distributed among the different allocated flows. Figure 1 describes the DRR process with two active users, considering that per Transmission Time Interval (TTI) the transmission is done to only one user. In the beginning of the simulation all bit counters are initialized with zero. The first transmission is done to user ‘A’, the amount of bits received by user ‘A’ is translated in a deficit for user ‘B’. In the next TTI the scheduler will choose the user with the bit counter with the higher value. In this case the transmission is done to user ‘B’. 3. MIMO-BLAST Receivers for WCDMA Systems Multiple Input Multiple Output (MIMO) schemes (presented in Figure 2) are used in order to push the capacity and throughput limits as high as possible without an increase in spectrum

228

A. Soares et al.

Figure 1. DRR scheduler scheme.

Figure 2. MIMO schemes.

bandwidth, although there is an obvious increase in complexity. The capacity limit of any DS-CDMA system is taken to be the resulting throughput obtained via the usage of the maximum number of codes. The codes are usually orthogonal to each other, so they are linearly independent. If any other code was to be used, it would be a linear combination of the other existing codes, and thus its content couldn’t be separable from the rest. However, if multiple transmit and receive antennas are employed, the capacity may be raised due to code re-usage across transmit antennas. If there are a sufficient number of receive antennas, it is possible to resolve all messages, as long as the channel correlation between antennas isn’t too high. For M transmit and N receive antennas, we have the capacity equation [6, 7],

 β  HH b/s/Hz (1) C E P = log2 det (I N ) + M where I N is an identity matrix of dimension N × N, H is the channel matrix, H is the transpose-conjugate of H and β is the SNR at any receive antenna. Foschini [7] and Telatar [8] both demonstrated that the capacity grows linearly with m = min(M, N ), for uncorrelated channels. Therefore, it is possible to augment the capacity/throughput by any factor, depending on the number of M and N antennas. The downside to this is the receiver complexity, sensitivity

MIMO Based Radio Resource Management for UMTS

229

Figure 3. MIMO receiver scheme.

to interference and correlation between antennas, which is more significant as the antennas are closer together. For a UMTS system, it is inadequate to consider more than 2 or 4 antennas at the UE/mobile receiver. Note that, unlike in CDMA where user’s signatures are quasi-orthogonal by design, the separability of the MIMO channel relies on the presence of rich multipath which is needed to make the channel spatially selective. Therefore, MIMO can be said to effectively exploit multipath. The receiver for such a scheme is obviously complex due to the number of antennas, users and multipath components. Different receivers were analyzed in [9, 10], in order to establish the tradeoff between performance and complexity for such systems. Two strategies were discussed; the Equalization and MRC (Maximum Ratio Combining) Based receivers. The main difference between both strategies is the fact that the equalization receivers operate on the whole block at once whereas MRC receivers work on tap/finger level, combining the taps later to form an estimate for the symbols. Figure 3 illustrates the main blocks from which the receiver is compiled.

4. Macro Diversity Combining Macro Diversity Combining (MDC) is proposed as an enhancement to the UMTS 3GPP Release 6 MBMS. In a point-to-multipoint (PtM) MBMS service the transmitted content is expected to be network specific rather than cell specific, i.e. the same content is expected to be multicast/broadcasted through the entire network or its large part. Therefore, a natural way of improving the physical layer performance is to take advantage of macro diversity. On the network side, this means ensuring sufficient time synchronization of identical MBMS transmissions in different cells; on the mobile station side, this means the capability to receive and decode the same content from multiple transmitters simultaneously. Basically the diversity combining concept consists of receiving redundantly the same information bearing signal over two or more fading channels, and combine these multiple replicas at the receiver in order to increase the overall received SNR. In macro diversity the received signals from different paths need to be processed by some sort of combining algorithm. In this study two different combining procedures are presented, namely Selective Combining (SC) and Maximal Ratio Combining (MRC).

230

A. Soares et al.

Figure 4. Selective combining.

4.1. S e l e c t i v e C o mbining Figure 4 shows a scheme of how selective combining operates at the receiver side. With SC the path/branch yielding the highest SNR is always selected. In order to guarantee that the receiver uses the path with the best quality a simultaneous and continuous monitoring of all diversity paths is required. The output of the diversity combiner will be: y(t) = gk .sm (t) + nk (t), with gk = max{|g1 |, . . ., |gN |}

(2)

Where gk is the maximum amplitude of the fading coefficients, and nk (t) is the Additive Gaussian White Noise (AGWN) which is independent from branch to branch. 4.2. M a x i m a l R at i o C o mbining The Maximal ratio combining (Figure 5), although being the most complex combining technique presented, is the optimum way to combine the information from the different paths/branches. The receiver corrects the phase rotation (and the fading amplitude, posteriously) caused by a fading channel and then combines the received signals of different paths proportionally to the strength of each path. Since each path undergoes different attenuations, combining them with different weights yields an optimum solution under an AWGN channel. The output of the receiver can be represented as: y(t) =

N 

|gj |2 sm (t) + nj (t)

(3)

j=1

5. Simulation Results Typically, radio network simulations can be classified as either link level (radio link between the base station and the user terminal) or radio network subsystem system level. A single

MIMO Based Radio Resource Management for UMTS

231

Figure 5. Maximal ratio combining.

Figure 6. HSDPA BLER vs. Eb/No without STTD.

approach would be preferable, but the complexity of such simulator – including everything from transmitted waveforms to multi-cell network – is far too high for the required simulation resolutions and simulation time. Therefore, separate link and system level approaches are needed. The link level simulator is needed for the system simulator to build a receiver model that can predict the receiver BLER/BER performance, taking into account channel estimation, interleaving, modulation, receiver structure and decoding. The system level simulator is needed to model a system with a large number of mobiles and base stations, and algorithms operating in such a system.

232

A. Soares et al.

Figure 7. HSDPA BLER vs. Eb/No with STTD.

5.1. S y s t e m L e v e l M o dels 5.1.1. Channel Model The channel model considers three types of losses: distance loss, shadowing loss and multi-path fading loss. The model parameters depend on the environment. For the distance loss it was used the COST-Walfisch-Ikegami-Model, LOS and NLOS, from COST 231 project. Shadowing is due to the existence of large obstacles like buildings and the movement of UEs in and out of the shadows. This is modelled through a process with a lognormal distribution and a correlation distance. The multi-path fading in the system level simulator corresponds to the 3GPP channel model, where the Vehicular A (3 km/h) environment was chosen as reference. These models are also used in the Link level simulator. In the radio network subsystem (RNS) system level simulator only the resulting fading loss of the channel model, expressed in dB, is taken into account. The fading model is provided by the link level simulator through a trace of fading values (in dB), one per TTI. For each environment where the mobile speed is the same several series of fading values are provided for each pair of antenna. The considered interference is the sum of intra-cell and inter-cell interference. Both have a noise-like character. This is mainly due to the large number of sources adding to the signal, which are similar in signal strength.

5.1.2. Traffic Model The traffic model for video broadcast/multicast used is based on the statistical analysis of MPEG4 traces presented in [11]. One example is presented in Table 1 based on the Star Wars

MIMO Based Radio Resource Management for UMTS

233

Table 1. MPEG4 Trace values QoS

High Medium Low

Compr. Ratio

Frame Size

Bit Rate

YUV:MP4

Mean

CoV [kbytes]

Peak/Mean

Mean [Kbps]

Peak [Mbps]

27.62 97.83 142.52

1.4 0.39 0.27

0.66 1.17 1.68

6.81 12.1 17.57

280 80 53

1.9 0.94 0.94

Table 2. HSDPA Simulation parameters Parameters

Value

Number of NodeB’s Cell to cell distance NodeB antenna gain + cable loss Propagation model Std of shadow fading Correlation between sites for slow fading Multipath fading NodeB total transmit power (sector) HSDPA used power Maximum used OVSF codes CQI’s Simulation Time Maximum UE Buffer size MBMS sessions Packet Schedulers algorithms HARQ Type No of retransmissions Link adaptation algorithm

10, with 3 sectors each 1000 m 14.5 dBi COST231 Hata-model 7 dB 0.5 3GPP VehicularA 3 km/h 43 dBm Variable 2 CQI 1 – 9 100 s 3s 1 Round Robin, Maximum CIR, DRR Chase Combining 1 Normal AMC

episode IV traces. In this document the simulations have only a low QoS video trace and one active MBMS session.

5.1.3. HSDPA results Table 2 shows the common parameters used in the HSDPA system level simulations. The geometry factor G is defined as the ratio of interference generated in the own cell to the interference generated in the other cells plus thermal noise. We know (see [12]) that for the studied macro-cellular scenario of Table 2 about 95% of the users experience a geometry factor of −6 dB or better, 80% experience a geometry of −3 dB or better and about 62% of the users experience a geometry of 0 dB or better. For HS-DSCH, user service and terminal equipment, Block Error Rate (BLER) curves were drawn for the various values of Eb/N0, so that the required amount of power for the required BLER level is obtained. Figure 1 shows the Vehicular A (3 km/h) BLER curves for

234

A. Soares et al. Table 3. HSDPA w/wo STTD, BLER Target = 0.01, Ec/Ior, for different CQI CQI

Bitrate [Kbps]

Eb/No (Single Antenna) [dB]

Ec/Ior (Single Antenna) [dB]

Eb/No (STTD) [dB]

Ec/Ior (STTD) [dB]

1 2 3 4 5 6 7 8 9

68.5 86.5 116.5 158.5 188.5 230.5 325 396 465.5

8.6 8.4 8.2 8.6 9.0 9.2 8.2 8.4 8.6

−4.74 −3.92 −2.83 −1.11 0.0 – – – –

8.1 8.2 7.6 7.2 7.4 7.7 6.6 6.9 7.2

−5.24 −3.72 −3.44 −2.51 −1.60 −0.37 0.00 – –

the Channel Quality Indicators (CQIs) CQI 1 to CQI 9 (QPSK modulated), all obtained with the link level simulator. Depending on the service, reference Eb/No values can be obtained in order to find the Signal to Interference Ratio (SIR) targets for each CQI, in this case the Eb/No value is obtained for a BLER of 1%, for real time services, and 10% for non-real time services. It’s also presented in Figure 2 the BLER curves for HS-DSCH with Space-Time Transmit Diversity (STTD). Table 3 shows the required Ec/Ior values (fraction of the total transmitted power from the base station) for each CQI (Figures 6 and 7). We can conclude from the link level results of Table 3 that the required fraction of total transmitted power from the base station (Ec/Ior) to assure the reference BLER=10−2 to 95% of all users (G = −6 dB) is excessive (notice that −1 dB = 0.8) for increasing bit rates even when using transmit diversity. Next the packet scheduler’s algorithms performance for video streaming broadcast is analyzed. The analysis, run for each scheduler, consists of measuring jitter performance, (by jitter we refer to the delay between frames, throughput and underflow probability). Let’s start by analyzing the throughput; Figure 8 shows the cumulative distribution function of the throughput for each scheduler when simulating 150 users (5/sector). The packet scheduler algorithm with the highest throughput is the Max C/I, due to its characteristic of transmit only to users in good channel conditions. The DRR scheduler has the lower throughput but at the same time it’s the most balanced. In terms of jitter (Figure 9), it’s noticeable that the Round Robin scheduler has the worst behavior, having more then 15% of the times a jitter bigger than 100 ms (which is the maximum frame delay allowed by this type of service). The DRR scheduler has less than 5% a delay higher than 100 ms. The jitter, among other factors, has a great impact on the underflow rate which is of great importance on the QoS perceived by users. On the receiver side, the client application stores the video frames in a buffer until it is time to show them. Typically, there is an initial buffering delay before the first frame is shown. Subsequent frames are taken from the buffer at a given codec frame rate. Re-buffering might be initiated if the receiver buffer runs empty (underflow) due to extensive transport delays. Figure 10 illustrates the number of occurring n underflows per user, in the simulation time interval, for each scheduler.

MIMO Based Radio Resource Management for UMTS

Figure 8. Schedulers throughput.

Probability of occurring n underflows

Figure 9. Schedulers frame jitter.

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

DRR RR Max C/I

1

2

3

4

5

n underflows

Figure 10. All scheduler’s probability of n underflows per user.

235

A. Soares et al. Probability of Occurimg 1 Underflow

236

0.25 0.2 0.15 0.1 0.05 0

5

6

7

8

Users/Sector

Figure 11. Probability of 1 underflow per user (33% of total transmit power).

The round robin fails again having a probability of 100% of occurring one underflow. In this situation the scheduler with a better underflow rate it’s the DRR having 0.6% probability of occurring one underflow. The Max C/I scheduler also shows that it’s unsuitable for this type of service. Next we will increase the number of users per sector and monitor the DRR behavior when using 33% and 66% of HSDPA allocated transmit power. Figure 10 shows the probability of occurring 1 underflow versus the number of users per sector. It can be seen that when using 33% of the total base station transmit power the DRR gives an acceptable when scheduling 6 users per sector, maintaining the probability of occurring 1 underflow below 5%. Figure 11 shows an increase in the system performance offering a good QoS when scheduling 10 users per sector, when using 66% of the total transmit power. We can conclude that even with DRR, only 10 users per sector are offered good QoS (throughput, jitter and delay). This means that HSDPA is the optimum choice for the pointto-point (PtP) mode of MBMS. For an increasing number of users, a better approach should be addressed, namely, MIMO systems with/without macro-diversity (Figure 12).

5.2. M I M O r e s u lt s Table 4 presents parameters which are commonly used in the subsequent sections. Results are presented in terms of Ec/Ior [dB] representing the fraction of cell transmit power necessary to achieve the corresponding BLER performance graduated on the vertical axis. The receiver for both SISO (Single Input Single Output) and MIMO is the same described in [9, 10]. For the reference value of BLER = 10−2 the difference of total transmitted power between the SISO (1 × 1) and MIMO (2 × 2) schemes is less than 1 dB (VehA) or equal to 1 dB (PedB). Considering that we are transmitting at 128 Kbps with SISO and 256 Kbps with MIMO we would expect the double of the transmitted power for MIMO. However we can conclude that the MIMO scheme is much more power efficient than SISO for BLER< 10−2 (Figure 13).

MIMO Based Radio Resource Management for UMTS

237

Table 4. Link level simulation parameters Parameters

Value

S-CCPCH Slot format Transport Block Size & number of transport blocks per TTI

12 (128 Kbps) Varied according to information bit rate (128 or 256 Kbps) and TTI value 16 bits 20 ms −10 dB (10%) −15 dB (3%) −15 dB (3%) Varied Used to sum the total Tx Ec/Ior to 0 dB (100%) Enabled Disabled Pedestrian B, 3 km/h Vehicular A, 3 km/h

CRC Transmission Time Interval (TTI) CPICH Ec/Ior P-SCH Ec/Ior S-SCH Ec/Ior Tx Ec/Ior OCNS Channel Estimation Power Control Channels

5.3. P e r f o r m a n c e u s i n g M a c r o D i v e rs ity Combining 5.3.1. Selective Combining and Maximal Ratio Combining Selective Combining (SC) and Maximal Ratio Combining (MRC) are enhancements for Release 6 PtM MBMS. In this section, the system level performance of selective combining and maximal ratio combining for MBMS is presented. The cross-correlation coefficients between antennas at Node B and UE were taken from 3GPP [12]. In Figures 14–18, the 1% BLER coverage vs MBMS channel power (Node B Tx Ec/Ior) with Selective Combining and Maximal Ratio Combining over 1, 2 and 3 radio links (RLs) are shown for various path models and TTI=20 ms. In Figure 14, the performance of the conventional one radio link (RL) reception is illustrated for comparison. For the reference average coverage between 90% and 95% the difference of required Ec/Ior between SISO (1 × 1) and MIMO (2×2) is between 5% and 10%. As expected the average coverage for the SISO scheme is always slightly better. However, this difference tends to decrease as the number of radio links increases. In Figures 19–23, the 1% BLER throughput vs. MBMS channel power (Node B Tx Ec/Ior) with Selective Combining and Maximal Ratio Combining over 1, 2 and 3 radio links are shown for various path models and TTI=20 ms. In Figure 19, the performance of the conventional one radio link (RL) reception is illustrated for comparison. To achieve the reference throughput between 90% and 95% of the total which is 256 Kbps for MIMO and 128 Kbps for SISO, the difference of required Ec/Ior between SISO (1 × 1) and MIMO (2 × 2) is less or equal to 5%. As expected the reference throughput is achieved with less Ec/Ior for the SISO scheme. However, this difference tends to decrease as the number of radio links increases.

A. Soares et al. Probability of Occurimg 1 Underflow

238

0.07 0.06 0.05 0.04 0.03 0.02 0.01 0

8

9 10 Users/Sector

11

Figure 12. Probability of 1 underflow per user (66% of total transmit power).

100

BLER

10–1

10–2

10–3

TXant=1, RXant=1, VehA TXant=1, RXant=1, PedB TXant=2, RXant=2, VehA TXant=2, RXant=2, PedB

10–4 –12

–11 –10

–9

–8

–7 –6 –5 Ec/Ior (dB)

–4

–3

–2

–1

0

Figure 13. BLER vs. Tx Power for SISO (1 × 1) and MIMO (2 × 2), the geometry is G = −3 dB.

6. Conclusions In this paper we analysed several effective radio resource management techniques to provide MBMS, through the use of HSDPA. HSDPA was not originally proposed to broadcasting and multicasting transmissions, but due to its flexibility we checked its capability to provide the point-to-point MBMS mode, MIMO and macro-diversity to guarantee the optimal distribution of QoS independently of the MBMS mode.

MIMO Based Radio Resource Management for UMTS

239

100 90

Average Coverage [%]

80 70 60 50 40 30 VehA, 1x1 (1RL)

20

PedB, 1x1 (1RL) VehA, 2x2 (1RL)

10

PedB, 2x2 (1RL)

0 0

10

20

30

40

50

60

70

80

90

100

S-CCPCH Ec/Ior [%]

Figure 14. Average coverage vs. Tx. power (1RL).

100 90

Average Coverage [%]

80 70 60 50 40 30 VehA, 1x1 (2RL - SC) PedB, 1x1 (2RL - SC) VehA, 2x2 (2RL - SC) PedB, 2x2 (2RL - SC)

20 10 0 0

10

20

30

40

50

60

70

80

90

100

S-CCPCH Ec/Ior [%]

Figure 15. Average Coverage vs. Tx. Power (2RL - SC).

The packet scheduler, Deficit Round Robin (DRR), which main objective is to provide the same bit rate to all allocated users was analysed. We could conclude that the DRR only assure a suitable multimedia broadcast multicast service up to 10 users per sector for the

240

A. Soares et al. 100 90

Average Coverage [%]

80 70 60 50 40 30 VehA, 1x1 (3RL - SC) PedB, 1x1 (3RL - SC) VehA, 2x2 (3RL - SC) PedB, 2x2 (3RL - SC)

20 10 0 0

10

20

30

40

50

60

70

80

90

100

S-CCPCH Ec/Ior [%]

Figure 16. Average coverage vs. Tx. power (3RL - SC). 100 90

Average Coverage [%]

80 70 60 50 40 30

VehA, 1x1 (2RL - MRC) PedB, 1x1 (2RL - MRC) VehA, 2x2 (2RL - MRC) PedB, 2x2 (2RL - MRC)

20 10 0 0

10

20

30

40 50 60 S-CCPCH Ec/Ior [%]

70

80

90

100

Figure 17. Average coverage vs. Tx. power (2RL - MRC).

urban macro-cellular environment. For more users per sector the user capacity decreases substantially with increasing delays resulting in unsatisfactory quality of service for a real time service. According to these results the DRR should be implemented as an additional point to point repair mechanism, for recovering lost packets, or to transmit an enhanced video layer instead of a stand alone operation to provide real time MBMS (point-to-multipoint mode).

MIMO Based Radio Resource Management for UMTS

241

100 90

Average Coverage [%]

80 70 60 50 40 30 VehA, 1x1 (3RL - MRC) PedB, 1x1 (3RL - MRC) VehA, 2x2 (3RL - MRC) PedB, 2x2 (3RL - MRC)

20 10 0 0

10

20

30

40

50

60

70

80

90

100

S-CCPCH Ec/Ior [%]

Figure 18. Average coverage vs. Tx. power (3RL - MRC).

256

Average Throughput [kbps]

224 192 160 128 96 64

VehA, 1x1 (1RL) PedB, 1x1 (1RL) VehA, 2x2 (1RL) PedB, 2x2 (1RL)

32 0 0

10

20

30

40

50

60

70

80

90

100

S-CCPCH Ec/Ior [%]

Figure 19. Average throughput vs. Tx. power (1RL).

We also presented the expected capacity gains that MIMO schemes with more complex receivers can provide to reduce the PtM MBMS transmission power. MIMO schemes are much more power efficient than SISO scheme for BLER < 10−2 . MIMO-BLAST receiversshould

242

A. Soares et al. 256

Average Throughput [kbps]

224 192 160 128 96 64

VehA, 1x1 (2RL- SC) PedB, 1x1 (2RL- SC) VehA, 2x2 (2RL- SC) VehA, 2x2 (2RL- SC)

32 0 0

10

20

30

40

50

60

70

80

90

100

S-CCPCH Ec/Ior [%]

Figure 20. Average throughput vs. Tx. power (2RL - SC).

256

Average Throughput [kbps]

224 192 160 128 96 64

VehA, 1x1 (3RL- SC) PedB, 1x1 (3RL- SC) VehA, 2x2 (3RL- SC) PedB, 2x2 (3RL- SC)

32 0 0

10

20

30

40 50 60 70 S-CCPCH Ec/Ior [%]

80

90

100

Figure 21. Average throughput vs. Tx. power (3RL - SC).

be built in the near future with or without the macro diversity combining already specified by 3GPP, as an effective mean to increase not only the throughput but also the number of simultaneous simulcast services.

MIMO Based Radio Resource Management for UMTS

243

256

Average Throughput [kbps]

224 192 160 128 96 64

VehA, 1x1 (2RL - MRC) PedB, 1x1 (2RL - MRC) VehA, 2x2 (2RL - MRC) PedB, 2x2 (2RL - MRC)

32 0

0

10

20

30

40 50 60 70 S-CCPCH Ec/Ior [%]

80

90

100

Figure 22. Average throughput vs. Tx. power (2RL - MRC).

256

Average Throughput [kbps]

224 192 160 128 96 64

VehA, 1x1 (3RL - MRC) PedB, 1x1 (3RL - MRC) VehA, 2x2 (3RL - MRC) PedB, 2x2 (3RL - MRC)

32 0 0

10

20

30

40

50

60

70

80

90

100

S-CCPCH Ec/Ior [%]

Figure 23. Average throughput vs. Tx. power (3RL - MRC).

Acknowledgements This work has been partially funded by the EU project B-BONE (Broadband and Multicasting Over Enhanced UMTS Mobile Broadband Networks - IST) [13].

244

A. Soares et al.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

3GPP, 25.803, “S-CCPCH Performance for MBMS”. 3GPP TS 25.308 V5.4.0, “High Speed Downlink Packet Access (HSDPA) Stage 2 - Release 5”, 2002–10. 3GPP TS 25.308 V5.4.0, “High Speed Downlink Packet Access (HSDPA) Stage 2 - Release 6”, 2004–03. M. Shreedhar and G. Varghese, “Efficient Fair Queuing Using Deficit Round-Robin”, IEEE Trans. Netw., Vol. 4, No. 3. 1996. D. Fernández and H. Montes, “An Enhanced Quality Of Service Method for Guaranteed Bitrate Services over Shared Channels in EGPRS Systems”, IEEE Veh. Tech. Conf., Spring 2002, 2/957.961 G.J. Foschini and M. J. Gans, “On limits of wireless communications in a fading environment when using multiple antennas”, Wireless Pers. Commun., Vol. 6, pp. 311–335, Mar. 1998. E. Telatar, “Capacity of multiantenna Gaussian channels”, AT&T Bell Laboratories, Tech. Memo., June 1995. I.E. Telatar, “Capacity of multiantenna Gaussian channels”, Eur. Trans. Commun., Vol. 10, No. 6, pp. 585–595, 1999. J.C. Silva et al., “Equalization Based Receivers for Wideband MIMO/BLAST Systems”, Wireless Pers. Commun., accepted for publication. J.C. Silva et al., “MMSE with Non-Uniform Modulations for a MIMO/BAST System”, Proc. Conf. Tele’ 2005, Tomar, Portugal, April 2005. F.H.P. Fitzek and M. Reisslein, “MPEG–4 and H.263 Video Traces for Network Performance Evaluation”, IEEE Netw., Vol. 15, No. 6, pp. 40–54, November/December 2001. 3GPP, 25.996-v6.1.0, “Spatial channel model for Multiple Input Multiple Output (MIMO) simulations”. FP6-IST-507607 Project B-BONE, “Broadcasting and multicasting over enhanced UMTS mobile broadband networks”, URL:http://b-bone.ptinovacao.pt/

Armando B. Soares graduated in telecommunication and computer science engineering at Instituto Superior de Ciências do Trabalho e da Empresa, Lisbon, Portugal. He is currently working for his M.Sc at the same university. He has been working as a researcher in the fields of radio resources optimization and efficient allocation for 3G UMTS networks and beyond. These research activities are being developed in ADETTI/ISCTE. Since 2003 he has been working in some EU funded telecommunications research projects, namely SEACORN and currently B-BONE.

MIMO Based Radio Resource Management for UMTS

245

João C. M, Silva graduated in aerospace engineering — avionics branch, in 2000 in Instituto Superior Técnico, Lisbon, Portugal, Mckinsey consultant (2000–2002). In 2002 is started his Ph.D. in electrical engineering in Instituto Superior Técnico, telecommunications area. His main research interests are: MIMO, CDMA, coding and modulation.

Nuno M. B. Souto graduated in aerospace engineering — avionics branch, in 2000 in Instituto Superior Técnico, Lisbon, Portugal. From November 2000 to January 2002 he worked as a researcher in the field of automatic speech recognition for Instituto de Engenharia e Sistemas de Computadores, Lisbon Portugal. He is currently working for his Ph.D. in electrical engineering in Instituto Superior Técnico. His research interests include wideband CDMA systems, channel coding, channel estimation and MIMO systems.

246

A. Soares et al.

Filipe M. L. Leitão graduated in telecommunication and computer science engineering at Instituto Superior de Ciências do Trabalho e da Empresa, Lisbon, Portugal. He is ccurrently working for his M.Sc at the same university. He has been working as a researcher in the fields of radio resources optimization and efficient allocation for 3G UMTS networks and beyond. These research activities are being developed in ADETTI/ISCTE. Since 2002 he has been working in some EU funded telecommunications research projects, namely SEACORN and currently B-BONE.

Américo M. C. Correia received the M.Sc and Ph.D. degrees in electrical engineering from Instituto Superior Técnico, Lisbon, Portugal, in 1990 and 1994, respectively. From 1991 to February 1999, he was with Instituto Superior Técnico (IST) as an Assistant Professor. He is currently with Instituto Superior de Ciências do Trabalho e da Empresa (ISTCE), Lisbon, Portugal. He visited Nokia Research Center from September to December 1998 as a visiting scientist. From September 2000 to August 2001 he has worked for Ericsson Eurolab Netherlands. His currently research interests include, Wideband CDMA systems, space-time coding for the transmission of high bit rates packets for the mobile terrestrial communication channel and software defined radio.