Tài liệu Adaptive WCDMA (P9) doc

9 CDMA network design 9.1 BASIC SYSTEM DESIGN PHILOSOPHY In Code Division Multiple Access (CDMA) systems, the capacity increase is based on how much interference the desired signal can tolerate. Prior to despreading, the signal level of a desired signal is always below the interference level. All the users have to share the same radio channel. If one user takes more power than needed, then the others will suffer and the system capacity will be reduced. In analog and TDMA systems, the most important key element is the carrier to inter- ference ratio (C/I ). There are two different kinds of C/I . One is the measured (C/I), which is used to indicate the voice quality in the system. The higher the measured value, the better it is. The other is called the specified (C/I )s, which is the required value for a specified performance of the cellular system. For example, the (C/I )s in the American mobile phone system (AMPS) is 18 dB. Since in analog and TDMA systems, owing to the spectral and geographical separations, the interference (I) is much lower than the received signal (C), sometimes we can utilize field strength meter to measure C to deter- mine the coverage of each cell. The field strength meter therefore becomes a useful tool in designing the TDMA system. In CDMA all the traffic channels are served solely by a single radio channel in every cell. In an m-voice channel cell, one of the m traffic channels is the desired channel and the remaining m − 1 traffic channels are the interference channels. In this case, at the receiver front end (prior to despreading) the interference is much stronger than the desired channel. C/I is hard to obtain by using the signal strength meter that will receive more interference than the desired signal. The key elements in designing a CDMA system are different from the key element in designing a TDMA system. We can design the CDMA system based on the specified E b /I 0 C I =  E b I 0  ·  R b B  · η ⇔ E b I 0 = G · C Iη (9.1) Adaptive WCDMA: Theory And Practice. Savo G. Glisic Copyright ¶ 2003 John Wiley & Sons, Ltd. ISBN: 0-470-84825-1 272 CDMA NETWORK DESIGN The left-hand side of equation (9.1) is derived from the right-hand side specifying that the signal-to-noise ratio (SNR) after despreading is G times higher than the input SNR. Values of E b /I 0 for the forward-link channels and for the reverse-link channels are different because of the different modulation schemes. In general, there will be two different requirements for C/I .One(C /I ) F for the forward-link channels and the other (C /I ) R for the reverse-link channels. In Chapter 8 we used (E b /I 0 ) R = 7dBand(E b /I 0 ) F = 5dB.So, for a given E b /I 0 the network design should make sure that the required C/I is guaranteed in each spot of the coverage area. In the first step, we start with a simple, very much approximative approach to the problem in order to get the very first initial insight into the system parameters. In the next iteration we will come up with a more detailed analysis. 9.1.1 Uniform cell-size scenario For the forward link a worst-case scenario is used to find the relation among the transmitted powers of cell sites. The position of the mobile for this case is shown in Figure 9.1. If we assume that the signal propagation losses can be approximated as R −4 (shadowing ignored at this stage) and if R is the cell parameter, then C/I at the mobile front end can be represented as C I = α 1 R −4 I(s)+ I(a)+ I(i)+ I(d) (9.2) where I(s) = I(self cell) = α 1 (m 1 − 1)R −4 I(a) = I(2 adjacent cells) = (α 2 m 2 + α 3 m 3 )R −4 I(i) = I(3 intermediate cells) = β(2R) −4 I(d) = I(6 distant cells) = γ(2.633R) −4 Home P t S Figure 9.1 CDMA system and its interference (from a forward link). BASIC SYSTEM DESIGN PHILOSOPHY 273 α i (1, 2, 3) is the transmitted power of each voice channel in the cell m i is the number of channels per cell β and γ are transmitted powers of the combined adjacent cells at a distance 2R and 2.633R, respectively. By solving the equation we get m 1 as follows: m 1 =  1 C/I + 1  −  α 2 m 2 + α 3 m 3 α 1  − β α 1 (2) −4 − γ α 1 (2.633) −4 (9.3) If there is no adjacent cell interference, α 2 = α 3 = β = γ = 0 in equation (9.3) and we have m 1 = 1 C/I + 1 (9.4) For C/I =−17 dB, we have m 1 = 51. If there is no interference other than from the two close-in interfering cells then α 1 = α 2 m 2 + α 3 m 3 [1/(C/I )] + 1 − m 1 (9.5) If C/I =−17 dB, m 1 = 30,m 2 = 25 and m 3 = 15, then α 1 = 25α 2 + 15α 3 51 − 30 = 1.19α 2 + 0.714α 3 (9.6) which gives the relationship among α 1 ,α 2 and α 3 . If the total transmitted power P in each cell site is P 1 = α 1 m 1 ,P 2 = α 1 m 2 ,P 3 = α 3 m 3 , when m 1 ,m 2 ,m 3 are given, then P 1 ,P 2 and P 3 are the maximum transmitted powers of three cells.  1 C/I + 1  · P 1 m 1 = P 1 + P 2 + P 3 (9.7) Following the same derivation steps  1 C/I + 1  P 2 m 2 = P 1 + P 2 + P 3  1 C/I + 1  P 3 m 3 = P 1 + P 2 + P 3 (9.8) The relationship of three maximum transmitted powers of three cells are P 1 m 1 = P 2 m 2 = P 3 m 3 274 CDMA NETWORK DESIGN Deduced from the equation, a design criterion that will be used in general for a CDMA system of N cellscanbeexpressedas P i m i = P j m j = constant For the reverse link the received signal from a desired mobile unit at the home cell site is C. Each signal of other m 1 channels received at the home site is also C (owing to power control). The interference power of certain mobile units, say r · m 1 , from the two adjacent cells comes from the cell boundary (see the worst case scenario in Figure 9.2). Because of the power control in each adjacent cell, the interference coming from the adjacent cell for each voice channel would roughly be C at the home cell site. So we have C I = C (m 1 − 1) · C + r 12 · m 2 C + r 13 · m 3 C = 1 m 1 − 1 + r 12 m 2 + r 13 m 3 (9.9) r 12 and r 13 are a portion of the total number of voice channels in adjacent cells that will interfere with the desired signal at the home cell, which is Cell 1. The worst-case scenario is when m 1 + r 12 · m 2 + r 13 · m 3 ≤ 1 C/I + 1 r 21 m 1 + m 2 + r 23 m 3 ≤ 1 C/I + 1 r 31 m 1 + r 32 m 2 + m 3 ≤ 1 C/I + 1 (9.10) Home I S I m 1 − 1 m 2 m 3 Figure 9.2 CDMA system and its interference (from a reverse-link scenario). BASIC SYSTEM DESIGN PHILOSOPHY 275 r depends on the size of the overlapped region in the adjacent cell and can be reasonably assumed to be 1/6 (which is 0.166) if the system is properly designed. If C/I =−17 dB, which is 50 −1 and r 12 = r 13 = 0.166, then m 1 + 0.166 · (m 2 + m 3 ) ≤ 51 (9.11) This is a relationship among the number of voice channels in each cell, m 1 ,m 2 and m 3 . From the reverse-link scenario, we can check to see whether all the conditions expressed in the equations can be met. The unknowns in these conditions come from the demanded voice channels, m 1 ,m 2 and m 3 . Then, on the basis of the forward-link equations, we can determine the maximum transmitted power of each cell. 9.1.2 Nonuniform cell scenario We may first assign the number of voice channels m in each cell owing to requirements from demographical data. Then we may calculate the total transmit power on the forward- link channel in each cell from the worst-case scenario as shown in Figure 9.3. The (C/I ) F received at vehicle 1 is  C 1 I 1  F = α 1 R −4 1 (m 1 − 1)α 1 R −4 1 + α 2 m 2 R −4 2 + α 3 m 3 R −4 3 + I a1 (9.12) I a is the interference coming from other interfering cells besides these three cells. This component is usually very small as compared to the two terms and can be neglected. Cell 1 Cell 3 Cell 2 R 1 R 2 R 3 m 1 – 1 m 3 – 1 m 2 – 1 a 3 a 1 a 2 I a Figure 9.3 The worst-case scenario on a forward-link channel. 276 CDMA NETWORK DESIGN (C 2 /I) F received at vehicles 2 and 3 can be expressed as  C 2 I 2  F = α 2 R −4 2 (m 2 − 1)α 2 R −4 2 + α 1 m 1 R −4 1 + α 3 m 3 R −4 3 + I a2 (9.13)  C 3 I 3  F = α 3 R −4 3 (m 3 − 1)α 3 R −4 3 + α 1 m 1 R −4 1 + α 2 m 2 R −4 2 + I a3 (9.14) If  C 1 I 1  F =  C 2 I 2  F =  C 3 I 3  F =  C I  F and I a1 = I a2 = I a3 = 0 (9.15) gives α 1 m 1 + α 2 m 2  R 2 R 1  −4 + α 3 m 3  R 3 R 1  −4 = α 1  1 (C/I ) F + 1  = α 1 · G α 1 m 1  R 1 R 2  −4 + α 2 m 2 + α 3 m 3  R 3 R 2  −4 = α 2 · G α 1 m 1  R 1 R 3  −4 + α 2 m 2  R 2 R 3  −4 + α 3 m 3 = α 3 · G (9.16) Solving these equations gives α 1 R −4 1 = α 2 R −4 2 = α 3 R −4 3 (9.17) Assume that the minimum values of α 1 , α 2 and α 3 will be α 0 1 ,α 0 2 and α 0 3 , respectively, then we have α 1 ≥ α 0 1 = C 0 R +4 1 /k 1 α 2 ≥ α 0 2 = C 0 R +4 2 /k 2 α 3 ≥ α 0 3 = C 0 R +4 3 /k 3 (9.18) where C 0 is the required signal received level at the vehicle location and k i is a constant gain related to the antenna heights at the cell sites. Now the total transmit power of each cell site will be P 1 = m 1 α 1 P 2 = m 2 α 2 P 3 = m 3 α 3 (9.19) BASIC SYSTEM DESIGN PHILOSOPHY 277 Cell 1 Cell 2 Cell 3 ( m 1 – 1)a 1 ′ r • m 2 • a 2 ′ I a r • m 3 • a 3 ′ a1 Figure 9.4 The worst-case scenario for reverse link. The worst-case scenario for reverse link is depicted in Figure 9.4. On the basis of the power control algorithm, all the signals will be the same on reaching the cell site.  C 1 I 1  R ≥ α  1 · R −4 1 (m 1 − 1)α  1 R −4 1 + r 12 m 2 α  2 R −4 1 + r 13 m 3 α  3 R −4 1 + ˙ I a1 (9.20) where α  1 ,α  2 and α  3 are the power of individual channels transmitted back to their cor- responding cell sites. r 12 and r 13 are the portion of the total number of voice channels in the adjacent cell that will interfere with the desired signal at cell 1. ˙ I a1 is the interference coming from other users in other cells that are not cell 2 and cell 3, which is a relatively small value and can be neglected. Similarly we have  C 2 I 2  R ≥ α  2 R −4 2 r 21 · m 1 α  1 R −4 2 + (m 2 − 1)α  2 · R −4 2 + r 23 · m 3 α  3 R −4 2  C 3 I 3  R ≥ α  3 R −4 3 r 31 m 1 α  1 R −4 3 + r 32 m 2 α 2 R −4 3 + (m 3 − 1)α 3 · R −4 3 (9.21) where r is the percentage of total channels from the interfering cell received by the home site. Simplifying the equations gives  I C  R ≥ (m 1 − 1) + r 12 m 2 α  2 α  1 + r 13 m 3 α  3 α  1 278 CDMA NETWORK DESIGN  I C  R ≥ r 21 m 1 α  1 α  2 + (m 2 − 1) + r 23 m 3 α  3 α  2  I C  R ≥ r 31 m 1 α  1 α  3 + r 32 m 2 α  2 α  3 + (m 3 − 1) (9.22) where  C I  R =  C 1 I 1  R =  C 2 I 2  R =  C 3 I 3  R (9.23) The minimum values of α  1 ,α  2 and α  3 can be defined as follows: α  1 ≥ α 0 1 = C 0 R 4 1 k 1 α  2 ≥ α 0 2 = C 0 R 4 2 k 2 α  3 ≥ α 0 3 = C 0 R 4 3 k 3 (9.24) where R 1 ,R 2 and R 3 are the radii of the three cells and k is a constant gain related to the antenna heights at the cell sites. Now equation (9.23) becomes  I C  R ≥ (m 1 − 1) + r 12 m 2  R 2 R 1  4 + r 13 m 3  R 3 R 1  4  I C  R ≥ r 21 m 1  R 1 R 2  4 + (m 2 − 1) + r 23 m 3  R 3 R 2  4  I C  R ≥ r 31 m 1  R 1 R 3  4 + r 32 m 2  R 2 R 3  4 + m 3 − 1 (9.25) which gives the basic design equation for the relation between the network parameters. From this we have m 1 ,m 2 or m 3 < 1 (C/I ) R + 1 (9.26) 9.2 CDMA NETWORK PLANNING In this section we provide more details on network planning and dimensioning. The approach is based on References [1–5]. WCDMA radio network dimensioning is the process through which the possible configurations and the amount of network equipment is estimated, on the basis of the operator’s requirements related to the following: CDMA NETWORK PLANNING 279 Coverage, which includes coverage regions, area type information, propagation conditions. Capacity, which includes spectrum available, subscriber growth forecast, traffic density information. Quality of Service, which includes area location probability (coverage probability), block- ing probability, end user throughput. Dimensioning activities include radio link budget and coverage analysis, capacity esti- mation, estimations on the amount of sites and base station hardware, radio network controllers (RNCs), equipment at different interfaces and core network elements (i.e. circuit-switched domain and packet-switched domain core networks). 9.2.1 Radio link budgets and coverage efficiency The interference margin is needed in the link budget because of the loading of the cell by other users. The load factor, which will be later related to (E b /N 0 ) R defined in equation (8.2) of Chapter 8, affects the coverage. The more loading is allowed in the system, the larger is the interference margin needed in the uplink, and the smaller is the coverage area. For coverage-limited cases a smaller interference margin is sug- gested, while in capacity-limited cases a larger interference margin should be used. In the coverage-limited cases the cell size is limited by the maximum allowed path loss in the link budget, and the maximum air interference capacity of the base station site is not used. Typical values for the interference margin in the coverage-limited cases are 1.0 to 3.0 dB, corresponding to 20 to 50% loading. Some headroom is needed in the mobile station transmission power for maintaining adequate closed-loop fast power control. This applies especially to slow-moving pedestrian mobiles in which fast power control is able to effectively compensate the fast fading. Typical values for fast fading margin are 2.0 to 5.0 dB for slow-moving mobiles. Handovers – soft or hard – give a gain against slow fading (lognormal fading) by reducing the required lognormal fading margin. This is because the slow fading is partly uncorrelated between the base stations, and by making handover the mobile can select a better base station. Soft handover gives an additional macro diversity gain against fast fading by reducing the required E b /N 0 relative to a single radio link, owing to the effect of macro diversity combining, as explained in Chapter 8, Section 8.7. The total soft handover gain is assumed to be between 2.0 and 3.0 dB in the examples given below, including the gain against slow and fast fading. The following system assumptions given in Tables 9.1 and 9.2 will be used in this section [1–5]. On the basis of this assumption, the link budget for three different services is shown in Tables 9.3 to 9.5. Table 9.1 Assumptions for the mobile station Speech terminal Data terminal Maximum transmission power 21 dBm 24 dB Antenna gain 0 dBi 2 dBi Body loss 3 dB 0 dB 280 CDMA NETWORK DESIGN Table 9.2 Assumption for the base station Noise figure 5.0 dB Antenna gain 18 dBi (three-sector base station) E b /N 0 requirement Speech: 5.0 dB 144-kbps real-time data: 1.5 dB 384-kbps non-real-time data: 1.0 dB Cable loss 2.0 dB Table 9.3 Reference link budget of adaptive multirate (AMR) 12.2-kbps voice service (120 km h −1 , in-car users, vehicular A type channel, with soft handover) 12.2-kbps voice service (120 km h −1 , in-car) Transmitter (mobile) Max. mobile transmission power (dBm) 21 A Body loss (dB) 3 B Equivalent isotropic radiated power (dBm) 18 c = a + b Receiver (base station) Thermal noise density (dBm Hz −1 ) −174 d Base station receiver noise figure (dB) 5 e Receiver noise density (dBm Hz −1 ) −169 f = d + e Receiver noise power (dBm) −103,2 g = f + 10 ∗ log(3840000) Interference margin (dB) 3 h Receiver interference power (dBm) −103,2 i = 10 ∗ log(10 ∗∗ [(g + h)/10 − 10 ∗∗ (g/10)] Total effective noise + interference (dBm) −100,2 j = 10 ∗ log[10 ∗∗ (g/10) + 10 ∗∗ (i/10)] Processing gain (dB) 25 k = 10 ∗ log(3840/12.2) Required E b /N 0 (dB) 5 l Receiver sensitivity (dBm) −120,2 m = l − k + j Base station antenna gain (dBi) 18 n Cable loss in the base station (dB) 2 o Fast fading margin (dB) 0 p Max. path loss (dB) 154,2 q = c − m + n − o − p Coverage probability (%) 95 Lognormal fading constant (dB) 7 Propagation model exponent 3,52 Lognormal fading margin (dB) 7,3 r Soft handover gain (dB), multicell 3 s In-car loss (dB) 8 t Allowed propagation loss for cell range (dB) 141,9 u = q − r + s − t [...]... (2000) WCDMA for UMTS New York: John Wiley & Sons, by permission of IEEE Definitions N Eb /N0 W Number of users per cell = number of users per cell ∗ (1+ soft handover overhead) Activity factor of user j at physical layer Signal energy per bit divided by noise spectral density that is required to meet a predefined Quality of Service (e.g bit error rate) Noise includes both thermal noise and interference WCDMA. .. (2000) WCDMA for UMTS New York: John Wiley & Sons, by permission of IEEE 170 Downlink 20 W Downlink 10 W Maximum path loss [dB] 165 160 155 Note: capacity gain depends on the maximum path loss 3-dB better coverage 150 10% (0.4 dB) higher capacity 145 200 400 600 800 Load [kbps] Figure 9.7 Effect on base station output to downlink capacity and coverage [1] Reproduced from Holma, H and Toskala, A (2000) WCDMA. .. investment in power amplifiers The power splitting approach requires that the operator’s frequency allocation allows the use of two carriers in the base station 9.3 SPECTRAL EFFICIENCY OF WCDMA The spectral efficiency of WCDMA can be defined either by the number of simultaneous calls of some defined bit rates as in Chapter 8 or more appropriately in third-generation systems, by the aggregated physical layer... arrival interval follows a Poisson distribution This approach can be used in dimensioning when calculating Erlang capacities There is an additional soft capacity in WCDMA if also the number of users in the neighboring cells is smaller WCDMA soft capacity is defined as the increase of Erlang capacity with soft blocking compared to that with hard blocking with the same maximum number of channels per cell... cell radius P – base station overall transmit power rik – portion of users from cell k interfering in cell i REFERENCES 1 Holma, H and Toskala, A (2000) WCDMA for UMTS New York: John Wiley & Sons 2 Wacker A et al (1999) Static simulator for studying WCDMA radio network planning issues Proceedings of VTC ’99 , Houston, TX, pp 2436–2440 3 Pace, A and Valentini, L (2000) System level performance evaluation... Indoor loss (dB) 80 12 3,52 4,2 2 15 Allowed propagation loss for cell range (dB) 133,8 s t u v =r −s+t −u It was assumed in Table 9.3 that mobile antenna gain is omnidirectional The coverage efficiency of WCDMA is defined by the average coverage area per site, in square kilometers per site, for a predefined reference propagation environment and supported traffic density From the link budgets above, the cell... and antenna pattern (e.g omni, 3-sector or 6-sector) The parameters are further explained in Table 9.6 The load equation is commonly used to make a semianalytical prediction of the average capacity of a WCDMA cell, without going into system-level capacity simulations This load equation can be used for the purpose of predicting cell capacity and planning noise rise in the dimensioning process For a classical... on base station output to downlink capacity and coverage [1] Reproduced from Holma, H and Toskala, A (2000) WCDMA for UMTS New York: John Wiley & Sons, by permission of IEEE 289 SPECTRAL EFFICIENCY OF WCDMA capacity improvement is smaller than the coverage improvement because of the load curve If we now keep the downlink path loss fixed at 153 dB, which is the maximum uplink path loss with 3 dB interference... dB corresponds to a 50% load factor and the noise rise of 6.0 dB 285 CDMA NETWORK PLANNING Table 9.6 Parameters used in uplink load factor calculation [1] Reproduced from Holma, H and Toskala, A (2000) WCDMA for UMTS New York: John Wiley & Sons, by permission of IEEE Definitions Recommended values Number of users per cell Activity factor of user j at physical layer N 0.67 for speech, assumed 50% voice... quality of service and propagation conditions The variation can be quite large (e.g 50–100%) Therefore, most system simulations that attempt to offer some indication of the average spectral efficiency of WCDMA reflect only the results for some predefined cell conditions and user behavior 9.3.1 Soft capacity Erlang capacity The traffic density can be measured in Erlang: Traffic density [Erlang] = Call arrival . specified E b /I 0 C I =  E b I 0  ·  R b B  · η ⇔ E b I 0 = G · C Iη (9.1) Adaptive WCDMA: Theory And Practice. Savo G. Glisic Copyright ¶ 2003 John Wiley. Holma, H. and Toskala, A. (2000) WCDMA for UMTS. New York: John Wiley & Sons, by permission of IEEE. SPECTRAL EFFICIENCY OF WCDMA 289 capacity improvement