Cell search procedure for time division duplex communication systems using code division multiple access

(74) Plenipotentiary (54) Name (56) Publications listed (57) Summary

Method for a user equipment for synchronizing time and a group code assigned to a base station in a time-shared duplex code-shared multiple access communication system

WO 99/12273 Al

A base station sends a synchronization signal in an assigned time window to a user equipment in a time-shared duplex code-shared multiple access communication system. The base station has an assigned code group based on a predetermined number of code groups. The base station sends a selected secondary synchronization code signal from a set of secondary synchronization code signals. The multiplicity of secondary synchronization code signal numbers is less than half of the predetermined number of code groups. The user equipment identifies the transmitted selected secondary code signals. Based in part on the identified secondary synchronization code signal, the assigned code group will be determined.

<img file="NO324236B1_D0001.tif" />

The present invention is generally in the area of ​​spread spectrum time-shared duplex (TDD) communication systems using code-shared multiple access (CDMA). More particularly, the present invention is in the field of Cellular Equipment Procedure of User Equipment (UE) within TDD / CDMA communication systems, according to the attached independent claims 1.11 and 22.

Further, advantageous embodiments of the present invention are set forth in the appended dependent claims.

Fig. 1 shows a wireless spread spectrum TDD / CDMA communication system. The system has a plurality of base stations 30] to 30₇. Each base station 30] has an associated cell 34] to 34γ and communicates with the user equipment (UEs) 32] to 32₃ in its cell 34].

In addition to communicating across different frequency spectra, TDD / CDMA systems carry different communications across the same spectrum. The multiple signals are separated from each other by their respective code sequences (codes). Also, to more efficiently utilize the spectrum, TDD / CDMA systems as illustrated in FIG. 2 repeating frames 38 divided into a number of time slots 36] to 36_nsuch as sixteen time slots 0 to 15.1 such systems, communications are sent in a selected time slot 36i to 36_n by using the selected code. Accordingly, a frame 38 is capable of carrying multiple communications separated from each other at both time slots 36] to 36_n and code.

For a UE 32] communicating with a base station 30], it is necessary to synchronize time and code. Fig. 3 is a flowchart of the cell search and synchronization process. Initially, UE 32] must determine which base station 30] to 30₇and cell 34] to 34₇ in which it is to communicate. In a TDD / CDMA system, all base stations are 30] to 30₇ time synchronized with a base station group. For synchronization with UEs 321 to 37₇ each base station must have 30] to 30₇ send a primary sync code (PSC) and several other sync code (SSC) signals in the same time slot dedicated to synchronization. The PSC signal has an associated bit code, such as an unmodulated 256 hierarchical code, and is transmitted in the dedicated time slot in step 46. To illustrate, a base station 30] can transmit in one or two time slots, such as for a system using time slots 0 to 15 in time slot K or time slot K + 8, where K is either 0, ...., 7.

WO 99/12273 A1 discloses synchronization to a base station and code allocation within a spread spectrum communication system, wherein the base station broadcast is divided into time slots, each time slot including a primary synchronization code and a secondary synchronization code which includes frame synchronization and encryption signals, and the signals received by the received signals. the user equipment and determines the base station's assigned code group.

One technique used to generate a PSC signal is to use two 16 hierarchical sequences, such as XI and X2 in Equations 1 and 2.

XI = [1,1, -1, -1,1, -1,1, -1, -1, -1, -1, -1,1,1,1, -1] Equation 1

X2 = [1,1, -1, -1, -1, -1,1, -1,1,1, -1,1,1,1, -1,1] Equation 2

Equation 3 illustrates an approach for generating a 256 hierarchical code, u (i), using XI and X2.

Y (i) = XI (versus 16) x X2 (in div 16), where i = 0, ..... 255 Equation 3

Using y (i), the PSC signal is generated by combining y (i) with the first row of length 256 in the Hadamard matrix, ho, which produces C_p(i) as shown in Equation 4.

C_p(in)⁼ y (0 ^x ho (i), where i = 0, ... 255 Equation 4

Since the first row of the Hadamard matrix is ​​all one, Equation 4 will be reduced to Equation

5.

Cp (i) = y (i), where i = 0 ...... 255 Equation 5

C_p(i) is used to produce a broad spectrum PSC signal suitable for transmission.

To prevent base station communications from interfering with each other, each base station 30] to 30₇ send their PSC signal with a unique time deviation, t_{of V}i, from the time slot limit 40. Various time deviations are shown in time slot 42 in FIG. 4. To illustrate, a first base station 30i has a first time deviation 44i, tawik.1, for the PSC signal, and a second base station 30₂ has a second time deviation 44₂, t_avvjk.₂.

To differentiate the different base stations 30] to 30₇ and cells 34] to 34₇, each base station has 30] to 30₇ within the group to which they are assigned a different group of codes (code group). An approach to assign a tawik to a base station that uses a n^tea code group 44n, t_avvequation 6 is equation 6.

tawik.n = n · 71T_C Equation 6

T_c is the bit duration and each window has a duration of 2560 bits. As a result, the deviation will 42_n for each sequential code group be separated by 71 bits.

Since initially that UE 32j and base stations 30] to 30γ are not time synchronized, UE 321 must look through each bit of frame 32 for the PSC signals. To accomplish this search, the received signal is fed to a custom filter that is matched to the PSC signal bit code. The PSC custom filter is used to search through all the bits in a frame to identify the PSC signal to the base station 30j which has the strongest signal. This process is referred to as step-1 in the cell search procedure.

After UE 32] identifies the PSC signal to the strongest base station 30], UE 321 needs to determine time slot 36i to 36n where the PSC signal and SSC signals are transmitted (referred to as the physical synchronization channel (PSCH) time slot) and the code group used in the identified base station 30]. This process is referred to as step-2 in the cell search procedure. To indicate the code group assignment to the base station 30i and the PSCH time slot index, the base station 30i sends signals that have selected secondary synchronization codes (SSCs) in step 48. The UE 32i receives these SSS signals in step 50, and identifies the base station code group and PSCH time slot index based on which SSCs received in step 52.

For a TDD system that uses 32 code groups and two possible PSCH time slots per frame, such a time slot K and K + 8 can be an approach for identifying the code group and the PSCH time slot index which is transmitted as a signal having one of 64 SSCs. Each of the synchronization codes corresponds to one of the 32 code groups and two possible PSCH time slots. This approach increases the complexity of UE 321 which requires at least 64 custom filters and extended processing. To identify the code group and the PSCH time slot index, 17,344 real-time additions and 128 real-time multiplications are required for each PSCH time slot and 64 real-time additions are required for the decision.

An alternative approach for step-2 in the cell search procedure uses 17 SSCs. These 17 SSCs are used to index the 32 code groups and two possible PSCH time slots per second.

frame. To implement this approach, at least 17 custom filters are required. To identify the code group and time slot, 1,361 real additions and 34 real multiplications are required for each PSCH time slot. In addition, 512 real additions are required for the decision.

WO 99/12273 discloses a system for synchronizing to the base station. A base station broadcast is divided into time slots. Each time slot includes a primary sync code and a secondary sync code that includes both frame and sync and distortion or long code information.

TR101146 Universal Mobile Telecommunications System 30.06 version 3.0.0 shows a base station synchronization system. A primary sync code is sent for a frame and a phase reference. Each of the 16 possible base station code groups is assigned to a unique secondary synchronization code. The secondary synchronization code is sent from the base station and identifies the code group for the base station.

Higuchi et al., "Fast Cell Search Algorithm in DS-CDMA Mobile Radio using Long Spreading Codes", shows a system for assigning long spread codes to a cell. A control channel is spread by combining the unique long code of the cell site and a short code common to all cell sites. Each cell's transmitted short code has a long code group identifier code to identify the long code.

It will be desirable to reduce the complexity required in UE 32j to perform the cell search procedure.

Fig. 1 illustrates a known TDD / CMDA system.

FIG. 2 illustrates repetitive time slots in a TDD / CDMA system.

FIG. 3 is a flowchart of cell searches.

Fig. 4 illustrates time deviations used by different base stations which transmit primary synchronization code signals.

Fig. 5 is a diagram of the simplified component of a user equipment and a base station using binary phase shift key modulation for cell searching.

FIG. 6 is a flowchart of a secondary synchronization code assignment.

Fig. 7 illustrates the simplified component of a user equipment and a base station which uses quadrature phase shift key modular circuits for cell searching.

Fig. 8 illustrates the simplified component of a user equipment and a base station that reduces the maximum number of secondary synchronization codes sent using quadrature phase shift key modulation.

Figures 9 to 17 are graphs showing the performance of different synchronization systems under different simulated channel conditions.

The preferred embodiment will be described with reference to the drawing figures in which like numbers represent like elements throughout. Fig. 5 shows the simplified circuit in a base station 30] and a UE 32] for use in cell search. During step-1 in the cell search, the base station 30] generates a PSC signal using a PSC scattered signal generator 66 which has the time deviation in the time slot 42 associated with the base station 30]. The PSC signal is combined with a combiner 63 with M SSC signals. The combined signal is modulated in a carrier frequency modulator 62. The modulated signal passes through an insulator 60 and is emitted in an antenna 58 or alternatively in a group antenna.

UE 321 receives the signals using an antenna 70 or alternatively a group antenna. The received signal goes through an insulator 72 where they are demodulated in a demodulator 74 to the fundamental frequency. During step-1 of the cell search, the PSC will use the custom filter 76 in the process 80 to search through all the bits in a frame 38 to identify the PSC signal for the base station 30] which has the strongest signal.

One way to approach the detection of a PSC signal placement in a frame is as follows. A selected number of positions in the received signal frame, such as forty, which has the highest number of accumulated bit adjustments (i.e., maximum signal strength), is repeatedly correlated to the same position in subsequent frames 38. From the selected position, the one with the highest number of cumulative adjustments (i.e. maximum signal strength) be identified as the location of the PSC signal.

For step-2 of the cell search procedure, the base station generates 30] SSC signals, SSCj to SSCm, using SSC scattered spectrum signal generators 68] to 68m. To reduce the complexity of UE 321, a reduced number of SSCs will be used. By reducing the SSCs, the number of custom filters required in UE 32] can also be reduced. In addition, the reduced number of SSCs will reduce the processing resources required to separate the different codes from one another. The reduced SSCs will also reduce the likelihood of incorrect detection of a code group number and PSCH time slot index (see Figs. 9-15).

An approach for reducing SSCs is shown in the flowchart of FIG. 6. The number of SSCs used, M, is based on the number of code groups and PSCH time slots used per. frame in step 54. The number of SSCs, M, is the base two logarithm of the number of the combination number rounded up to the second highest integer in step 56, as shown in Equation 7.

M = log2 (# code groups x # PSCH time slots per frame) Equation 7 The base station 30i, using the SSC signal generators 68i to 68m, generates the SSC signals associated with the base station code group and the number of PSCH time slots per frame. frame. The SSC signals are combined with each other as well as with the PSC signal at the combiner 63. Accordingly, the combined signal will be modulated in the modulator 62 passing through the insulator 60 and emitted into the antenna 58. UE 32i receives the transmitted signal passing through isolator 72 and demodulated in demodulator 74. Using the corresponding SSCi to SSC_M custom filters 78i to 78_M, process 80 will determine the binary code with which the SSCs are modulated. Based on the particular binary code, the base station code group and PSCH time slot index in the frame will be determined. To illustrate a system that uses 32 code groups and two possible time slots per. frame, such as windows K and K + 8, the number of binary bits needed to modulate SSCs, M, will be six (log2 64). In such a system, six SSCs will be modulated by six bits using binary phase shift key modulation (BPSK). The six SSCs are selected from the 256 rows in the Hadamard matrix Hg. The Hadamard matrix is ​​generally sequential as in Equations 8 and 9.

Ho = (1) Equation 8

H1 =, t -1, ...., 8

Equation 9

A specific code, Ck,_n(i), where n is the code group number associated with an SSC is produced using Equation 10. The six rows of the Hadamard matrix, Hg, are r (k) = [24,40,56, 104,120,136].

Ck, n (i) - b_k, nxh_{r (k)}(i) xy (i) where r = 0.1 ..... 255 and k = 1 ..... 6

Equation 10

The values ​​of b₂ to b₆ are shown in Table 1.

Table 1

<td>Code group (s)</td><td>ask, n</td><td>b5, n</td><td>b ^ n</td><td>b₃n</td><td>b₂n</td>

<td> 1</td><td> +1</td><td> +1</td><td> +1</td><td> +1</td><td> +1</td>

<td> 2</td><td> +1</td><td> +1</td><td> +1</td><td> +1</td><td> -1</td>

<td> 3</td><td> +1</td><td> +1</td><td> +1</td><td> -1</td><td> +1</td>

<td> -</td><td></td><td> -</td><td></td><td></td><td></td>

<td> 32</td><td> -1</td><td> -1</td><td> -1</td><td> -1</td><td> -1</td>

The value of bee,_n are shown in Table 2.

Table 2

<td>PSCH time slot order in the frame</td><td>b, n</td>

<td>K, where K = 0, ..., 7</td><td> +1</td>

<td>K + 8</td><td> -1</td>

Each code corresponds to an SSC, SSCi to SSC₆. To distinguish the different base station SSC signals from each other, each base station SSC signal will have the same deviation in its PSC signal. The IUE 32i of step-2 of the cell search procedure (i.e., the code group number and the PSCH window order detection) is performed as follows. The received base signal is first correlated with C_p as shown in Equation 4 to find the phase reference. This correlation is performed with a PSC-matched filter 76 in FIG. 5. The phase reference is found by normalizing the correlation value found at the output of the PSC-matched filter 76. The received base signal is also correlated with Ci, the window order in the frame. This correlation is performed in the SSC-matched filters 78j-78m of FIG. 5. These custom filter outputs are rotated back before BPSK demodulation. The rotation back is performed by complex multiplication of the complex conjugate of the phase reference. The retracted SSC-custom filter outputs are then BPSK demodulated. The BPSK demodulation is performed by a hard limiter on the real portion of the retracted SSC-adapted filter outputs. As a result, the real portion of the retracted SSC-adapted filter outputs will be greater than zero and demodulated as +1. Otherwise, it will be demodulated as -1. The demodulated binary data represents the code group of the base station 30] and the PSC time slot sequence in the frame as shown in Table 1 and Table 2, respectively. To facilitate the detection of the six SSCs, the UE 32i will accumulate the back-rotated outputs of the SSC-adapted filters 78j-78m over a number of PSCH time slots such as four or eight.

Using six SSCs for 32 codegroups and two possible PSCH time slots, 653 real-time additions and 28 real-world multiplications are required in UE 32] to identify the codegroup / PSCH time slot index. No additions or multiples are required for the decision. Consequently, the reduction in the number of transmitted SSCs in the PSCH time slot will reduce the processing in UE 32i.

Alternatively, to further reduce the number of SSCs, quadrature phase shift key (QPSK) modulation can be used. To reduce the SSC number, each SSC signal will be sent either to an I phase (I) or quadrature (Q) component to the PSCH. An extra bit of data associated with using either the I or Q carrier is used to separate the code group / PSCH time slots. As a result, the number of SSCs, M, required in Equation 6 will be reduced by one.

For example, to separate 32 code groups and two possible PSCH time slots, five SSCs (M = 5) are required. The code groups are divided into two (code groups 1-16 and code groups 17-32). When the SSCs are sent in the I carrier, this will limit the code group to the lower half (code groups 1-16), and when the SSCs are sent on the Q carrier, it will be limited to the code group of the upper half (code groups 17-32). The five SSCs are separated from each other by the remaining sixteen possible code groups and two possible PSHC time slots.

A simplified base station 30i and UE 32i using QPSK modulation is shown in FIG. 7. Base station 30] generates the correct SSC signals for its code group and the PSCH time slot using SSC spread spectrum signal generators 681 to 68m. Also based on the base station code group / PSCH time slot index, switches 90i to 90m will either switch the outputs of generators 68i to 68m to an I combiner 86 or to a Q combinator 88. The combined I signal which includes the PSC signal is modulated in an I modulator 82 prior to transmission. The combined Q signal is modulated in a Qmodulator 84 prior to transmission. One approach to producing the Q carrier for modulating the signals is to delay the I carrier ninety degrees in a delay device 98. UE 32; demodulates the received signals with both an Id modulator 92 and a Q demodulator 94. Similar to base station 30i, UE 32i can produce a Q carrier for demodulation using a delay device 96. Finding binary data representing the lower or higher portions of the 16 code groups and the PSCH time slot index is the same as applying BPSK demodulation to I and the Q components of the received signal, respectively. The Custom Filters 100j to 100_M is used in process 80 to determine if any SSC signals are transmitted on the I component to the PSCH. A decision variable, Idvar, is found like this using Equation 11.

Idvar = | rxi I + | rx₂| + ... + | rx_m| Equation 11

The absolute value | rxj | is the order of magnitude of the real component (I component) of the i^teaSSC custom filter output. Similarly, the Q-matched filters 102₂ to 102_M be used in process 80 to determine if an SSC signal was sent on the Q component to the PSCH. A decision variable Qdvar is found like this using Equation 12.

Qdvar = | ixi I + | ix2 | + · · · + | IXm | Equation 12

The absolute value | ixj | is the order of magnitude of the imaginary (Q component) of the i^teaSSC custom filter output.

If Idvar is greater than Qd_var, the SSC signals will be sent on the I component. Otherwise, the SSC signals will be sent on the Q component.

Another approach is to use QPSK modulation to reduce the number of SSC signals sent as shown in FIG. 8.1 instead of sending the number of SSCs as in FIG. 7, the number of SSCs, M, representing the code group number and the PSCH time slot index will be reduced by one. To recover the one lost bit of information by reducing the SSCs, two sets of M SSCs will be used. For example, using 32 code groups and two possible PSCH time slots, one set, SSCn to SSCu, can be assigned to the lower code group such as code groups 1 to 16, and the second set, SSC21 to SSC24, can be assigned to the upper code group , such as code groups 17 through 32. For the lower code group, sending SSCs to SSC14 on the I carrier limits the code groups to 1 to

8. The Q carrier limits the code groups to 9 to 16. Similarly, the upper code group in phase SSC21 to SSC24 will limit the code groups to 17 to 24 and Q SSC21 to SSC₂₄ limiting the code groups to 25 to 32. As a result, the maximum number of SSCs sent in a time slot will be reduced by one. By reducing the number of SSCs, the interference between SSC signals will be reduced. Reduced interference between SSCs allows higher output power level for each SSC signal that simplifies detection in UE 32].

A simplified base station 30] and UE 32] implementing the reduced SSC approach is shown in FIG. 8.1 base station 30], two sets of M SSC spread spectrum signal generators 104] 1 to 1042m will generate the SSC signals corresponding to the base station code group and PSC time slot. The corresponding SSC signal is switched by using switches 106h to IO62M to either an I-82 or Q modulator 84 suitable for this base station code slot and PSCH time slot. In UE 32], an I-set of custom filters 10811 to 1 08₂q be used to determine if the SSCs were sent on

I carrier. A Q set of custom filters 110] 1 to 11 0₂m is used to determine if the SSCs were sent on the Q carrier. By detecting the transmitted I and Q SSCs, the process will determine the base station code group and PSCH time slot.

An approach to determine which of the 32 code groups and two possible PSCH time slots used in the base station 32] is followed here. After process 80 has accumulated data from the custom filters 110n to 1024, the code group set, either SSCn to SSC14 or SSC21 to SSC24, will be determined using Equations 13 and 14.

var.set 1 = | rxn | + | ix₁₂| + ... +1 rxi₄| + | ixi₄| equation 13 var.set 2 = | rx2i | + I1X22I + · · +1 rx₂₄| + I1X24I Equation 14

The values, rxi in to 1x24, are the number of accumulated adjustments for a respective SSC, SSC11 to SSC24, received in the I channel. Similarly, ixn to 1x24 is the number of accumulated adjustments for the Q channel for SSCn to SSC24, Equations 13 and 14 require a total of 16 real additions. Set 1 represents the total accumulated values ​​of the first SSC set, SSCn to SSC14. Set 2 represents the total accumulated values ​​of the second SSC set, SSC21 to SSC24. The process 80 compares var set 1 with var set 2 and the larger of the two variables is assumed to be the SSC set sent from base station 321.

To determine whether the SSCs were transmitted on the I or Q channels, Equations 15 and 16 are used.

our. I = | rx_pin | + ... +1 rxp4 | Equation 15 var. Q = | ixpi I + ... + | ixp4 | Equation 16

If set 1 is chosen to be larger than set 2, the value of p is one. Conversely, if var set 2 is larger, the value of p will be two. var.1 are the accumulated values ​​for the selected set on the I carrier and var.Q are the accumulated values ​​on the Q carrier. The larger of the two variables, var.I and var.Q, is believed to be the channel on which the selected set was transmitted. By arranging the additions in Equations 13 and 14, the values ​​of var.I and var.Q can be determined simultaneously with var.set 1 and var.set 2. Accordingly, by determining whether the I or Q carrier is used, require no addition of additions. As a result, using QPSK modulation and two SSC sets will require only 803 real additions and 36 real multiplications in each time slot and 16 real additions for the decision.

Figures 9 to 15 are graphs illustrating the performance separating the 32 code groups / two PSCH time slots of systems using 32 SSCs 128.17 SSCs 124 and 6 SSCs 126. The graphs show the performance of different simulated channel conditions. The simulations accumulated the SSC adaptations in UE 32; over four or eight PSCH time slots and compares the probability of incorrect synchronization to the channel's signal to noise ratio (SNR) in decibels.

The simulation of FIG. 9 shows an additive white Gaussian noise (AWGN) channel and accumulation over eight PSCH time slots. The simulation of FIG. 10 shows a single-way Rayleigh fading channel at six kilohertz (kHz) frequency deviation and accumulation over four PSCH time slots. The simulation of FIG. 11 shows the same as in FIG. 10 simulation except that the accumulation was performed over eight PSCH time slots. The simulation of FIG.

uses a three-way ITU channel that a UE 32i travels between at 100 kilometers per hour (km / h) and accumulation over eight PSC time slots. Fig. 13 shows simulation using an ITU channel with three multi-paths having six kilohertz (kHz) frequency deviation and a UE 32i moving at 500 kilometers per hour when accumulating over eight PSCH time slots. The simulation of FIG. 14 shows a single-way Rayleigh channel that has 10 kHz frequency deviation when accumulating over eight PSCH time slots. The simulation of FIG. 15 shows an ITU channel with three multi-paths having 10 kHz frequency deviation and UE 32i moving at 500 km / h with accumulation over eight PSCH time slots.

Under the simulated conditions of FIG. 14 and 15, 6 SSCs 128 will outperform the other techniques 124,126. As shown in FIG. 9 to 13, 6 SSCs 128 will have a favorable performance compared to the other techniques 124,126.

Fig. 16 is a graph of the simulated performance of 6 SSCs 114 using BPSK and two sets of 4 SSCs 112 using QPSK modulation. The simulation used an eight PSCH time slot accumulation of the customizations for each SSC and broadcast over an AWGN channel. As shown, two sets of QPSK modulation 112 will exceed 6 SSC BPSK modulation 114.

Fig. 17 illustrates the performance of BPSK and two sets of QPSK modulation that accumulate adjustments over four and eight PSCH time slots. The SSCs were simulated as being broadcast over a single Rayleigh channel. The performance of both modulation plans increases with the addition of time slot correlations. Two sets of QPSK modulation for four PSCH time slots 116 and eight PSCH time slots 120 outperform the BPSK modulation for four PSCH time slots 118 and eight PSCH time slots 122, respectively.