ITRM20110253A1 - METHOD OF ACQUISITION AND SYNCHRONIZATION OF CODE. - Google Patents

Description

"CODE ACQUISITION AND SYNCHRONIZATION METHOD"

DESCRIPTION

The present invention refers to a method of acquisition and synchronization of code in LTE (Long Term Evolution) environments, in particular to procedures to be adopted for the calculation of a test variable.

The code acquisition and synchronization phase is a procedure of fundamental importance in all radio-mobile systems. Before connecting to the network, a terminal must perform a series of preliminary operations to synchronize correctly, both in time and in frequency with the radio base station relating to the cell to which it belongs.

Frequency synchronization is necessary to compensate for accuracy errors due to the imperfections of the oscillators present in mobile terminals, while time synchronization is essential to correctly receive and transmit the desired information, respecting the particular structure of the telecommunication system adopted. .

The set of these operations is generally indicated with the term "cell search". The cell search is necessary in two main situations: when the terminal accesses the system for the first time (Initial cell search) and when the terminal is in an inactive state (idle mode search) or in an active state (active mode search) , to monitor neighboring cells and identify those candidates for handover (Search for target cell).

The cell search phase strongly affects the performance of the telecommunications system, as it affects the switch-on delay of the terminal devices, the stand-by time and the quality of the radio link. To assist the terminal in this phase, â € œsynchronizationâ € channels have been introduced in the various telecommunications systems.

The signals, transposed on these channels, have very precise structures, the definition of which is one of the most critical aspects in the design of a telecommunications system.

In the state of the art, the time taken by a terminal for the acquisition and synchronization of code, or Cell Search time, certainly constitutes a bottleneck of the entire system, both in terms of access to the network, both in general terms of system performance and in terms of mobile terminal performance.

Therefore, the object of the present invention is to solve the problems still left open by the known art and this is obtained through a method of acquisition and synchronization of code as defined in claim 1.

A further object of the present invention is a computer program as defined in claim 15.

Still further object of the present invention is a telecommunications terminal apparatus which adopts a method according to the present invention, as defined in claim 16.

Further characteristics of the device of the present invention are defined in the corresponding dependent claims.

This description will be conducted with particular reference to the cell search in the Long Term Evolution (LTE) cellular system, presenting an innovative solution whose performance exceeds those of conventional strategies, while at the same time guaranteeing a completely negligible increase in computational complexity. The Long Term Evolution (LTE) system presents itself as the evolution of third generation systems and guarantees high data-rates, mobility support higher than that of current standards, and excellent levels of Service Quality.

However, it is to be understood that the present invention can be used in other areas, where the same need is felt and where similar procedures for accessing a network are applied.

The present invention, overcoming the problems of the known art, entails numerous and evident advantages.

In particular, the present invention allows to effectively increase the performance of conventional receivers through operations that can be performed with low computational complexity algorithms.

Furthermore, the performance of the mobile terminals is increased thanks to the possibility of rapidly completing all the phases of cell search, code acquisition and synchronization in time and frequency with the base station relating to the cell to which it belongs.

A further advantage is given by the fact that the innovation is inserted on the receiver side and therefore does not require any modification of the standards (LTE or other) already implemented at the market level. Furthermore, it is important to underline how the invention can be effectively applied to all the LTE standards proposed for the definition of the P-SCH (non-repetitive pattern, repetitive pattern and symmetrical-periodic repetitive pattern).

The advantages deriving from the adoption of this invention are evident and find application in various fields, not only technical or industrial but also commercial. It allows, in fact, to speed up the synchronization and code acquisition operations and this involves a faster execution of the initial cell search process, reduced start-up and stand-by times, high quality of the radio link and faster management. of handover operations. Mobile network operators and device manufacturers can clearly benefit from the adoption of the invention that allows them to increase the quality of the services offered and maximize the satisfaction of end users.

These and other advantages, together with the characteristics and methods of use of the present invention, will become evident from the following detailed description of its preferred embodiments, presented by way of non-limiting example, with reference to the figures of the attached drawings, in which:

- Figures 1a to 1d illustrate examples of P-SCH (Primary Synchronization Channel) channels, on which the PSS signal is transmitted;

- figure 2 is an example block diagram of a typical power detector; - figure 3 is a representation of the autocorrelation module of a ZC sequence;

Figure 4 illustrates an example of random ordering performed on the samples of the synchronization signal PSS; And

- Figures 5a to 5c are graphs showing the probability of detection Pd as a function of SNR, for 10 <5> Montecarlo simulations and Pfa = 10 <-3>, respectively for the best case, worst case and middle case, for different values (128, 256) of the length of the IFFT.

The present invention will be described in detail below with reference to the figures indicated above.

Cell search in an LTE system

The cell search in an LTE system is based on the detection of two synchronization signals which are periodically transmitted in the downlink frames.

Before describing the structure of these two signals, a brief overview of the characteristics of the LTE physical layer is appropriate, with particular reference to the downlink section.

The medium access technique in LTE is the Orthogonal Frequency-Division Multiple Access (OFDMA). This is a multiple access technique, based on OFDM modulation, which consists in assigning to each user a subset of the sub-carriers into which the available band is divided, for a given time interval. The downlink transmission takes place according to a structure in time and in a well-determined frequency: the data are organized in frames, each frame has a duration of 10ms and consists of 10 subframes (of 1ms) each of which contains two time-slots (of 0.5 ms). Each time slot includes a number of OFDM symbols which depends on the length of the cyclic prefix (6 or 7 symbols).

In the frequency domain the resources are organized into sub-carriers, each slot is constructed from 12 sub-carriers, which together are referred to as Resource Block (RB). The smallest unit in the frequency domain is the Resource Element (RE) which consists of a subcarrier for the duration of an OFDM symbol.

The LTE system supports both Frequency Divison Duplexing (FDD) and Time Divison Duplexing (TDD), the structure of the frames and the division into blocks are the same in the two cases but the assignment of the slots for the various purposes is different in the two cases.

The cell search in an LTE system is based on two synchronization signals, broadcast in the downlink frames of each cell: the Primary Synchronization Signal (PSS) and the Secondary Synchronization Signal (SSS).

The detection of these two signals not only allows synchronization in time and frequency, but also provides the terminal with other fundamental information such as: the cell identity, the length of the cyclic prefix (CP) of the OFDM symbols and also allows to determine if the cell operates in TDD or FDD mode.

The two synchronization signals are periodically transmitted every 5ms, therefore twice within a single frame, always in the same position (which varies according to the TDD or FDD mode). The position of the PSS is chosen in such a way as to allow the terminal to identify the start of the slot (i.e. the last symbol of the first and eleventh slot), without previously knowing the length of the cyclic prefix.

The SSS signal is at a fixed distance from the PSS signal and this allows its coherent detection, in the reasonable assumption that the coherence time of the channel is greater than the time interval that elapses in the transmission of the two synchronization signals .

While the PSS transmitted in a cell is always the same, the two repetitions of the SSS are different from each other and this allows the mobile terminal to correctly acquire the beginning of the frame. The two synchronization signals also make it possible to identify the identity of the cell. There are 504 physical layer identities (NlD <cell>), grouped into 168 groups (phsical-layer-identities-groups NID <(1)> of three identities NID <(2)>. These three identities are generally assigned to the cells below control of the same node B. The three NIDs <(2)> are identified by three possible PSSs, while 168 different SSSs indicate the identity of the NID group <(1)>.

Hence, the cell search in an LTE system consists of two phases:

1. PSS signal detection: from the PSS signal detection the terminal learns the identity of the physical cell layer and the slot start instant (i.e. initial code acquisition)

2. detection of the SSS signal: from the detection of the SSS signal the terminal learns the frame synchronization, the cell identity, the length of the cyclic prefix.

Initial code acquisition is the first critical step in the synchronization process and must be performed with low delays and high performance.

The characterization of the PSS is an extremely critical phase. The PSS is notoriously generated as a Zadoff-Chu sequence of length NZC equal to 63, according to the following formula:

ï £ ± Ï € a (n 1)

∠’j

63

d () ï £ ́ n = 0,1, ..., 30

u n = e

ï £ ² Ï € u (n + 1) (n2) j (1) ï £ ́ ∠’

ï £ ³e 63 n = 31.32, ..., 61

The three different sequences, chosen to represent the three possible PSS, are generated by varying the parameter u called the root of the sequence.

Table 1: Root indices for the PSS signal

N (2)

Root ID - u

0 25

1 29

2 34

The ZC sequences were selected for their excellent auto and cross-correlation properties:

1. They are periodicals of the NZC period;

2. The discrete Fourier transform DFT of a sequence of ZC is still a sequence of ZC, conjugated, scaled in amplitude and translated in time;

3. The cyclic autocorrelation is ideal;

4. The cross-correlation between two different sequences and with NZCprimo is constant; 5. Two translated versions of the same sequence are orthogonal to each other.

The PSS signal obtained with the formula (1) in the frequency domain is then reported in the time domain through an IFFT of suitable size, preferably 128 or 256 points.

The channel on which the PSS signal is transmitted, called Primary Synchronization Channel (P-SCH), must be structured in such a way as to facilitate the detection of the correct time and frequency offset. For this purpose, various solutions have been proposed in the known art.

Typically, the P-SCH is made up of different time repetitions of the same signal, in the example in Figure 1a and 1b, the P-SCH is made up of K blocks of the same length (in Figures K it is respectively 2 or 4) , and the cyclic prefix CP is inserted at the beginning of the OFDM symbol which constitutes the P-SCH.

A further known solution, similar to the previous one, is the one illustrated in Figure 1c, in which the symbol B represents the symmetrical opposite of A. Finally, a solution has also been proposed that provides a non-periodic structure, in which the signal is repeated only once, as depicted in Figure 1d.

The optimal receiver, in the presence of white Gaussian noise, is represented by an adapted filter receiver (power detector or energy detector), of which figure 2 shows a classic block diagram.

The detection is based on the maximum likelihood rule: the received signal, at the output of the adapted filter, is cross-correlated with a time-shifted reference signal to search for the time offset that corresponds to the maximum of the cross -correlation. To ensure greater robustness to the random phase variations introduced by the channel, the so-called non-coherent approach is followed, according to which the receiver calculates the square modulus of the estimated cross-correlation, averaging the results obtained over a number W of realizations:

1 W

Zk (Ï ") = ∑ Rw (Ï ") <2> (2)

W w = 1

This value is then compared with a suitable predetermined threshold to determine the presence or absence of the signal of interest. The threshold value is chosen following the constant false alarm probability procedure (CFAR) in which the threshold is established so as not to exceed a predetermined limit of probability of false alarm. Typical threshold values are 10 <-3> or 10 <-4>.

According to the present invention, the classic power detector does not represent the optimal solution for the acquisition of code in an LTE network.

In particular, two main problems can be identified related to the properties of the ZC sequences themselves.

The cyclic autocorrelation of the ZC sequences is substantially ideal, but the same cannot be said of the linear self-correlation, which indeed has non negligible secondary peaks, as shown in figure 3.

Therefore, the present invention is an alternative to the methodology implemented up to now, capable of overcoming these difficulties, through an extremely more effective methodology and, at the same time, devoid of greater computational complexity, a characteristic that allows not to burden resources. already very limited hardware of mobile terminals.

In particular, a method according to the present invention is implemented by means of a computer, for example in a terminal apparatus of a telecommunications network.

The method is used to calculate a test variable Zk (Ï „) within a terminal code acquisition and synchronization procedure in a digital telecommunications network.

The procedure takes place through the processing of the PSS synchronization signal transmitted by a base station on the network and received by the terminal and a copy generated locally at the PSS signal receiver and used as a PSSR reference signal.

The method foresees the execution of a random ordering of the samples of the PSS synchronization signal received by the terminal, so as to obtain a randomized synchronization signal PRND.

Preferably, this random ordering of the samples of the synchronization signal PSS comprises a permutation of the random and / or pseudorandom type of the same.

Alternatively, the ordering of the samples can also be obtained by means of a sub-sampling of the signal, preferably of the random type, for example a subsampling of the aperiodic type.

Then, the synchronization signal, thus randomized, PRND, is divided into at least two synchronization blocks BSw (FIG.4). In the following the number of BSw synchronization blocks will be indicated with WB.

The WB number of BSw synchronization blocks is determined according to the specific project requirements. For example, it can be selected randomly.

However, in order to achieve better performance, as will become evident, it is preferable to use the minimum number of synchronization blocks, ie WB = 2.

The length of each BSw synchronization block is also determined on the basis of specific needs and desired results, therefore, for example, randomly. However, since the PSS signal is currently used according to the standard in force (and therefore the randomized signal PRND) consisting of 62 samples, it is considered preferable to divide the signal into blocks of equal length, preferably into two synchronization blocks, each comprising 31 samples. .

Similarly, the present invention provides that similar operations are also performed on the PSSR reference signal generated locally at the receiver.

In particular, therefore, the method provides for the execution of a random ordering of the samples of the reference signal PSSR generated locally at the terminal, so as to obtain a randomized reference signal SRND.

Preferably, this random ordering of the samples of the reference signal PSSR comprises a permutation of the random and / or pseudo-random type of the same. Alternatively, the ordering of the samples can also be obtained by means of a sub-sampling of the signal, preferably of the random type, for example a subsampling of the aperiodic type.

Preferably, the random (or pseudo-random) ordering of the PSSR reference signal is carried out with the same algorithm adopted for the PSS synchronization signal. In fact, in this way, the codes are synchronous at a single time offset, even if they are always different from block to block.

Then, the reference signal, thus randomized, SRND, is divided into at least two reference blocks BRw. The number and length of the reference blocks BRw will be equal to that of the synchronization blocks BSw already determined.

At this point, for each determined synchronization block BSw, the method provides for calculating a cross-correlation Rw with a respective reference block BRw.

A test variable Zk (Ï „) is then calculated as a function of the cross correlations Rw.

Preferably, the test variable Zk (Ï „) is calculated for each synchronization and / or reference block, according to the following relationship:

1 W

Z Ï „2

k <()> =

W ∠‘Rw (Ï")

w = 1,

where W is the predefined number of realizations on which to average the result.

Although so far the operations of random ordering and division into blocks have been described as operated in the time domain, it is understood that the same operations can be performed, both for the synchronization signal and for the reference signal, also in the frequency domain of the OFDM coefficients or also in further ortho-normal transformed domains, bi-uniquely determinable starting from the temporal domain and / or the frequency domain. Naturally, if one operates in domains other than time, the test variable can be evaluated in the same domain, in fact, also the DFT-FFT and its inverse, defined by the OFDM modulation, are, appropriately scaled, orthonormal transformations.

Therefore, in the light of this consideration, we will continue for descriptive simplicity to refer to operations performed in the time domain.

From the above it can be seen that the values of the cross-correlations Rw, unlike what happens in the known art, are different for each replica of the synchronization signal, precisely by virtue of the random ordering introduced.

The value of the test variable is then compared with a predefined threshold to determine the presence or absence of the signal of interest.

The threshold value is preferably chosen following a constant false alarm rate (CFAR) procedure in which the threshold is established so as not to exceed a predetermined False Alarm Probability limit (typical values for applications in this field are 10 <-3> or 10 <-4>).

This threshold value is also what, at the same time, allows the system to maximize the probability of acquiring the code (and the time offset) sought. This strategy allows to minimize the maximum cost errors (false alarms in fact) that would otherwise force the mobile terminal to perform synchronization and time tracking operations (with very expensive operations from the point of view of the computational complexity required) to acquire a code which will reveal, only after a long time (a penalty time in fact), a false acquisition due to noise.

In general, the code acquisition and synchronization process has a dual function: not only to determine the presence of the signal of interest (one of the three possible Zadoff-Chu sequences transmitted by the radio-base station), but also to restore the synchronism between transmitter and receiver. The time delay Ï „, in correspondence with which the test variable exceeds the threshold, also provides the mobile terminal with the information necessary to identify the temporal beginning of the frame and thus recover the correct phase offset.

In a serial search, the receiver continues to sequentially try all possible offsets for each of the three ZC sequences until the correct one is found, only then can the transmission begin.

We therefore understand the need to conclude the code acquisition and synchronization phase in the shortest possible time.

The solution offered by the present invention achieves this result, through a processing that provides for a random permutation of the signals which, preferably, varies randomly at each iteration of the algorithm and a subdivision into blocks of the same randomized signals (FIG. 4). In this way the blocks are made up each time of different samples of the PSS signal of interest. The cross-correlation between the synchronization block and the corresponding reference block gives a different result each time, because the values on which the cross-correlation are calculated are different. In this way, the subsequent averaging operation carried out on a certain number of repetitions of the input signal not only allows to mitigate the effects due to the presence of noise on the channel, but effectively eliminates the cross-correlation queues.

Below are the results of various numerical simulations, performed to verify the validity of the invention.

The graphs shown in figures 5a to 5c show the Probability Detection (Pd) as the signal to noise ratio (SNR) varies, for three different cases of the time delay between transmitter and receiver (perfect synchrony; complete asynchrony; typical intermediate operating case).

The threshold value, to identify the correct synchronization offset, was chosen following the CFAR procedure, corresponding to a typical Pfa value (Pfa = 10 <-3>). In order to thoroughly analyze the effective applicability of the proposed solution, the best case was first considered, ie the ideal case of perfect synchronization between the transmitter and the receiver (figure 5a). Subsequently, to evaluate the performance of the new method even in extremely critical situations, the so-called worst case was considered (figure 5b). This represents the most critical operating condition and corresponds to a synchronization delay between the transmitter and the receiver, equal to half a chip.

Finally, we also considered a third intermediate case, middle case, which represents the typical operational situations (Figure 5c).

The graphs show the comparison between the conventional method and the innovative solution in the three application cases, for IFFT on 128 and 256 points.

The graphs show how the performance of the proposed solution is superior to that of the classic detector in all scenarios considered. These highlight the advantages deriving from the adoption of our invention which allows to reach high performances, without any increase of the computational complexity.

It is of course to be understood that the method according to the present invention can be applied to all LTE standards for PSS: non-repetitive pattern, repetitive pattern and symmetrical-periodic repetitive pattern with two or four repetitions.

The present invention has been described heretofore with reference to its preferred embodiments. It is to be understood that other embodiments may exist which pertain to the same inventive nucleus, all falling within the scope of the protection of the claims set forth below.

Claims (16)

CLAIMS 1. Method implemented by a computer, to calculate a test variable (Zk (Ï „)) within a procedure for acquiring and synchronizing the code of a terminal in a digital telecommunications network, through the processing of a synchronization signal (PSS) transmitted by a base station of said network and received by a terminal and a reference signal (PSSR) generated locally on said terminal, the method comprising the following steps: - performing a random ordering of the samples of the received synchronization signal (PSS), obtaining a randomized synchronization signal (PRND); - dividing the randomized synchronization signal (PRND) into at least two synchronization blocks (BSw); - performing a random ordering of the samples of the generated reference signal (PSSR), obtaining a randomized reference signal (SRND); - dividing the randomized reference signal (SRND) into reference blocks (BRw), in number and length equal to said synchronization blocks (BSw); - for each synchronization block (BSw) calculating a cross-correlation (Rw) with a respective reference block (BRw); - calculate said test variable (Zk (Ï „)), as a function of said calculated cross-correlations (Rw). Method according to claim 1, wherein said steps of random ordering, subdivision into blocks and calculation, are performed in the time domain. Method according to claim 2, wherein said test variable (Zk (Ï „)) is calculated, for each synchronization and / or reference block (BSw, BRw), according to the following relationship: 1 W Zk <(> Ï „<)> = W ∠‘R 2 w (Ï „) w = 1, where W is the predefined number of realizations on which to average the results obtained. Method according to claim 1, wherein said random ordering, blocking and calculation steps are performed in the frequency domain. Method according to claim 1, in which said random ordering, subdivision into blocks and calculation steps are performed in an orthonormal transform domain, bi-uniquely determinable starting from the time domain or the frequency domain. Method according to one of claims 1 to 5, wherein said random ordering of the synchronization signal samples (PSS) comprises a permutation of the same random type. Method according to one of claims 1 to 6, wherein said random ordering of the synchronization signal samples (PSS) comprises a pseudo-random permutation of the same. Method according to one of claims 1 to 7, wherein said random ordering of the synchronization signal samples (PSS) comprises a random subsampling. Method according to claim 8, wherein said random subsampling is of the aperiodic type. Method according to one of claims 1 to 9, wherein said random ordering of the samples of the reference signal (PSSR) is carried out with the same algorithm adopted for the synchronization signal (PSS). Method according to one of claims 1 to 10, wherein said number (WB) of synchronization blocks (BSw) is equal to two. Method according to one of claims 1 to 11, wherein said number (WB) of synchronization blocks (BSw) is selected at random. Method according to one of claims 1 to 12, wherein the length of each of the synchronization blocks (BSw) is determined randomly. 14. Method according to claim 11, wherein said randomized synchronization signal (PRND) is composed of 62 samples and the length of each of the two synchronization blocks (BSw) is equal to 31 samples. Computer program adapted to implement a method according to any one of claims 1 to 14 when executed on a computer. 16. Terminal apparatus for telecommunications, characterized in that it comprises means for executing a program according to claim 15.