DE19919545A1 - Signal sequence forming method for synchronising digital mobile radio system - Google Patents

Abstract

The method involves providing signal sequence (K(i)) which is based on a first signal subsequence (K1(j)) with the length n1 and a second signal subsequence (K2(k)) with the length n2. The second signal subsequence is repeated n1 times and is modulated with the first signal subsequence. At least one of the signal subsequences is a Golay sequence. Preferably, the signal sequence has a length of 256 and the first and second signal subsequences have a length of 16. The modulation rule is as follow: K(i) = K2(I mod n2) * K1(I div n2), for I = 0...n1*n2-1.

Description

The invention particularly relates to a method of formation one for the purpose of synchronizing at least two overs transmission units to be transmitted signal sequence, and a Method for determining a signal sequence that can be formed in this way and corresponding transmitting or receiving units.

In signal transmission systems, such as mobile radio systems, it is required that one of the communications partner (first transmission unit) certain specified Si gnale recognizes that from another communication partner (second transmission unit) are sent. It can are, for example, so-called synchronization bursts (Synchronization radio blocks) to synchronize two Synchronization partners, such as radio stations, or so-called access bursts.

To receive signals of this type in relation to the ambient noise it is be knows, the received signal continuously over a specified Correlate time duration with a predetermined signal sequence and the correlation sum over the period of the predetermined To form signal sequence. The range of the received signal, the results in a maximum correlation sum corresponds to the ge were looking for signal. The synchronization signal from the base station tion of a digital mobile radio system is, for example Signal sequence upstream as a so-called training sequence, the in the manner just described in the mobile station Correlation recorded with the stored signal sequence or is determined. So the mobile stations with the base station can be synchronized.  

Such correlation calculations are also in the base station for example in the Random Access Channel (RACH) - Detection required. In addition, a correlation be calculation to determine the channel impulse response and the Signal propagation times of received signal bursts carried out.

The correlation sum is calculated as follows:

where E (i) is a reception derived from the received signal signal sequence and K (i) is the predetermined signal sequence, i runs from 0 to n-1. The correlation sum Sm is on each other consequently for several staggered from reception signal obtained signal sequences E (i) calculated, and then the maximum value of Sm determined. Should be k consecutive Correlation sums are calculated, the calculation is effort k * n operations, whereby a multiplication and Addition is counted together as one operation.

The calculation of the correlation sums is therefore very expensive dig and requires, especially in real-time applications such as Voice communication or video telephony or in CDMA systems, powerful and therefore expensive processors used in the Be have high electricity consumption. For example is for synchronization of the standardization sensitive UMTS mobile radio system a known signal sequence the length is 256 chips (with CDMA a transmitted bit is also Called chip). The result will be every 2560 chips repeated. Since the mobile station is initially asynchronous to the Chip clock works, the received signal must be oversampled to be sufficient even with an unfavorable scanning position to get the signal. This leads to the sampling of the I and Q components for 256 * 2560 * 2 * 2 = 2621440 operations.  

The invention is also based on the object, method and Specify arrangements that allow signal sequences to be bil the, and thus to indicate signal sequences that are transmitted in Receive signal sequences are easy to determine. The invention is also based on the task of a procedure and arrangement conditions that allow these signal sequences to be specified by the Formation of correlation sums comparatively easy to do average.

The task is solved by the characteristics of the independent Claims. Further training is to the subclaims remove.

The invention is based on the idea of bil signal sequences by a second signal sub-sequence of length n2 n1 times is repeated, with the first signal sequence modu is gated, and at least one of the signal sequences a Go lay sequence is.

This allows signal sequences to be formed which, when in a received signal sequence are easily determined can be. In particular, the use of Golayse sequences are advantageous because they are used to calculate the correlation very effective algorithm is known.

For example, when using a hierarchy correlation sequence of length 256, which consists of 2 constituent 16 Golay sequences is built for the PSC of a UMTS system the computing effort compared to a conventional Realization using a Golay sequence of length 256 from 15 to 14 additions per calculated correlator output worth be reduced.

A further development of the invention provides that the permutation P ₁ , P ₂ , P ₃ , P ₄ and unit size W ₁ , W ₂ , W ₃ , W _{4 used} to form the signal sub-sequence following set of permutation-unit size pairs ( P ₁ P ₂ P ₃ P ₄ , W ₁ W ₂ W ₃ W ₄ ): 3201, + 1-1 + 1 + 1; 3201, -1-1-1 + 1; 3201, -1-1 + 1-1; 3201, + 1-1-1-1; and that the permutation used to form the second signal sequence (P ₁ P ₂ P ₃ P ₄ ) is 3201.

A particularly favorable implementation variant of the Invention in ASICs are made possible.

By specifying the procedure for forming signal sequences are also the signal sequences caused by such a Ver driving can be formed or are available in the frame the invention. In particular, their use in data transmission systems, in particular for the purpose of synchronization sation of a mobile station with a base station.

To determine a contained in a received signal sequence predetermined signal sequence by determining correlates tion sums becomes a partial correlation sum sequence of the second Partial signal sequence with corresponding parts of the received signal sequence calculated. To calculate a correlation sum who selected the n1 elements of the partial correlation sum sequence and in the sense of a scalar product with the first signal part follow multiplied.

In a further development of the invention, calculations are made once partial correlation sums are saved and used for calculation other correlation sums are used.

So it is possible when calculating further correlations to sum previously calculated partial correlation sums turn and thus reduce the computational effort enormously.

The received signal sequence is also understood to mean a signal sequence which, for example, by demodulation, filtering, dero tation, scaling or analog / digital conversion from one received signal was derived.

In the following the invention is based on various embodiments Rungsbeispiele described in more detail to explain them Figures listed below serve:

FIG. 1 is schematic illustration of a mobile radio network;

Fig. 2 block diagram of a radio station;

Release attachments Figure 3 is a conventional method for calculating Korrelati.

Fig. 4 representation of signal sequences according to the invention and signal sequences;

Fig. 5 is a schematic representation of the formation of the signal sequence according to the Invention;

Fig. 6, 7 and 8 are schematic illustration of a method for calculating a correlation sum;

FIGS. 9 and 10 schematic representation of a Ausführungsva riante a method for forming the correlation sum;

Fig. 11 block diagram of an efficient hierarchical Go lay-correlator.

In Fig. 1 is a cellular mobile network, such as the GSM (Global System for Mobile Communication) system shown, which consists of a variety of mobile switching centers MSC, which are networked with each other, or access to a fixed network PSTN / ISDN produce. Furthermore, these mobile switching centers MSC are each connected to at least one base station controller BSC, which can also be formed by a data processing system. A similar architecture can also be found in a UMTS (Universal Mobile Telecommunication System).

Each base station controller BSC is at least with a base station BS connected. Such a base station BS is a radio station that uses a radio interface Radio connection to other radio stations, so-called mobile can build MS stations. Between the mobile stations MS and the base station BS assigned to these mobile stations MS can use radio signals to transmit information within radio channels f, which lie within frequency bands b, over will wear. The range of the radio signals of a base station tion essentially define a radio cell FZ.

Base stations BS and a base station controller BSC can be combined into a base station system BSS. The Base station system BSS is also for radio channel ver administration or allocation, data rate adjustment, monitoring of the radio transmission link, hand-over procedures, and in the case of a CDMA system for the allocation of the spreading code sets, responsible and transmits the nö term signaling information to the mobile stations MS.

In the case of a duplex system, FDD (Frequency Divi sion duplex) systems, such as the GSM system, for the uplink and (Mobile station (transmitting unit) to the base station (receiving unit)) other frequency bands than for the Downlink d (base station (transmitter unit) to the mobile station (Receiving unit)). Within the different frequency bands b can be replaced by an FDMA (Frequency Division Multiple Access) method, multiple frequency channels f can be realized.

In the context of the present application, one understands about carrying unit also communication unit, transmitting unit, Receiver unit, communication terminal, radio station, mobile station or base station. Use in the context of this registration Terms and examples often refer to GSM mobile radio system; however, they are in no way concerned limits, but can based on the description of one Specialist also easily to other, possibly future,  Mobile radio systems, such as CDMA systems, especially wide-band CDMA systems are mapped.

Using multiple access methods, data can be accessed via a Radio interface efficiently transmitted, separated and one or several specific compounds or the corresponding Participants will be allocated. A time multiple can do this attacked TDMA, a frequency division multiple access FDMA, a code lot subject access CDMA or a combination of several of these Multiple access procedures are used.

In FDMA, frequency band b is divided into several frequency channels f disassembled; these frequency channels are multiplied by time access TDMA divided into time slots ts. The inside egg nes time slot ts and a frequency channel f transmitted Signals can monitor the data through connection-specific dulated spreading codes, so-called CDMA codes cc who separated the.

The resulting physical channels are assigned to logical channels according to a defined scheme. There are two basic types of logical channels:
Signaling channels (or control channels) for the transmission of signaling information (or control information) and traffic channels (Traffic Channel TCH) for the transmission of user data.

The signaling channels are further divided into:

• - broadcast channels
• - Common control channels
• - Dedicated / Access Control Channel DCCH / ACCH.

Broadcast channels belong to the group of broadcast channels Control Channel BCCH, through which the MS radio information mations from the base station system BSS, the frequency Correction Channel FCCH and the Synchronization Channel SCH. The Random Access Chan belongs to the Common Control Channels nel BACH. Which are used to implement these logical channels carried radio blocks or signal sequences can be used for under  different purposes signal sequences K (i) so-called correlation sequences contain, or on these logical channels for un Different purposes signal sequences K (i) are transmitted.

The following is an example of a method for synchronizing tion of a mobile station MS with a base station BS tert: During a first step of the initial base station search or cell search (initial cell search procedu re) the mobile station uses the primary synchronization channel (primary synchronization channel SCH (PSC)) to a Time slot synchronization with the strongest base station too to reach. This can be done with a matched filter (matched filter) or a corresponding circuit ensures who the one to the primary synchronization code cp, that of al len base stations is broadcast, is adjusted. Doing so the same primary synchronization from all base stations BS tion code cp length 256 sent.

The mobile station uses correlation to determine the received signal sequences K (i) from a received sequence according to a principle which is explained in FIGS . 6 to 11 and the associated description. Peaks are output at the output of a matched filter (matched filter) for each received signal sequence of each base station located within the reception range of the mobile station. The detection of the position of the strongest peak enables the timing of the strongest base station to be determined modulo the slot length. In order to ensure greater reliability, the output of the adapted filter can be accumulated non-coherently over the number of time slots. The mobile station thus carries out a correlation over a signal sequence with a length of 256 chips as a matched filter operation.

The synchronization code cp is formed in accordance with a signal sequence K (i) according to a principle, as explained in FIG. 5 and associated description, or can be formed in such a way or is obtainable in this way. The signal sequence K (i) or the synchronization code cp of length 256 is formed from two signal sub-sequences K1 (j), K2 (k), each of which has a length of 16, or can be formed in this way. These partial signal sequences form a pair of partial signal sequences (K1 (j); K2 (k)).

A signal sequence K (i) obtainable in this way can also "hierarchical signal sequence" are called. A signal part sequence can also be called "short correlation sequence".

At least one partial signal sequence is a Golay sequence X _n (k), which can be formed by the following relationship:

X ₀ (k) = δ (k)
X ' ₀ (k) = δ (k)
X _n (k) = X _n-1 (k) + W _n .X ' _n-1 (kD _n )
X _n (k) = X _n-1 (k) - W _n .X ' _n-1 (kD _n ),

k = 0, 1, 2,. . ., 2 ^NX -1
n = 1, 2,. . ., NX
D _n = 2 ^Pn

With
nx = 2 ^NX
δ (k) Kronecker's delta function
P _n , n = 1, 2,. . ., NX, is any permutation of the numbers {0, 1, 2,. . ., NX-1} for the X sequence,
W _n weights for the X sequence (+1, -1, + i or -i).

W _n can therefore assume the values +1, -1, + i or -i or assume the values +1 or -1 for generating binary Golay sequences. In the context of the present application, W _{n is} also referred to as a unit size.

In one embodiment of the invention, at least one signal sub-sequence is a Golay sequence optimized in terms of frequency error, in particular of length 16, the permutation P ₁ , P ₂ , P ₃ , P ₄ and unit size W ₁ , W ₂ , W used to form the sequence ₃ , W ₄ , a set of permutation unit size pairs is taken, which is specified in one and / or more of claims 5, 6 or 7.

Through the use of a so formed or signal sequence or synchronization code cp that can be formed Computational effort in the correlation sum calculation for Er averaging of the signal sequence K (i) in the receiving mobile status on MS significantly reduced for the purpose of synchronization become.

The autocorrelation function of one by two signal subfol gene signal sequence K (i) has in contrast to an orthogonal one used in conventional methods Gold code generally has poorer autocorrelation properties create. For example, it has higher secondary maxima and egg higher rms value of the secondary minima. Also show UMTS link level simulations that when using such Signal sequences K (i) in PSC- for slot synchronization at one Frequency offset between transmitter and receiver of the Synchroni sation error as opposed to using an orthogonal Gold codes are generally higher.

Through elaborate simula created especially for this purpose tion tools, however, could be made from at least one Golayse existing signal partial sequence pairs (K1 (j); K2 (k)) telt, on the basis of which, as explained above, Signalfol gen K (i) can be formed or can be formed, in particular especially for synchronization between base station and mobile station even with a higher frequency offset between sen the and recipient can be reliably determined. Here was also used by one in the simulations for the UMTS system Frequency offset of 10 kHz assumed. By using it Such signal sequences K (i) become the calculation effort Calculation of the correlation sums significantly reduced without a simultaneous increase in the synchronization error in To have to buy. It can also be more expensive to use  No crystals in the receiver for frequency stabilization become.

Due to the complex simulations, a lot of Go Lay sequences of length 16 described by a set of Permutation-one-size-pairs, which in one and / or meh reren of claims 5, 6 or 7 is specified, determined those on the basis of which signal sequences K (i) can be formed, which are so probably with zero frequency offset between transmitter and receiver, as well as with a larger frequency offset when used for Synchronization purposes a small synchronization error exhibit.

With complex simulations it turned out that the Use of a signal sequence K (i), both in smaller as well as with larger frequency errors (frequency offset) Has synchronization properties particularly advantageous is. From this follows a preferred choice of permutation One size pairs, from which signal sequences and finally signal sequences K (i) are available or can be formed. The calculation of the autocorrelation function depending the frequency error turned out to be particularly ge is suitable for assessing the synchronization properties ner Si formed by a permutation-one size pair signal sequence K (i).

The following criteria can also be used to select these partial signal sequence pairs (K1 (j); K2 (k)):

• - Autocorrelation function: The calculation of the autocorrelation function, taking into account a frequency offset between the transmitter and receiver unit, can also be carried out according to the following formula:
  

The values a (κ) for κ = 0. . . n-1 can be calculated. If there are several pairs of signal sequences, the same good ratio of the main maximum to the maximum secondary maximum in the autocorrelation function of the resulting signal fol ge K (i) result, the Si Signal sequence pairs that have a lower effective value of the Side minima have to be selected. In doing so the ratio of the main maximum to the maximum secondary maximum be as large as possible and the effective value of the secondary minima possible as small as possible. Through subsequent link-level simulations for for example, the UMTS system can have partial signal sequences can be determined in the event of frequency errors 0 kHz and 5 kHz and 10 kHz are similar in terms of synchronization error behave well, like a conventional orthogonal gold code, which is not hierarchical, and for the Synchronisa tion is known to have very good properties.

• - Missed detection rate: select the signal substring pairs by comparing the missed detection rate when performed complete simulations.
• - Detection probabilities for a given frequency error and given SNR on AWGN channels.

Within the scope of the invention there are 10 to 10ˆ2 permutation One size pairs, which are basically possible Set of permutation unit size pairs selected were. The selected partial signal sequences therefore only form egg  ne very small subset of the basically possible amount of usable for the formation of 16-digit Golay sequences Permutation one size pairs.

It has proven to be particularly advantageous to use a signal sequence K (i) which gives the permutation P ₁ , P ₂ , P ₃ , P ₄ and unit size W ₁ , W ₂ , W ₃ , W _{4 of the} following amount of permutation used to form a signal sub-sequence - Unit size pairs (P ₁ P ₂ P ₃ P ₄ , W ₁ W ₂ W ₃ W ₄ ) is taken from:

0213, + j + j + j-1; 0213, -j + j + j-1; 0213, + 1-j + j-1; 0213, -1-j + j-1;
0213, + 1 + yy-1; 0213, -1 + yy-1; 0213, + yyy-1; 0213, -jjj-1;
0213, + j + j + j + 1; 0213, -j + j + j + 1; 0213, + 1-j + j + 1; 0213, -1-j + j + 1;
0213, + 1 + j-j + 1; 0213, -1 + j-j + 1; 0213, + yy-j + 1; 0213, -jj-j + 1;
3120, + 1-j + j-1; 3120, -1-j + j-1; 3120, + 1 + yy-1; 3120, -1 + yy-1;
3120, + 1 + j + j + j; 3120, -1 + j + j + j; 3120, + 1-j-j + j; 3120, -1-j-j + j;
3120, + 1 + j + jj; 3120, -1 + j + jj; 3120, +1-yyy; 3120, -1-yyy;
3120, + 1-j + j + 1; 3120, -1-j + j + 1; 3120, + 1 + j-j + 1; 3120, -1 + j-j + 1.

It has proven to be particularly advantageous to use a signal sequence K (i) which gives the permutation P ₁ , P ₂ , P ₃ , P ₄ and unit size W ₁ , W ₂ , W ₃ , W _{4 of the} following amount of permutation used to form a signal sub-sequence - Unit size pairs (P ₁ P ₂ P ₃ P ₄ , W ₁ W ₂ W ₃ W ₄ ) is taken from:

3201, + 1-1 + 1 + 1; 3201, -1-1 + 1 + 1; 3201, + 1-1-1 + 1; 3201, -1-1- 1 + 1; 3201, + 1-1 + 1-1; 3201, -1-1 + 1-1; 3201, + 1-1-1-1; 3201 1-1-1-1; 1023, + 1 + 1-1 + 1; 1023, -1 + 1-1 + 1; 1023, + 1-1-1 + 1; 1023, -1-1-1 + 1; 1023, + 1 + 1-1-1; 1023, -1 + 1-1-1; 1023, + 1-1-1- 1; 1023, -1-1-1-1.

It has proven to be particularly advantageous to use a signal sequence K (i) which gives the permutation P ₁ , P ₂ , P ₃ , P ₄ and unit size W ₁ , W ₂ , W ₃ , W _{4 of the} following amount of permutation used to form a signal sub-sequence - Unit size pairs (P ₁ P ₂ P ₃ P ₄ , W ₁ W ₂ W ₃ W ₄ ) is taken from:

3201, + 1-1 + 1 + 1; 3201, -1-1-1 + 1; 3201, -1-1 + 1-1; 3201, + 1-1-1-1 and the permutation (P ₁ P ₂ P ₃ P ₄ ) used to form the second signal sub-sequence is equal to 3201.

If partial signal sequences (constituent sequences) are used for the PSC of UMTS Golay sequences with a length of 16, the weights W _n = 1, -1, i, -i and the delays being any permutation from D _n = {1, 2, 4 , 8}, there are more than 2 ¹² different possibilities for each of the two constituent sequences. So a total of 2 ²⁴ options. Due to the selected permutation-unit size pairs, only 16 to 10 ˆ2 different constitutive sequences and thus only a very small proportion of fundamentally possible constitutive sequences can be used for the formation.

Fig. 2 shows a radio station, which can be a mobile station MS, consisting of an operating unit or interface unit MMI, a control device STE, a processing device VE, a power supply device SVE, a receiving device EE and possibly a transmitting device SE.

The control device STE essentially consists of a program-controlled microcontroller MC, the writing and le send can access SPE memory modules. The Microcon troler MC controls and controls all essential elements and functions of the radio station.

The processing device VE can also by a digita len signal processor DSP be formed, also on Memory modules SPE can access.

In volatile or non-volatile memory chips SPE is the program data used to control the radio station and the communication process, especially the Signali procedures, are needed and during processing  information generated by signals is stored. Au In addition, it can signal sequences K (i) that lead to correlation used for purposes and interim results by Korrela sum calculations can be saved. The under the Signal sequences K (i) according to the invention can thus be in the Mo bilstation and / or the base station can be stored. It it is also possible that one or more of the above Permutation unit size pairs or Si derived therefrom Signal sequences or signal sequence pairs (K1 (j); K2 (k)) in the mobile station and / or the base station are. It is also possible that in the mobile station and / or the base station a signal sequence K (i) from a signal part follow pair (K1 (j); K2 (k)) and / or a signal sub-sequence a permutation unit size pair is formed.

In particular, in one base station or in all base Stations of a system stored a signal sequence K (i) be that at fixed or variable intervals to Synchronisa tion purposes is broadcast. That is in the mobile station MS Signal partial sequence pair (K1 (j); K2 (k)), from which those in the base station stored signal sequence K (i) can be formed or ge can be formed, saved and becomes the Synchronisati on the mobile station with a base station for computing wall-friendly correlation sum calculation.

The high-frequency part HF may consist of the transmitter SE, with a modulator and an amplifier V and an Emp catching device EE with a demodulator and also one Amplifier. Through analog / digital conversion, the analog Audio signals and the analog from the receiving device EE originating signals converted into digital signals and di processed signal processor DSP. After processing If necessary, the digital signals are processed by digital / analog conversion into analog audio signals or other output signals and analog signals to be supplied to the transmission device SE changes. For this purpose, a modulation or demodulation may be used lation carried out.  

The transmitting device SE and the receiving device EE the frequency of a voltage control via the synthesizer SYN ten oscillator VCO supplied. By means of the voltage control The oscillator VCO can also be used to clock the system clock Processor devices of the radio station are generated.

To receive and send signals over the air section The place of a mobile radio system is an antenna device ANT provided. In some known mobile radio systems, such as the GSM (Global System for Mobile Communication) Signals received and pulsed in time in so-called bursts Posted.

The radio station can also be a base station BS act. In this case, the speaker element and that Microphone element of the MMI control unit through a connection to a mobile network, for example via a base status onscontroler BSC or a switching device MSC er puts. In order to simultaneously transmit data with several mobile stations MS to replace, the base station BS has a corresponding a large number of transmitting or receiving device.

In Fig. 3, a received signal sequence E (l) in which it can also be a derived from a received signal burst, the length w is illustrated. To calculate a first correlation sum S0 according to the formula given above, elements of a first section of this received signal sequence E (l) are multiplied in pairs by the corresponding elements of the signal sequence K (i) of length n, and the length of the resulting partial results is added to the correlation sum S0 .

To calculate a further correlation sum S1, the Signal sequence K (i) as shown in the figure by one Element shifted to the right and the elements of the signal follow K (i) with the corresponding elements of the signal sequence  E (l) multiplied in pairs, and by a summation of the resulting partial results again the correlation sum S1 educated.

The pairwise multiplication of the elements of the signal sequence with corresponding elements of the received signal sequence and the subsequent summation can also be described in vector notation as the formation of a dot product, provided that the elements of the signal sequence and the elements of the received signal sequence are combined to form a vector of a Cartesian coordinate system:

The maximum can be found in the correlation sums S determined in this way are searched, the maximum of the correlation sums S with ei be compared to a predetermined threshold value, and thus determined be determined whether in the received signal E (l) the predetermined Signal sequence K (i) is included and if so, where in the reception signal E (l) it is, and so two radio stations with are synchronized with each other or data to which an individual dual spread code in the form of a signal sequence K (i) aufmodu was detected.

In FIG. 4, the received signal sequence E (l) and as a correlation sequence, a signal sequence K (i), the gnalteilfolgen on the Si K1 (j) K2 (k) is again based, is shown.

In Fig. 5 the formation of a signal sequence K (i) is Darge is based on two signal sequence K2 (k) of length n2 and K1 (j) of length n1. For this purpose, the signal sub-sequence K2 (k) is repeated n1 times and modulated by the signal sub-sequence K1 (j). The formation of the signal sequence K (i) can also be expressed mathematically by the following formula:

K (i) = K2 (i mod n2) * K1 (i div n2), for i = 0. . . n1 * n2-1.

Here mod denotes the integer remainder of a division and div the integer result of a division.

This is illustrated by a sequence f2 that consists of the repeated signal sequences that are shown one after the other K2 (k), and a sequence f1, which is represented by an expanded Si Signal sequence K1 (j) is shown above the sequence f2.

By multiplying the elements of sequence f2 by the corresponding elements depicted over the sequence f2 Sequence f1 creates the new signal sequence K (i) of length n The generation of a signal sequence K (i) is still in the picture below using an example of two binary signal sequences the length 4 shown.

Signal sequences K (i) formed in this way can be simplified Calculation of correlation sums of these signal sequences K (i) can be used with received signal sequences E (l).

A schematic representation of such a simplified and thus also faster and less expensive calculation of correlation sums S is shown in FIGS . 6 to 8, which will be discussed below.

First, a partial correlation sum TS (z) is formed. For this purpose, for example for the first element of the partial correlation sum sequence TS (0), the correlation sum of the second signal partial sequence K2 (k) is formed with the corresponding section of the received signal sequence E (l).

For the second element of the partial correlation sum sequence TS (1), the second signal partial sequence K2 (k) is shifted by one element, as illustrated, and the correlation sum is also formed with the corresponding element of the received signal sequence E (l), etc.

The nth element of the partial correlation sum sequence TS (n1 * n2-1) is calculated accordingly after n-1 shifts of the second signal partial sequence K2 (k) with respect to the received signal sequence E (l).

The resulting partial correlation sum sequence TS (z) is shown in the upper area of FIG. 7. Each n2-th element is now selected from this partial correlation sum sequence and multiplied in pairs with the corresponding element of the first signal partial sequence K1 (j).

If the selected elements of the partial correlation sum sequence TS (z) and the first signal partial sequence K1 (j) are combined into vectors, the first correlation sum S0 is generated by the dot product of these two vectors.

Fig. 7 shows in the lower region, the corresponding calculation of further correlation sums S1 and S2 by selecting n2 th to 1 and 2 to the right of the selected first ele ments lying elements:

By storing the partial correlation sum calculated once men TS can use this for the later calculation of distance their correlation sums are used, and thus the corresponding calculation steps are dispensed with.

Depending on the design variant, the com complete partial correlation sum sequence TS (z) over the whole emp catch signal sequence E (l) are calculated and then the individual Correlation sums or only when necessary to calculate a new correlation sum, the corresponding additional partial correlation sums are calculated.  

FIG. 8 again shows the two-step method for calculating correlation sums S, this time using the example of two binary signal sequences of length 4 shown in FIG. 5.

In a first step, the partial correlation sums TS (z) correspond to the second signal sub-sequence K2 (k) ++ - + calculated the sections of the received signal sequence E (l), and then in a second step every fourth element of the so he witnessed partial correlation sum sequence TS (z) selected with the corresponding element of the first signal sub-sequence K1 (j) + - + multiplied and summed up to the correlation sequence S0.

The thick lines represent the new ones leading calculation steps for the calculation of a further correlation sum S1, in the event that the others Partial correlation sums TS previously calculated and saved were saved.

This variant can be as memory efficient as possible be carried out if every n2-th partial correlati sum is calculated. For this purpose, the samples between cached.

FIGS. 9 to 10 illustrate another embodiment for simplified calculation of correlation sums S based on the already above-mentioned example of two binary signal sequence of length 4 before.

First every 4th element of the received signal sequence E (l) is selected and the partial correlation sum sequence TS (z) so selected elements with the partial signal sequence K1 (j) ge forms. From the resulting partial correlation sequence TS (z) are each 4 consecutive elements selects the signal part in pairs with corresponding elements follow K2 (k) multiplied and the resulting partial results nisse summed up to the correlation sum S. Do like  which the thick lines draw the additional necessary Steps to calculate another correlation sum S1 represents, in the event that the other partial correlation sums TS have already been calculated and saved.

Fig. 10 shows again the calculation of a first correlation sum S0 in which every fourth element of the received signal sequence E (l) is selected, these elements are multiplied with corresponding elements of the first signal sub-sequence K1 (j) + - + and the partial correlation sum TS (0) is calculated by summing the partial results. In a second step, the first four successive elements of the partial correlation sum sequence TS (z) are multiplied by the corresponding elements of the second signal partial sequence K2 (k) ++ - + and the resulting partial results are summed up to the correlation sum S0.

In this embodiment variant, less memory becomes shared the partial correlation sums required if the Totals are calculated successively.

Another embodiment of the invention is made by the regular construction principle of the signal sequence K (i) be conditioned regular (almost periodic) structure of the ape periodic autocorrelation function of this signal sequence Ge need. This means that when looking for a signal not only gives a main maximum, but in regular Intervals also occur secondary maxima. For accelerated su che according to the signal sequence in the received signal sequence take advantage of the regularity of the location of the maxima. Once a Secondary maxima were found, based on the periodizi predict the location of the other maxima, d. H. one calculates the correlation sum only at these points. On in this way you can quickly detect the main maximum. However, it can be the supposed secondary maximum even by a random increase (because of the noise component) Act worth. In this case, one will look at the potential  The expected main maximum is actually not a maximum Find. Therefore in this case the hypothesis is rejected and the calculation continues conventionally.

One can do this through the construction principle of the signal sequences conditional regularity of the secondary maxima but also to the elimi nation and correction of disturbing secondary maxima in the correlation Take advantage of the result. After detection of the maximum can one can calculate the secondary maxima from the maximum and this value subtract from the corresponding correlation results. In this way you get the correlation result of a (Hypothetical) sequence with perfect autocorrelation function on. This results from the regularity of the secondary measure xima a very simplified calculation.

Variants of the invention are used for calculation of scalar products, correlation sums and / or partial correla sum of sums Efficient Golay correlators used.

FIG. 11 shows an efficient hierarchical correlator for signal sequences, with X1, Y2 sequence sequences being used as constituent sequences K1, K2. It consists of two matched filters (a) connected in series, each of which is formed as an efficient Golay correlator. b) shows the matched filter for sequence X and c) shows the matched filter for sequence Y.

Definition:
a _n (k) and b _n (k) are two complex sequences of length 2 ^N ,
δ (k) is the Kronecker delta function,
k is an integer representing time,
n is the iteration number,
D _n is the delay,
P _n , n = 1, 2,. . ., N, is an arbitrary permutation of the numbers {0, 1, 2,. . ., N-1},
W _n can assume the values +1, -1, + i, -i as weights and is also referred to as a unit size.

The correlation of a Golay sequence of length 2 ^N can be carried out efficiently in the following way:
The sequences R _a ⁽⁰⁾ (k) and R _b ⁽⁰⁾ (k) are defined as R _a ⁽⁰⁾ (k) = R _b ⁽⁰⁾ (k) = r (k), where r (k) the received signal is (or is already the output of another correlation level).

The following step is carried out N times, n runs from 1 to N:
Calculate

R _a ⁽ⁿ⁾ (k) = W * _n * R _b ^(n-1) (k) + R _a ^(n-1) (k - D _n ).

And

R _b ⁽ⁿ⁾ (k) W * _n * R _b ^(n-1) (k) - R _a ^(n-1) (k - D _n ).

Where W * _{n is} the complex conjugate of W _n . If the weights W are real, W * _{n is} identical to W _n .

R _a ^(N) (k) is then the correlation sum to be calculated.

In the above formula, the quantities R _{b are} multiplied by the weights. Alternatively, you can also multiply the sizes R _b by weights W ', whereby the weights W' can be calculated from the weights W. Possibly. the input signal itself, or the output signal is multiplied; this multiplication can also be omitted for use as a correlator, since a constant factor for the application of the result is often irrelevant. Alternatively, this multiplication can be combined with other mostly necessary scalings.

To build an Efficient Golay correlator for a PSC code of length 256 (2 ⁸ ) chips in the receiver, you need 2 * .8 - 1 = 15 complex adders.

With the combination of hierarchical correlation and Effi cient Golay correlator, you only need 2.4 - 1 + 2.4 - 1 = 14 complex adders for a hierarchical code - described by two constituent sequences X and Y - of length 256 (2 ⁴ .2 ⁴ ) (even if one uses four-valued constituent sequences).

This reduces the calculation effort by 7%. This is enormous, especially since the computing effort for the primary Synchronista tion in CDMA mobile radio systems is very high.

Variants of the invention are given below:
Two constituent Golay sequences of length nx = 2 ^NX and ny = 2 ^NY are used to form a code sequence of length 2 ^{NX + NY} and are hierarchically structured as described above.

We use as weights for the constituent Golayse sequences +1 and -1 and thus generate binary sequences.

We use as weights for the constituent Golayse sequence +1, -1, i or -i and thus generate 4-valued sequences Zen.

Real Golay sequences are used.

Complex Golay sequences are used.

Two constituent Golay sequences of the same are used Length.

Two complementary Golay sequences are used.

Only an Efficient Golay correlator is used, possibly with Programmable delays for the optional calculation of one or both complementary Golay sequences.  

You use a sequence as described, but still add additional values. When calculating these values must be like be accumulated as usual. The rest of the calculation can be carried out efficiently as described. This allows the generation of sequences of any length.

Two constituent parts are used.

One uses several constituent partial sequences.

You only use a Golay for a part of the partial sequences Episode.

These sequences are used for the synchronization channel in UMTS.

Constituents optimized for frequency errors are used Golay sequences.

Two filters are used to calculate the correlation, one being a matched filter on the Golay sequence X and the other a matched filter on the Golay sequence Y with spread delays ny.DX _n .

For the calculation of the correlation, two filters connected in series are used, one being a matched filter on the Golay sequence X and the other a matched filter on the Golay sequence Y with spread delays ny.DX _n and the output signals of the filters corresponding to the Efficient Golay - Correlator algorithm can be calculated.

One uses the to calculate the partial correlation sums Efficient Golay correlation algorithm and to determine the overall correlation the algorithm for hierarchical Correlation.  

The invention is not based on radio transmission systems limits, but can also when using other transmissions delivery process z. B. acoustic method (ultrasound), ins special for the purposes of sonography or optical processes ren, for example infrared measurement based on Lidar principles be used. Another area of application is the Un investigation of changes in the spectral composition of backscattered signals.

The formation of signal sequences, their transmission and the calculation of correlation sums of these signal sequences with received signal sequences can be used in different technical fields:

• - For the purpose of synchronizing two transmission units ten, such as radio stations, in particular the Ver application of these sequences in the synchronization channel in CDMA Mobile radio systems, as is the case in standardization sensitive UMTS system,
• - During data transmission by means of the signal sequence spread transmission symbols or data in spread spectrum (spread spec trum) systems, in particular for determining transmission symbols or data to which such a signal form has been modulated de,
• - in measuring technology for distance and object measurement,
• - to determine the transmission properties of the between Transmission units, such as the sending unit and receiving unit lying transmission channel, in radar measurement technology, around the Location of an object and / or more from the geometry and depend on the specific reflection properties of the object to determine common parameters
• for determining transmission properties of the transmission channel located between the transmitter and receiver, in radar measurement technology for determining parameters of a backscattering medium, in particular the ionosphere, in particular by incoherent scattering,
  for determining transmission properties of the transmission channel lying between transmission units, such as transmitting unit and receiving unit, in particular for determining multipath propagation in measurement technology or communication technology. The correlation results during communication are used to determine the time-varying properties of the transmission channel (channel impulse response). In particular, additional paths of multipath propagation are determined. For this purpose, the signal sequences K (i) can also be transmitted in the form of a midamble within a radio block. This knowledge can then be used in an otherwise conventional receiving unit.

Claims (25)

1. Method for forming a signal sequence K (i) of length n, in which
the signal sequence K (i) is based on a first signal sub-sequence K1 (j) of length n1 and a second signal sub-sequence K2 (k) of length n2, wherein
the second signal sub-sequence K2 (k) is repeated n1 times and is modulated with the first signal sub-sequence K1 (j), and
at least one of the partial signal sequences is a go-lay sequence. 2. The method according to claim 1, wherein
the signal sequence K (i) has a length of 256, and
the signal sequence K (i) is based on a first signal sub-sequence K1 (j) of length 16 and a second signal sub-sequence K2 (k) of length 16, wherein
the second signal sub-sequence K2 (k) is repeated 16 times and is modulated with the first signal sub-sequence K1 (j). 3. The method as claimed in one of the preceding claims, in which the signal sequence K (i) is formed by modulating the second signal sequence K2 (k) in accordance with the following rule:
K (i) = K2 (i mod n2) * K1 (i div n2), for i = 0. . . n1 * n2-1. 4. The method as claimed in one of the preceding claims, in which at least one of the partial signal sequences is a Golay sequence X _n (k) of length nx, which can be formed by the following relationship:
X ₀ (k) = δ (k)
X ' ₀ (k) = δ (k)
X _n (k) = X _n-1 (k) + W _n .X ' _n-1 (kD _n )
X _n (k) = X _n-1 (k) - W _n .X ' _n-1 (kD _n ),
k = 0, 1, 2,. . ., 2 ^NX -1
n = 1, 2,. . ., NX
D _n = 2 ^Pn
With
nx = 2 ^NX
δ (k) Kronecker's delta function
P _n , n = 1, 2,. . ., NX, is any permutation of the numbers {0, 1, 2,. . ., NX-1} for the X sequence,
W _n weights for the X sequence (+1, -1, + i or -i). 5. The method according to any one of the preceding claims, wherein the permutation used to form a signal sequence P ₁ , P ₂ , P ₃ , P ₄ and unit size W ₁ , W ₂ , W ₃ , W ₄ following set of permutation unit size pairs (P ₁ P ₂ P ₃ P ₄ , W ₁ W ₂ W ₃ W ₄ ) is taken from:
0213, + j + j + j-1; 0213, -j + j + j-1; 0213, + 1-j + j-1; 0213, -1-j + j-1;
0213, + 1 + yy-1; 0213, -1 + yy-1; 0213, + yyy-1; 0213, -jjj-1;
0213, + j + j + j + 1; 0213, -j + j + j + 1; 0213, + 1-j + j + 1; 0213, -1-j + j + 1;
0213, + 1 + j-j + 1; 0213, -1 + j-j + 1; 0213, + yy-j + 1; 0213, -jj-j + 1;
3120, + 1-j + j-1; 3120, -1-j + j-1; 3120, + 1 + yy-1; 3120, -1 + yy-1;
3120, + 1 + j + j + j; 3120, -1 + j + j + j; 3120, + 1-j-j + j; 3120, -1-j-j + j;
3120, + 1 + j + jj; 3120, -1 + j + jj; 3120, +1-yyy; 3120, -1-yyy;
3120, + 1-j + j + 1; 3120, -1-j + j + 1; 3120, + 1 + j-j + 1; 3120, -1 + j-j + 1. 6. The method according to any one of claims 1 to 4, wherein the permutation used to form a signal sub-sequence P ₁ , P ₂ , P ₃ , P ₄ and unit size W ₁ , W ₂ , W ₃ , W ₄ following set of permutation unit sizes Pairs (P ₁ P ₂ P ₃ P ₄ , W ₁ W ₂ W ₃ W ₄ ):
3201, + 1-1 + 1 + 1; 3201, -1-1 + 1 + 1; 3201, + 1-1-1 + 1; 3201, -1-1- 1 + 1; 3201, + 1-1 + 1-1; 3201, -1-1 + 1-1; 3201, + 1-1-1-1; 3201, 1-1-1-1; 1023, + 1 + 1-1 + 1; 1023, -1 + 1-1 + 1; 1023, + 1-1-1 + 1; 1023, -1-1-1 + 1; 1023, + 1 + 1-1-1; 1023, -1 + 1-1-1; 1023, + 1-1-1-1; 1023, -1-1-1-1. 7. The method according to any one of claims 1 to 4, wherein the permutation used to form the signal sub-sequence P ₁ , P ₂ , P ₃ , P ₄ and unit size W ₁ , W ₂ , W ₃ , W ₄ following set of permutation unit sizes Pairs (P ₁ P ₂ P ₃ P ₄ , W ₁ W ₂ W ₃ W ₄ ):
3201, + 1-1 + 1 + 1; 3201, -1-1-1 + 1; 3201, -1-1 + 1-1; 3201, + 1-1-1-1, and
the permutation used to form the second signal sub-sequence (P ₁ P ₂ P ₃ P ₄ ) is equal to 3201. 8. The method according to any one of the preceding claims, in which the formation and / or transmission of the signal sequence K (i) to Purposes of synchronization of at least two transmissions units. 9. A method for determining a predetermined signal sequence K (i) contained in a received signal sequence E (l), which can be obtained by a method according to one of claims 1 to 7,
by determining the correlation sums S of the signal sequence K (i) with corresponding sections of the received signal sequence E (l), in which
a partial correlation sum sequence TS (z) of the signal partial sequence K2 (k) with corresponding parts of the received signal sequence E (l) is calculated, and
to calculate a correlation sum S n1, elements of the partial correlation sum sequence TS (z) are selected and multiplied by the signal partial sequence K1 (j) in the sense of a scalar product. 10. The method according to claim 9, wherein to calculate a correlation sum S n1 each n2-th Elements of the partial correlation sum sequence TS (z) selected become. 11. A method for determining a predetermined signal sequence K (i) contained in a received signal sequence E (l), which can be obtained by a method according to one of claims 1 to 7,
by determining the correlation sums S of the signal sequence K (i) with corresponding sections of the received signal sequence E (l), in which
a partial correlation sum sequence TS (z) of the signal partial sequence K1 (j) with selected elements of the received signal sequence E (l) is calculated, and
to calculate a correlation sum S n2 elements of the partial correlation sum sequence TS (z) in the sense of a scalar product are multiplied by the signal partial sequence K2 (k). 12. The method of claim 11, wherein to calculate a partial correlation sum TS n1 n2- te elements of the received signal sequence E (l) can be selected. 13. The method according to any one of claims 9 to 12, in which calculated partial correlation sums TS are stored and used to calculate a further correlation sum S. become. 14. The method according to any one of claims 9 to 13, in which at least one dot product using an Efficient Golay Correlator (EGC) is calculated. 15. The method according to any one of claims 9 to 14, in which at least a partial correlation sum or at least one Partial correlation sum sequence TS (z) by an Efficient Golay Correlator (EGC) is calculated. 16. The method according to any one of claims 9 to 15, in which at least one correlation sum S from the partial correlation total sequence TS (z) through an Efficient Golay Correlator (EGC) is calculated. 17. A method for synchronizing a base station (BS) with a mobile station (MS), in which
the base station detects a signal sequence K (i) obtainable by a method according to one of claims 1 to 8, and
the mobile station determines the signal sequence K (i) according to one of claims 9 to 16. 18. Sending unit (BS) with
Means (SPE) for storing a signal sequence K (i) obtained by a method according to one of claims 1 to 8, and
Means for transmitting this signal sequence K (i) for the purpose of synchronization with a receiving unit (MS). 19. Sending unit (BS) with
Means (SPE) for storing a permutation-one-size pair taken from a set of permutation-one-size pairs specified in claims 5 to 7,
Means for forming a signal sequence K (i) according to a method according to one of claims 1 to 8, and
Means for transmitting this signal sequence K (i) for the purpose of synchronization with a receiving unit (MS). 20. Receiver unit (MS) with
Means (SPE) for storing a permutation-one-size pair taken from a set of permutation-one-size pairs specified in claims 5 to 7,
Means for receiving a received signal sequence E (l), and
Means for determining a signal sequence K (i). 21. Receiver unit (MS) with
Means (SPE) for storing a pair of signal sequences, at least one signal sequence being a Golay sequence,
Means for receiving a received signal sequence E (l), and
Means for determining a signal sequence K (i). 22. receiving unit (MS) according to claim 21 with Means (SPE) for storing a pair of signal sequences, where at least one signal sub-sequence is a Golay sequence, use a permutation-one-size pair to form them bar which is a set of permutation unit sizes Is taken pairs, specified in claims 5 to 7 is. 23. receiving unit (MS) according to one of claims 20 or 22 With Means for determining a signal sequence K (i) according to one of the Claims 9 to 16. 24. receiving unit (MS) according to one of claims 20 to 23 With Means (SPE) for storing intermediate results (TS). 25. receiving unit (MS) according to one of claims 20 to 24 With two matched filters connected in series, called Ef ficient Golay correlators are trained.