FI106502B - Symbol synchronization method - Google Patents

Description

106502

TECHNICAL FIELD OF THE INVENTION - Förfarande for symbolsynkronisering

The invention relates to receiver structures, namely structures and methods for synchronizing a receiver with received data symbols.

BACKGROUND OF THE INVENTION

Continuous Phase Modulation (CPM) comprises a class of modulation methods that are simultaneously efficient in terms of both power and bandwidth. In all CPM modulations, the envelope of the RF signal is constant and the phase varies continuously. The standard envelope allows the use of nonlinear amplifiers, which simplifies receiver and transmitter design. TFM (Tamed Frequency Modulation) is a CPM modulation system. One of the major advantages of TFM is the very narrow bandwidth required compared to most other modulation systems. CPM signals can be represented by 15 s (t) = (2Eb / Τι>) 0 · 5οο5 (27 ^ 0ΐ + φ (ί> α) + φ0) (1) where the crossing function <p (t, ä) is given by «*. · · F. \: φ (ί, α) = 2π $ Σ αλ g (r- iTb) dx • '-00 i 4 | • · 1 -2n ^ iaihiq (t-iTb) i nTb <t <(n + l) Tb (2) / = o • · • · · * ♦ »* · • · ·: where ... {at} are M-th degree data symbols, M is an even pair of 20 character sets ± 1, ± 3, ...., ± (M-1), * · · · ..]. · Ht is a modulation index that can vary from j to one, • · # q (i) is a phase response function, · · «\ ''. g (t) is the frequency response, · ·

Eb is bit energy, 2 106502

Tb is the bit interval, fQ is the carrier frequency, and φ0 is the arbitrary start phase.

For TFM, M is 2, which makes the data symbols a ,. = ± 1, and h = 0.5. For a TFM, a 5 bit period is as long as a symbol period.

TFM modulation is characterized in that the phase shift of the modulated carrier over one bit period is determined not only by the current bit but by three consecutive binary signals according to the coding rule = (3) 10, where / „represents binary data at time t = nTb, and In = ± 1, i.e. for example -1 corresponds to bit 0, and +1 corresponds to bit 1. From this coding rule, one can see that the phase change is π / 2 if three consecutive bits have the same polarity, and the phase remains constant if the polarities of three consecutive bits alternate- VAT. The phase shifts π / 4 are related to the bit configurations ++ -, + -, - H and - +. Figure 15 a shows a signal state diagram of a TFM. When the correlation of the transmitted bits is included, the TFM signal produces a narrower power spectrum than, for example, the MSK signal, since phase changes are smoother in TFM. TFM modulation is explained in more detail in "Tamed Frequency Modulation, Novel: /: Method to Achieve Spectrum Economy in Digital Transmission", by: ··· Cornells B. Dekker, 16 Trans.on Comm. Vol COM-26, NO.5, May 1978, pp. 534 -542. CPM modulation is explained in more detail in “Digital Phase Modulation”, • *

John B. Anderson, Tor Aulin and Carl-Erik Sundberg, Plenum Publishing Corporation, 233 Spring Street New York, N.Y. 10013, pp. 15 - 53.

• · ·: Direct conversion receivers are receivers that do not use intermediate frequencies to filter and detect received signals. The radio frequency (RF) signal received at the direct conversion receiver is mixed with a local oscillator signal having a frequency corresponding to the carrier frequency of the RF signal. Direct conversion receivers have many advantages. For example, bandwidth filtering can be done at low frequencies, that is, audio frequencies, which allow very narrow bandwidths to be realized at the sharpest corners. Also, intermediate frequency filters are not required. However, direct conversion receivers have not been used to receive TFM signals due to the inherent problem of their direct conversion receiver structures 3 106502, namely the presence of a dc voltage deviation (DC) at the mixed output due to imperfections in the mixer structure. The DC offset is typically due to leakage of the local oscillator into the RF port of the mixer, and subsequent mixing of the leaked signal with the local oscillator 5 signal itself. random variations in the leakage signal lead to a relatively slow and randomly varying DC offset signal.

In addition to the DC offset problems inherent in direct downconverting applications, other types of low frequency noise often occur in the baseband signal. An example is low frequency phase noise due to transmitter phase noise or local oscillator phase noise. The problem of DC offset can be thought of as very low frequency phase noise.

Figure 2 illustrates the problem caused by DC offset in the detection of TFM modulated data. The signal received without DC deviation has a star pattern in white circles. The vectors Si and s2, plotted with thick dashed lines, 15 represent signals expressed at two consecutive sampling times. Without DC deviation, the expression of the vectors Si and s2 is straightforward. The DC deviation changes the situation significantly. The DC offset works by shifting the signal in a star pattern on the IQ chart as shown in black circles. The vectors s1 and s'2 represent the respective sampled signal vectors in the presence of a DC offset. Clear-20 stitch with any detector optimized for V ·; sex, there are difficulties in identifying the vectors s \ and s'2. Therefore, the occurrence of the DC- • * ♦ * · '«deviation easily leads to high error rates of expression.

* · «» M • «·: ··: The only known solution to this problem has so far been not to use a direct downconverter because this DC offset problem does not occur in 25 heterodyne receivers.

* · ♦ * Different coherent detectors or ··· ent differential detectors have been studied for TFM modulated signals. These detectors are based on signal phase or. : T: For phase difference over one bit period, however, do not work well with DC offset because the phase of a given point in the constellation changes due to DC offset • · -. 30 dos, and likewise the phase difference of two consecutive signal vectors changes due to the DC offset.

»* T · This issue is resolved for MSK-type modulation. WO94 / 28662 describes the structure of a dual differential detector, which is explained below. If rn is defined as a signal vector at time nTb, then the difference vector between two received vectors over one bit period is defined as K = rn ~ rn-1 (4) and the phase difference δφη between the vectors Δ „and ÄB_j is defined as 5 δφη = Ζ (Αη) -Ζ (Αη_1 ) (5) wherein the decision could be made as follows if 0 <δφη <135 °, then the transmitted bit bn = 1; if 135 ° <δφη <-135 °, then the transmitted bit bn! = &n-i; if -135 ° <δφ „<0, then the transmitted bit bn = 0.

This method works with MSK because in MSK the phase changes during each bit period.

The situation is slightly different when considering the idea of using a dual differential detector for TFM modulation. For TFM, δφη is not a sufficient measure to perform the decision since the difference vector An may be zero. As above . 15 explained, in TFM modulation, the phase does not change according to certain information bit patterns. Therefore, δφη may also be zero for some bit patterns, • · <'; '; which makes it impossible to make a decision on bits with a large δφ „base * ·" <support.

«

Symbol synchronization refers to adjusting the sampling timing of the receiver • · 20 when the received signal parameters, such as phase and amplitude, are at points in the constellation of the modulation. If the receiver timing is incorrect, ... the samples can be taken with distant constellation points, for example the constellation • · φ in Figure 1 is between points. In: "Simple Baseband Symbol Synchronization Scheme", · Bellini, Electronics Letters, March, 1997, vol *: ··: 25 30 No. 7, pp. 548-549, discloses a symbol synchronization method in which information about a timing error: •: · is taken from a phase or \ phase difference between two vectors received in close proximity. When a DC offset occurs, the star of the received signal will creep in *; * ·. away from the expected position that significantly affects the phase or phase difference of the signal over one bit period. Therefore, this method is 106502 5 sensitive to DC offset and cannot be used to synchronize TFM symbols on a receiver experiencing a DC offset problem.

SUMMARY OF THE INVENTION

It is an object of the invention to provide a symbol synchronization method for TFM signals, which also works well in the presence of DC offset and other types of low frequency noise.

The objectives are achieved by using as a synchronization variable a parameter which essentially depends on the phase difference of the sample vectors, such as the phase difference itself or the length of the difference vector between two received sample vectors, and searching for the average maximum of said pa-parameter.

The symbol synchronization method according to the invention is characterized by what is defined in the characterizing part of the independent method claim. The symbol synchronization system according to the invention is characterized in what is defined in the characterizing part of the independent claim of the symbol synchronization system. The mobile communication device according to the invention is characterized by what is defined in the characterizing part of the independent claim on the mobile communication device. The base station according to the invention is characterized in what is defined in the characterizing part of the independent claim on the base station. The radio link system according to the invention is characterized by the feature defined in the characterizing part of the independent patent claim for a radio link system. Other preferred embodiments of the invention are disclosed in the dependent claims.

• ♦ * ♦ ♦ * · * · * Since in most cases the DC offset changes relatively slowly, it can be assumed that the DC offset is constant over one bit period. Therefore, the effect of the DC offset can be eliminated by taking the difference between two closely received: T: vectors. In the following, rn (x) = r (nTb + τ) means the signal received at t = nTb + r, where Tb is the time length of the symbols, rn.} (X) = r ((nl) Tb + τ) means at time t = (nl) ) Tb + r received signal. Then • ·: ': Λ (τ) = rn (x) - rn.i {x) (6) • y: 30 represents the difference between the signals r »(x) and r« .j (x), where τ, 0 <τ <Tb means :: a timing deviation from nTb.

106502 6

When the DC offset signal rDC changes very slowly relative to the symbol length Tb, then rDC (nTb + x) «rDC ((n - V) Th + τ), 0 <τ <Tb, resulting in (r) = (r) + rDC > n (r)) - (rn_x (r) + rDW (r)) «r„ (r) - rn_x (τ) (7) Therefore, the difference signal Αη (τ) has no effect of DC deviation. The other in-5 data obtained from the difference vector also has no effect on the DC offset. In addition, the inventor found that the length of the difference vector, i.e., | Δη (t) | can provide the required timing error information. Namely, the quantity | Δ „(Υ) |, ie the quantity | Δ„ (γ) | the mean has the maximum at the right sampling time nTb, n = 0.1, 2, .... Therefore, in order to obtain the correct timing, a preferred embodiment of the invention employs a trace-10 loop of magnitude | for the maximum, or something else substantially equal to | Δ „(r) | dependent on retrieving other large extremes.

In a preferred embodiment of the invention, the average of the square of the length of the difference vector, i.e. | Δ „(Υ) | 2, is used as a variable to provide the correct timing. The invention is not limited to | Δ „(γ) | to find the maximum of an essentially dependent function. As an example, in another preferred embodiment of the invention, the function 1- | Δ „(γ) | / | Δ„ | minimum. In the previous equation | Δ «| max is for] Δ„ (τ) | maximum value. In various embodiments of the invention, any of the following can be used to find the correct symbol synchronization: · ί function, which essentially depends on | Δ „(γ) | and having an extreme value substantially • I 11 «« <I 1 ..... 1- i "· 20 when | Δ„ (τ) | has a maximum. In addition, in various embodiments of the invention. • · · based on a predetermined number of variables, i.e., using a window of predefined lengths to average • * · * *.

: * · *; In order to increase | Δ „(r) | the maximum is easier to find, the signal can be over: T: 25. Preferably, the received signal vector can be sampled at a frequency fs = r / Tb, where r is preferably an integer equal to or greater than 2.

• ·. For each sample, the difference vector is calculated by counting the sig- * on the one hand. "The difference between the nal vector and the previously sampled signal vector for the time period Tb. Figure 3 illustrates an example of a set of values where the values of-106502 7 are crossed, Figure 3 illustrates 11-fold oversampling, i.e., sampling at r-11. In one preferred embodiment, sampling times are adjusted according to the maximum mean time offset obtained.

In another preferred embodiment of the invention, the samples s (nTb - At) and s (nTb + At) 5 are taken with a time offset Δt before and after the sampling time nTb. The difference of the mean lengths of the respective difference vectors A „(nTb - At) and A„ (nTb + At) m / O) = K (nTb - At + τ) I - | Δ „(nTb + At + τ) \ (8) is used. to control the adjustment of the sampling time. In equation (8), the variable τ means the control of the sampling time. The value of τ is changed until the difference / (r) is zero. Finding the zero of the difference f (r) can be done in many ways and with many different search strategies. In fact, this method is equivalent to finding the zero for a large derivative of \ An (nTb + f) \. Many different search strategies for finding a zero for a derivative of a function are known to one of skill in the art and, therefore, such strategies are not described in further detail. The time offset Δt may preferably be Δt = Tb / 4.

It may be noted here that instead of averaging the lengths of the difference vectors as described above, the averaging can equally be performed by the results of the difference function. In other words, first we compute the equation 8 'for a use sample of • · <20: • «<·: ··: f (^) = \ A„ (nTb-At + T) \ - \ A „(nTb + At + T) \ (8 ') •' · · · '' and the results are averaged to give the result / (r).

• · · •

In another preferred embodiment of the invention, oversampling is not used. : T: Instead, the sampling time nTb is slightly modified, and the resultant. : T: 25 to | Δ „|, or some other substantially | Δ„ | dependent value, the change in value is monitored. If the resulting change is negative, the sampling moment is reversed. If the resultant change is positive, the sampling time is further modified in the same direction. If the resulting change is approximately zero, it is concluded that the sampling time is close enough: / ·· 30 optimum times. Of course, the thresholds defining how large a positive change is defined as a positive change, how large a negative change is defined as a negative change, and which values of the resulting 106502s changes are defined as approximately zero can be set in any desired manner so that the control method performance can be optimized. The optimum values for the limit values can be found, for example, by simulation or experimentation.

In another preferred embodiment of the invention, over-sampling position interpolation is used to provide calculated signal samples representing samples taken before and after the actual sampling time. According to such an embodiment, for example, compute the difference f (j) by interpolation using three consecutive difference vectors: f (O = I ch (kTb + τ) + dM (t +1) T „+1; | -1 (cä (kTb + τ) ; + mk - \) Th + t *) | (9) 10 where c can be, for example, 0.75, and d is 0.25 These values c and d correspond to At = 7 * / 4. as an example of linear interpolation, but in other embodiments of the invention any other type of interpolation and any other values for coefficients c and d can be used to obtain other values of A. For example, parabolic or third power interpolation may be used. needed because the bandwidth of TFM modulated signals is very small, roughly 0.5 / 7 *, so a sampling rate of 1/7 * is sufficient to meet the Nyquist condition.

«

The length of the resolution vector can be used in many ways for optimal symbol sync • <:. ' · I 20 to find the answer. In general, these methods correspond to finding the maximum length of the difference vector by varying the timing offset parameter.

·: · The difference vectors can be sampled with different values of the timing offset parameter: *:; oversampling and / or interpolating the sample values. These methods for finding optimal synchronization, for which some examples in the previous paragraphs are illustrated, can be divided into two main groups: feed-in-type methods and feedback-type methods.

• »» •: In feed-through type methods, a number of difference vector lengths are obtained with different values of the timing offset parameter. Thereafter, optimal timing can • •. to find which timing deviation gives the longest averages, or t *. 30 at which timing deviation values the derivative of the resulting distribution, that is, adjacent! * the differences between the mean values are closest to zero.

»« «

In feedback-type methods, the derivative approach, i.e. the difference approach, is most preferred. Methods of this type use the difference of the average lengths of the 106502 9 difference vectors at the two offsets used for observation to adjust the synchronization timing to bring the differences in the average lengths close to zero, i.e., to obtain the maximum difference between the average lengths of the difference vectors. 5 The difference vectors taken at the two observation time deviations can be obtained by oversampling and / or interpolation. Preferably double over-sampling is used, although other types of over-sampling can equally well be used.

When the DC offset or some other type of low frequency noise does not interfere too much with the signal, the phase difference between the sample sample vectors of the signal can be used to find its optimal synchronization timing in the same way as the length of the resolution vector. The correct symbol timing τ maximizes the mean of the expression (10): \ <f> (nTb + 7) - φ ((η -1) Tb + t) \ (10), or the mean of the phase change of the signal over the period Tb. In the previous embodiments teaching the use of lengths of difference vectors, phase differences can also be used.

In general, the inventive method maximizes or minimizes the value of a parameter that is substantially dependent on the phase difference of the sample vectors so that optimal synchronization is found. Preferably, the parameter is the phase difference itself.

«In case of DC offset or other low frequency noise, • t 'is preferred. j 20 uses the length of the difference vector as said parameter.

«« • · l

'·': BRIEF DESCRIPTION OF THE DRAWINGS

• 4 (11

• I

The invention will now be explained in more detail with reference to the accompanying drawings, in which: Fig. 1 illustrates the star pattern of a symbol for TFM modulation, «· ·«: ·: T: 25 illustrates the effect of a DC offset, Figure 3 shows a block diagram of a preferred embodiment of the invention, * *: Figure 5 shows a preferred embodiment of the invention applied to a receiver structure using the length of the difference vectors for detecting transmitted data, Fig. 6 is a block diagram of a preferred embodiment of the invention adjusting the sampling timing of the received signals, Fig. 7 is a block diagram of a mobile communication device according to the invention, Fig. 8 is a block diagram of a preferred embodiment of the invention a.

In the pictures, the same reference numerals are used for like entities. Figures 1, 2 and 3 were previously described.

DETAILED EXPLANATION

Fig. 4 shows a block diagram of a receiver structure according to a preferred embodiment of the invention. Such a receiver structure can be used, for example, in cellular networks and base stations of cellular networks. The receiver structure comprises a sampling section 200 and a synchronization section 400. The sampling section receives an RF signal and converts the received signal into baseband I and Q signals by means of se-210s, 210b. The local oscillator 205 produces a signal which is mixed with the received RF signal, and the phase shifter 206 provides a 90 ° phase shift to the local oscillator signal which is applied to the mixer 210b. The output signals of the mixers • · «201a and 210b are applied to matched filters 215a, 215b whose filtering characteristics are optimized for filtering TFM signals. The output signals of the filters 215a, 215b are applied to switching elements 225 which sample the I and Q signals in the 90 degree phase shift. Each of the I and Q samples: Y The corresponding pair defines a signal sample vector. The switching elements 225 are controlled by a sampling oscillator 220. The sampling portion 200 is an example of a structure that can be used in a receiver structure according to the invention. The invention is not limited to the use of structure 200 of the sampling section * shown in Fig. 4. Sampling is performed by switching switches *: ** · The elements 225 typically include analog / digital converters or other types of "" i sampling instruments.

The resulting I and Q samples are introduced into synchronization section 400, where the samples are introduced into interpolation blocks 430a, 430b. The interpolation block 430a, 430b produces calculated samples based on samples of the I and Q signals and performs the calculation based on the value of the control signal obtained from the loop filter block 344. The interpolated 10650211 I and Q signal samples are applied to a delay block 435a, 435b and to a summing block 440a, 440b, which delay and summing blocks calculate the difference between successive I and Q samples. The delay blocks 435a, 435b delay the I and Q signals by one T 1 period so that the differences are calculated from the samples taken at the corresponding moments of the symbol periods if oversampling is used. The difference results are applied to the difference vector length calculation block 445, which calculates the difference vector length corresponding to the differences of each I and Q sample pair. The calculated lengths are applied to the difference function calculation block 450, which may calculate the difference function f (j) according to, for example, equation (8 ').

The result is applied to a loop filter block 455, which performs averaging of differences 10. The / (Y) or equivalent value is then used to control the interpolation of the blocks 430a, 430b.

The embodiment of Figure 4 illustrates an example of how computation, i.e., interpolation, produces sample signal vectors that correspond properly to the sample vectors synchronized to the symbols. In some embodiments of the invention, properly synchronized sample vectors are obtained, not by counting, but by adjusting sampling times. An example of such an embodiment will be described later in this specification in connection with Figure 6.

Figure 4 further illustrates an example of combining symbol synchronization and demodulation to obtain the actual transmitted data. Figure 4 shows a demodulator. 20 blocks 410, which receive the interpolated I and Q sample vectors to their inputs, and which perform demodulation based on the interpolated sample vectors. The inputs of the demodulate * I block 410 may equally be taken from the outputs of the summing blocks 440a, · · 44b, whereby demodulation must be made on the basis of the received sample vectors * *.

Another advantageous way to demodulate the received data is to use the output of the difference vector length computation block 445. An example of such an embodiment is further explained in connection with Figure 5.

• ♦ · ♦ ♦ ♦ v; Figure 5 shows another preferred embodiment of the invention. In this embodiment, the output of the difference vector length calculation block 445 is used to determine the transmitted data 30. The lengths of the separation vectors are introduced into the Viterbi detector block 480, • «. which indicates the transmitted data based on the lengths of the difference vectors. The determination of the transmitted data on the basis of the lengths of the difference vectors is explained in more detail in the patent application entitled "Demodulation Method", the filing date and the applicant of which are the same as in the present patent application 10650212. The operation of the synchronization section 400 and other components of Figure 5 are explained in conjunction with Figure 4 and need not be explained in further detail here.

In another preferred embodiment, the difference calculations are performed before interpolation, i.e., the interpolation is performed for sample differences. In such an embodiment, the delay and summing blocks 435a, 440a; 435b, 440b would be placed in synchronization section 400 and interpolation blocks 430a; 430b between I and Q sample current inputs.

Fig. 6 illustrates another preferred embodiment of the invention in which the timing of the sampling oscillator 220 is adjusted due to symbol synchronization. The sampling oscillator 220 receives from the synchronization section 400 a control signal which adjusts the timing of the sampling oscillator 220. Figure 6 also shows an embodiment of the invention in which interpolation according to equation (9) is applied instead of oversampling. .

6, the I and Q signal samples from the sampling section 200 are applied to a delay block 435a, 435b and to a summing block 440a, 440b, which delay and summing blocks calculate the difference between 15 consecutive I and Q samples. The delay blocks 435a, 435b delay the I and Q signals by one TV cycle. The difference samples are introduced into calculation block 460, which calculates the early and late difference vector by interpolating from three consecutive samples, and which calculates the length of the early and late difference vector. In other words, the computation block computes the two terms of equation (9), i.e. the early difference vector 20 \ (cA (kTh + t) + dA ((k -1) Tb + r)) | length and late resolution vector \ (cA (kTb + r) + dA ((k + 1) Tb + r)) | length. The length of the early difference vector is output to I, and the length of the late difference vector is output to L, from which they are introduced to the summing block 465 which reduces the length of the early difference vector: V: the length of the late difference vector. The averaged result is applied to a sampling oscillator 220 for adjusting the sampling oscillator timing. 30, when the length of the early difference vector is less than the length of the late difference vector, the resulting control signal from the loop filter 455 may cause the sampling oscillator 220 to advance the sampling time. Of course, some signal processing, such as the scaling of the average result from the loop filter 455, may be required to obtain the control signal to adjust the oscillator 220 as desired. In the example of Figure 6, such signal processing is performed in the sampling oscillator block 220.

The other components of the sampling section 200 of Figure 6 and the operation of the sampling section 200 and the demodulation block 410 are described in connection with the other figures and need not be described in further detail herein.

Fig. 7 is a block diagram of a digital mobile communication device according to a preferred embodiment of the invention. The mobile communication device comprises a microphone 301, a keypad 307, a display 306, a headset 314, an antenna duplexer or switch 308, an antenna 309, and a control unit 305, all of which are typical components of conventional mobile communication lines. The mobile communication device further typically comprises transmission and reception blocks 304, 311. The transmission block 304 comprises the functions required for speech and channel coding, encryption and modulation, and the necessary RF circuits to amplify the signal for transmission. Receiver block 311 comprises the necessary amplifier circuits and functions needed to demodulate the signal and decrypt it, as well as for channel and speech decoding. The signal produced by microphone 301 is amplified in amplifier stage 302 and converted to digital form by analog / digital converter 303, after which the signal is applied to transmitter block 304. The transmitter block encodes the digital signal and produces a modulated and amplified RF signal, then the RF signal is applied to the antenna. 20 309 via duplexer or switch 308. The receiver block 311 demodulates the received signal and decrypts and channel encodes. The resulting speech signal is converted to analog form by a digital / analog converter 312, whereby the output signal is amplified by the amplifier 313, followed by the amplified signal: the control unit 305 controls the functions of the mobile communication device, reads the commands given by the user via the keypad 307, and displays messages to the user via the display 307. In the mobile communication device according to the invention, the mobile communication device comprises a synchronization part 400 400 is preferably the structure shown in Figure 4. However, other structures may be used for symbol synchronization 30 according to the invention.

• ·

Figure 8 shows an example of an embodiment of the invention. In the example of Figure 8, the synchronization part 400 according to the invention is used in at least some of the cellular networks. in base stations 360 for synchronizing * · * symbols received from mobile communication means 350. Furthermore, Figure 8 illustrates a base station controller 370 that controls base stations 360, 35, and two radio link units 371 for connecting base station controller 370 to other mobile communication network 380. Figure 8 also illustrates another preferred embodiment of the invention, namely the use of synchronization portions 400 in radio links. The demodulation method according to the invention is very advantageously used in continuous high speed communication, which requires continuous symbol synchronization, symbol synchronization tracking and data demodulation. High speed radio links are one example of such a preferred embodiment of the invention.

Fig. 9 is a block diagram of a preferred embodiment using phase difference instead of length of difference vectors as the decision variable. Figure 9 illustrates the structure of Fri-10 using the early / late port. The sensitivity of the timing error detector is the best, i.e. the error signal is greatest if the timing difference between the delayed and the advancing branch is the same as the minimum to maximum distance of the mean of the expression (10). Since the distance between the extremes is 7) / 2, one branch should be delayed by 7) / 4 and the other branch should be promoted by 7) / 4. In the case of TFM, one sample per bit 15 is sufficient for signal reconstruction: the delay 7), / 4 and the promotion 7), / 4 can be done by linear interpolation, even when the sampling frequency is only one sample per bit.

Since only phase differences or difference vector lengths over period 7) are used in the inventive method, the method can be used when a frequency error occurs.

Another advantage is that the method can be used without knowing bit decisions.

4 I

· '* ·.; Although the present invention allows symbol synchronization in the presence of DC offset and other types of low frequency noise, the inventive symbol synchronization method can also be used in conventional TFM receiver structures where the amount of DC offset and low frequency noise is not disturbing. great. In addition, while the above examples use the difference between successive received sample vectors to achieve synchronization, the difference in length over two successive bits of * can also be used to achieve synchronization. Similarly, while the previous description describes synchronization to a TFM modulated signal, the inventive synchronization method can equally be used to provide synchronization to MSK, GMSK, and GTFM ·: ··: modulated signals. In particular, the invention may be applied to Viterbi type *. > TFM receivers for digital microwave radio links.

In the following claims, the term sample vector refers to a pair of corresponding I and Q signal samples.

106502 15

In light of the above description, it will be apparent to one skilled in the art that various modifications may be made within the scope of the invention. Although a preferred embodiment of the invention is described in detail herein, it should be apparent that many modifications and modifications can be made, all within the real spirit and scope of the invention.

• · «• <· 4 4 4 4 4 4 4 f 4 4 4 f 4 4 t 4 4 f

4 I

4 • • • 4 • • • • • • 4 4 • • • • • 4 4 • 4 • 4 4 * V4t4 4 4

Claims (17)

A method of symbol synchronization for synchronization of the timing of the symbols of a TFM modulated signal, characterized in that the method comprises the steps 5. to obtain sample vectors from the signal according to a controlling parameter, - to calculate no mentioned sample vectors parameters which are essentially dependent on the phase difference. of said sample vectors, - to calculate a function substantially dependent on said parameters, and - to adjust said controlling parameter to obtain an extreme value of said function. Method according to claim 1, characterized in that said parameter is the phase difference between successive sample vectors. Method according to claim 1, characterized in that said parameter is the length of a difference vector which is calculated as the difference between two sample vectors. Method according to claim 3, characterized in that said differential vectors are calculated in consecutive sample vectors. . Method according to claim 1, characterized in that said function comprises'; '' Calculating the average of these parameters. • «* v» t; * ·,! 20 6. Symbol synchronization system for synchronization to the symbols of a TFM modulated signal, characterized in that the system comprises - means for producing signal sample vectors from a TFM modulated signal according to a control signal, means for calculating parameters which is substantially dependent on the phase difference ... of said sample vectors, means for calculating a function substantially dependent on said parameters, and * means for producing said control signal. to adjust the production of signal sample vectors to give the company an extreme value of this function. y 30 7. A system according to claim 6, characterized in that the system further comprises: means for calculating difference vectors from said sample vectors, and said means for calculating parameters being arranged to calculate the length of said difference vectors. 10650 / System according to claim 7, characterized in that said means for calculating the value of a function calculates the average of the length of said difference vectors. Mobile communication means, characterized in that it comprises a system for symbol synchronization comprising means for producing signal sample vectors from a TFM-modulated signal according to a control signal, - means for calculating parameters which are substantially dependent on the phase difference of said sample vectors, 10. means for calculating a function substantially dependent on said pair of meters, and - means for producing said control signal for adjusting the output of signal sample vectors to find an extreme value of said function. Mobile communication means according to claim 9, characterized in that said symbol synchronization system further comprises means for calculating difference vectors from said sample vectors, and that said means for calculating parameters is arranged to calculate the length of said difference vectors. Mobile communication means according to claim 10, characterized in that said means for calculating the value of a function calculates the average of the length of said difference vectors. 12. Base station in a mobile communication network, characterized in that it comprises a system for symbol synchronization which comprises means for producing signal sample vectors from a TFM-modulated signal according to ***** a control signal, ♦ means for controlling calculate parameters which are substantially dependent on the phase difference of said sample vectors, ··· *: - means for calculating a function substantially dependent on said para: *: *: meters, and - means for producing said control signal for to adjust the output of signal • ·. 30 sample vectors to find an extreme value of said function. • «.j. Base station according to claim 12, characterized in that said symbol synchronization system further comprises means for calculating difference vectors from said sample vectors, and that said means for calculating parameters is arranged to calculate the length of said difference vectors. 106502 Base station according to claim 13, characterized in that said means for calculating the value of a function calculates the average of the length of said difference vectors. Radio link system, characterized in that it comprises a system for symbol synchronization which comprises means for producing signal sample vectors from a TFM modulated signal according to a control signal, - means for calculating parameters which are substantially dependent on the phase difference of said sample vectors, means for calculating a function substantially dependent on said pair of meters, and - means for producing said control signal for adjusting the production of signal sample vectors to give an extreme value of said function. Radio link system according to claim 15, characterized in that said symbol synchronization system further comprises means for calculating difference vectors from said sample vectors, and that said means for calculating parameters is arranged to calculate the length of said difference vectors. Radio link system according to claim 16, characterized in that said means for calculating the value of a function calculates the average of the length of said 411 difference vectors. (<* CI <• · • · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·