DE102014016446A1 - Method for correcting a symbol clock frequency and phase of a data signal modulated on a carrier wave by means of a continuous phase shift keying - Google Patents

Abstract

The invention relates to methods of correcting a symbol clock frequency and phase of a data signal modulated on a carrier wave by a continuous phase shift keying in a wireless data transmission between a transmitter and a receiver by means of an incoherent block detector, comprising the steps of: receiving the one of a plurality of successive ones Data blocks formed by the data receiver signal generator, generating an identical data signal to another data signal, which is formed from a plurality of identical to the first data blocks second data blocks, filtering the data signal by means of a first filter, and filtering the further data signal by means of a second different from the first filter Filtering, selecting a filtered first data block and a filtered second data block corresponding thereto, wherein the filtered first data block is comprised of a plurality of first symbols and the filtered second data block lock is formed of a plurality of second symbols, forming time-derived first symbols, or forming first sums of time-derived first symbols, forming conjugate complex second symbols, or forming second sums of conjugate complex second symbols, multiplying each time-derived first symbol, or each first sum, each having a weighting factor, wherein weighted time-derived first symbols, or weighted first sums are formed, multiplying each conjugate complex second symbol, or each second sum, each with a weighting factor, where weighted conjugate complex second symbols, or weighted second sums multiplying each weighted time-derived first symbol, or each weighted first sum, by each weighted conjugate complex second symbol, or with each weighted second sum, forming products summing each one of the second Real part of the products, wherein a resulting third sum is a measure of the symbol clock phase error e, and continuous correction of the symbol clock phase using the symbol clock phase error e and concomitant correction of the symbol clock frequency.

Description

The invention relates to a method for correcting a symbol clock frequency and phase of a data signal modulated on a carrier wave by means of a continuous phase shift keying.

According to the prior art, such methods are well known, for example from the publication "Timing recovery based on the PAM representation of CPM" by E. Perrins, S. Bose and M. Wylie-Green, published in the IEEE Military Communications Conference, November 2008, pages 1-8 , The known method is suitable for correcting a data signal, which is transported by means of a carrier wave of known frequency and phase. However, if the frequency and / or the phase of the carrier wave change, ie, if the transmission channel is time-variant, the known method no longer yields useful results.

A time-variant transmission channel results, for example, in a relative movement of the receiver to the transmitter. For example, the receiver may be in a mobile phone that is traveling at high speed in a car, train, or airplane. The receiver can also be located in a remotely controllable projectile. In addition, the temporal variability of the transmission channel may be due to a vibration of the transmitter or receiver. A temporal variance of the transmission channel can, for. B. also occur when the data signal is reflected in the transmission.

The object of the invention is to eliminate the disadvantages of the prior art. In particular, a method for correcting a symbol clock frequency and phase of a data signal modulated on a carrier wave is to be specified, with which error-free reception is possible even with a time-variant transmission channel.

This object is solved by the features of claim 1. Advantageous embodiments of the invention will become apparent from the features of claims 2 to 10.

According to the invention, the following steps are proposed for correcting a symbol clock frequency and phase of a data signal modulated on a carrier wave by means of a continuous phase shift keying in a wireless data transmission between a transmitter and a receiver by means of an incoherent block detector:
Receiving the data signal formed by a plurality of successive first data blocks by the receiver,
Generating an identical to the data signal further data signal, which is formed from a plurality of identical to the first data blocks second data blocks,
Filtering the data signal by means of a first filter, and filtering the further data signal by means of a second filter different from the first filter,
Selecting a filtered first data block and a filtered second data block corresponding thereto, wherein the filtered first data block is formed from a multiplicity of first symbols and the filtered second data block is formed from a multiplicity of second symbols,
Forming time-derived first symbols, or forming first sums of time-derived first symbols,
Formation conjugates complex second symbols, or formation of second sums of complex second conjugated symbols,
Multiplying each time-derived first symbol, or each first sum, each with a weighting factor, whereby weighted time-derived first symbols, or weighted first sums are formed,
Multiplying each conjugate complex second symbol, or every other sum, each with a weighting factor, whereby weighted conjugate complex second symbols, or weighted second sums are formed,
Multiplying each weighted time-derived first symbol, or each weighted first sum, by each weighted conjugate complex second symbol, or by each weighted second sum, thereby forming products,
Summing each of the real part of the products, a resulting third sum being a measure of the symbol clock phase error e, and
continuous correction of the symbol clock phase using the symbol clock phase error e and concomitant correction of the symbol clock frequency.

The "continuous phase shift keying" is a modulation method used in data transmission. For the continuous-phase-Umtastung also the name CPM (English for continous phase modulation, ie Phasenumtastung with steady phase) is common. Unlike other digital phase modulation techniques, where the carrier phase can jump abruptly at the start of each symbol, the carrier phase is steadily continued during the continuous phase shift keying. The continuous phase characteristic in the continuous-phase shift keying leads to an improved spectral efficiency compared to the other methods.

In particular, the data transmission can take place by means of one of the following CPM methods: CPFSK (Continuous Phase Frequency Shift Keying), MSK (Minimum Shift Keying), GMSK (Minimum Gaussian Shift Keying), SOQPSK (Shape-Offset Quaternary Phase-shift keying) with the embodiments SOQPSK-TG (SOQPSK Telemetry Group according to IRIG 106) or SOQPSK-MIL (SOQPSK according to MIL-STD 188-181), ARTM Tier 0 / I / II (Advanced Range Telemetric Animal 0 / I / II) and PCM / FM (English for pulse code modulation / frequency modulation). A multi-h partial-response CPM method can be used. In contrast to this, a full-response and / or single-h CPM method can also be used.

An "incoherent block detector" is a detector which detects several symbols at the same time. The simultaneously captured symbols form a data block. The incoherent block detector needs no information about the phase of the carrier wave. It is sufficient a rough information about their frequency. The incoherent block detector includes an incoherent block-timing-error-detector (Block-TED) and a data detector. The incoherent block error detection detector detects the symbol clock phase error e.

A "symbol" is understood to mean a pulse containing one or more bits of information content.

In the formation of first sums and in the formation of second sums, a plurality of time-derived first symbols or a plurality of complex conjugate second symbols are combined as summands. Conveniently, all first sums and all second sums are formed from the same number of summands. Expediently, a first sum corresponds to each first sum and a first sum to every second sum, wherein both sums each have corresponding summands. Thus, several symbols are correlated with each other. This correlation is expediently carried out by a filter designed as a coherent partial correlator. A coherent fractional correlator allows a gradual transition from incoherent detection of the data signal to coherent detection of the data signal. This allows a trade-off between the benefits of coherent detection and the benefits of incoherent detection. On the one hand, robustness against noise is better for coherent detection. On the other hand, the incoherent detection is more robust against changes in the transmission channel.

With the proposed method, even with a strongly time-variant transmission channel, the symbol clock frequency and the symbol clock phase of a data signal modulated on a carrier wave can be corrected, so that error-free reception is possible.

According to the invention, the following steps are proposed for correcting a symbol clock frequency and phase of a data signal modulated on a carrier wave by means of a continuous phase shift keying in a wireless data transmission between a transmitter and a receiver by means of an incoherent block detector as an alternative to the already proposed method according to the invention.
Receiving the data signal formed by a plurality of successive first data blocks by the receiver,
Filtering the data signal by means of a first filter,
Generating a further filtered data signal which is identical to the filtered data signal and which is formed from a multiplicity of filtered second data blocks identical to the filtered first data blocks,
Selecting a filtered first data block and a filtered second data block corresponding thereto, wherein the filtered first data block is formed from a multiplicity of first symbols and the filtered second data block is formed from a multiplicity of second symbols,
Forming time-derived first symbols, or forming first sums of time-derived first symbols,
Formation conjugates complex second symbols, or formation of second sums of complex second conjugated symbols,
Multiplying each time-derived first symbol, or each first sum, each with a weighting factor, whereby weighted time-derived first symbols, or weighted first sums are formed,
Multiplying each conjugate complex second symbol, or every other sum, each with a weighting factor, whereby weighted conjugate complex second symbols, or weighted second sums are formed,
Multiplying each weighted time-derived first symbol, or each weighted first sum, by each weighted conjugate complex second symbol, or by each weighted second sum, thereby forming products,
Summing each of the real part of the products, a resulting third sum being a measure of the symbol clock phase error e, and
continuous correction of the symbol clock phase using the symbol clock phase error e and concomitant correction of the symbol clock frequency.

In contrast to the method according to the first alternative, the filtering in the method according to the second alternative takes place before the generation of the further data signal. The method according to the second alternative differs from the method according to the first alternative in that the data signal and the further data signal pass through the same filters. The method according to the second alternative requires only half of the filter operations compared to the method according to the first alternative. The method according to the second alternative is particularly simple and fast.

According to an advantageous embodiment of the invention, the first filter is part of a first filter bank, and / or the second filter is part of a second filter bank. A filter bank generally comprises several filters. Expediently, the first filter bank and the second filter bank each have an equal number of filters.

According to a further advantageous embodiment of the invention, it is provided that the respective impulse responses of the first filter are determined by first pulses of an AMP decomposition of the continuous phase shift keying and / or the respective impulse responses of the second filter by second pulses of an AMP decomposition of the continuous one Phase-shift are fixed. The AMP decomposition of the continuous-phase shift keying takes place according to the publication "Timing recovery based on the PAM representation of CPM" by E. Perrins, S. Bose and M. Wylie-Green, published in the IEEE Military Communications Conference, November 2008, pages 1-8 , and the publications cited therein. PAM (English for pulse-amplitude modulation) is another name for AMP (English for amplitude modulated pulse).

Pulses are generally unbalanced and complex valued. The respective impulse response is usually the time reversed and conjugated complexes of the respective pulse.

Furthermore, several or all weighting factors may be the same. The weighting factors are generally complex. In particular, the weighting factors of mutually corresponding temporally derived first symbols and conjugate complex second symbols may be equal to one another or complex conjugates to one another.

According to a further advantageous embodiment of the invention

The index k traverses the first symbols in their chronological order, assigning a discrete time to each first symbol, the index κ traversing the second symbols in their chronological order, assigning each second symbol a discrete time, the index 1 traversing the first one filter the first filter bank, the index λ passes through the second filters of the second filter bank, L denotes the partial response length, L ₀ denotes the block length in symbols, R denotes both the number of first filters in the first filter bank and the number of second ones Filter in the second filter bank, b * / l, k and b _{λ, κ} denote the respective weighting factors, ẋ _{l, k} denotes the time-derived first symbols of the l-th first filter of the first filter bank at time k, and x _{λ, κ} denotes the second symbols of the λ-th second filter of the second Filterbank at time κ.

According to a further advantageous embodiment of the invention, each first pulse is divided into a plurality of first pulse sections and every other pulse into a plurality of second pulse sections, wherein

The index k traverses the first symbols in their chronological order, assigning a discrete time to each first symbol, the index κ traversing the second symbols in their chronological order, assigning each second symbol a discrete time, the index 1 traversing the first one Filter of the first filter bank, the index λ passes through the second filter of the second filter bank, the index m passes through the first pulse sections, the index μ passes through the second pulse sections, L ₀ denotes the block length in symbols, R denotes both the number of first filters in the first filter bank as well as the number of second filters in the second filter bank, P designates both the number of first pulse sections and the number of second pulse sections, b * / l, k + m and b _{λ, k + μ} denote the respective weighting factors, ν. _{l, m, k} denotes the time-derived first symbols of the l-th first filter of the first filter bank for the m-th first pulse section at the time k, and ν * / λ, μ, κ denotes the conjugate complex second symbols of the λ-th second filter of the second filter bank for the μ-th second pulse section at the time κ.

The number of the respective pulse sections P results from the AMP decomposition. It is preferably greater than or equal to the partial response length L, ie P ≥ L. Particularly preferably, the number of the respective pulse sections P is slightly larger than the partial response length L, in particular P = L + 1.

According to a further advantageous embodiment of the invention, the weighting factors are obtained from the AMP decomposition.

According to a further advantageous embodiment of the invention, each real part of products is multiplied prior to summation each with a forgetting factor which is exponentially dependent on the time interval of the symbols multiplied together. A forgetting factor is expediently formed from a real number.

According to a further advantageous embodiment of the invention, the data signal is transmitted by means of the transmitter using a spectral spreading method.

The partial response length L is preferably 1 to 10. Particularly preferred values of L are values of 1 to 3. The block length in symbols L ₀ is preferably 1 to 100. For L ₀ , values of 3 to 5 are particularly preferred. The number of filters in the filter bank R is preferably 1 to 20. Particularly preferred values of R are values of 1 to 10.

Embodiments of the invention will be explained in more detail with reference to drawings. Show it:

1 schematically illustrated synchronization circuit for correcting the symbol clock phase by means of a feedback,

2 schematic flowchart for determining the symbol clock phase error e according to a first embodiment,

3 schematic flow diagram for determining the symbol clock phase error e according to a second embodiment and

4 schematic representation of multiplications carried out by means of a cross multiplier.

1 shows a schematically illustrated synchronization circuit for correcting the symbol clock phase by means of a feedback. An antenna 1 receives a carrier wave with a modulated data signal. A demodulator 2 demodulates the data signal. The reference number 3 denotes a rough synchronization of the carrier wave frequency. For a frame synchronization 4 a start of a data transfer frame is detected. An interpolator 5 corrects the symbol clock phase and supplies the corrected data signal to an error detection detector (= TED) 6 , The processes within the TED 6 be in 2 explained in more detail. Through the TED 6 a symbol clock phase error e is determined. The symbol clock phase error e becomes a loop filter 7 fed. The output side is the loop filter 7 with a numerically controlled oscillator (= NCO) 8th connected. From the NCO 8th off, the symbol clock phase in turn becomes the interpolator 5 fed. The interpolator 5 then correct the symbol clock phase once more. The NCO 8th also gives - illustrated by the direct connection to the TED 6 - the clock phase of the decimation before. Due to the feedback with continuous correction of the symbol clock phase, a correction of the symbol clock frequency is achieved at the same time.

2 shows schematically the processes within the TED 6 for determining the symbol clock phase error e according to a first embodiment. In the TED 6 a data signal is received. In the in 2 shown first embodiment, the filtering takes place before generating a further data signal. Therefore, in 2 no second filter bank provided with second filters. The TED 6 includes a first filter bank 10 For example, with four first filters, so it is R = 4. The data signal is therefore split into four paths. The reference number 11 denotes the impulse response c * / 0 (-t) of the first filter with index l = 0. The impulse response c * / 0 (-t) is the temporally inverse and conjugate complexes of a corresponding first pulse of an AMP decomposition of the continuous phase shift keying. c * / 1 (-t) . c * / 2 (-t) and c * / 3 (-t) are the impulse responses of the other first filter with the indices l = 1, l = 2 and l = 3, respectively. The reference numeral 12 denotes a subsequent step in which a downsampling of the impulse response c * / 0 (-t) done on the symbol clock. At the exit of the downsampling step 12 the symbols x _{0, k of} the first filter with index l = 0 are passed on at time k. In 2 For example, the time k = 2 is considered. Accordingly, a downsampling of the impulse responses takes place in each case c * / 1 (-t) . c * / 2 (-t) and c * / 3 (-t) to x _1.2 , x _2.2 and x _3.2 respectively. The reference number 13 denotes a split of the path at x _0.2 into two subpaths. In the first part path becomes at the step 14 the time-derived first symbol ẋ _{0,2 of} the 0th first filter of the first filter bank 10 formed at the time k = 2. In the second part path is at step 15 the conjugate complex second symbol x * / 0.2 of the 0th first filter of the first filter bank 10 formed at the time k = 2. Analogously, a splitting of the paths to x _1.2 , x _2.2 and x _{3.2 takes place} in two subpaths. In the respective first partial paths, a time derivative of x _1,2 , x _2,2 and x _3,2 respectively takes place and in the respective second partial paths the conjugated complexes of x _1,2 , x _2,2 and x _{3, 2} formed. The reference number 16 denotes the subsequent step in which the time-derived first symbol ẋ _{0.2 of} the 0th first filter at time k = 2 is multiplied by a weighting factor b _0.2 . The conjugate complex second symbol x * / 0.2 of the 0th first filter at time k = 2 is compared with the conjugate complexes of the same weighting factor b _0.2 , ie with b * / 0.2, multiplied (reference numeral 17 ). Analogously, in the two partial paths to x _1.2 a multiplication by b _1,2 or the conjugated complexes of b _1,2 , in the two partial paths to x _2,2 with b _2,2 and the conjugated complexes of b _2,2 and the two partial paths to x _3,2 with b _3,2 and the conjugated complexes of b _3,2 . The weighting factors b _{l, k} are available as results of the AMP decomposition. Then the values of the first partial paths are added up (reference numeral 18 ). The result of the addition is u. ₂ . The values of the second subpaths are also added up (reference numeral 19 ). The result of the addition is here u * / 2.

This in 2 illustrated first embodiment does not include a coherent Teilkorrelator. In exemplary embodiments not illustrated here, which comprise a coherent partial correlator, a sum of values of u. _{k are formed} at a certain number of successive times k.

Accordingly, in each of these embodiments, a sum of values of u * / k in each case to form the same successive times κ.

The first filter bank 10 has a first exit 10a and a second exit 10b , The exits 10a . 10b the first filter bank 10 are with inputs 20a . 20b a cross multiplier 20 connected. It reaches u. ₂ over the first exit 10a the first filter bank 10 to the first entrance 20a of the cross multiplier 20 and u * / 2 over the second exit 10b the first filter bank 10 to the second entrance 20b of the cross multiplier 20 , The cross multiplier 20 includes a memory. In the memory of the cross multiplier 20 already values are stored at the previous times. For a simplified representation, only the values associated with the times k = 0 and k = 1 are considered below, namely u. ₀ , u * / 0, u. ₁ and u * / 1. In addition, u. ₂ and u * / 2 in the memory of the cross multiplier 20 saved. The memory of the cross multiplier 20 includes a line memory and a column memory. The values of u. ₀ , u. ₁ and u. ₂ , for example, in the line memory and the values of u * / 0, u * / 1 and u * / 2 stored in the column memory.

As in 4 is illustrated, each value of the line memory is multiplied by each value of the column memory. In the embodiments illustrated in the figures, the number of stored values is three. Thus, nine multiplications are carried out and nine products are formed. For each multiplication has the cross multiplier 20 one output each. The cross multiplier 20 has in the illustrated embodiments according to nine outputs. From the cross multiplier 20 output products then the real part is formed (reference numeral 21 ). The real parts of the products are finally added up (reference numeral 22 ). As a result, this addition provides the symbol clock phase error e.

3 shows schematically the processes within the TED 6 for determining the symbol clock phase error e according to a second embodiment. In the TED 6 a data signal is received. In the in 3 In the second embodiment shown, the filtering takes place after the generation of an identical further data signal. The reference number 29 denotes a splitting of the data signal, wherein the identical further data signal is generated. The data signal is transferred to a first filter bank 30 and the further data signal in a second filter bank 40 guided. The first filter bank 30 includes, for example, four first filters and the second filter bank 40 four second filters, so it is R = 4. The process within the first 30 and the second filter bank 40 is essentially the same as in 2 shown first embodiment. The data signal and the further data signal are each split into four paths. The reference numerals 31 and 41 denote the impulse response c A * / 0 (-t) of the first filter with index l = 0 or the impulse response c B * / 0 (-t) of the second filter with index l = 0. The terms A and B are here to better distinguish the first 30 and the second filter bank 40 serve. c A * / 1 (-t) . c A * / 2 (-t) and c A * / 3 (-t) are the impulse responses of the other first filters with the indices l = 1, l = 2 and l = 3, respectively. c B * / 1 (-t), c B * / 2 (-t) and c B * / 3 (-t) are the impulse responses of the other second filters with the indices λ = 1, λ = 2 and λ = 3. In particular, corresponding impulse responses of the first 30 and second filter bank 40 be the same, so it can apply c A * / 0 (-t) = c B * / 0 (-t) ≡ c * / 0 (-t), c A * / 1 (-t) = c B * / 1 (-t) ≡ c * / 1 (-t), c A * / 2 (-t) = c B * / 2 (-t) ≡ c * / 2 (-t) and c A * / 3 (-t) = c B * / 3 (-t) ≡ c * / 3 (-t). The reference numerals 32 and 42 denote the subsequent step of downsampling the impulse response c A * / 0 (-t) respectively. c B * / 0 (-t) on the symbol clock. Also in 3 By way of example, the time k = κ = 2 is considered. The same happens in the first filter bank 30 a downsampling of the impulse responses c A * / 0 (-t), c A * / 1 (-t), c A * / 2 (-t) and c A * / 3 (-t) at x A / 0,2, x A / 1,2, x A / 2,2 respectively. x A / 3,2 and in the second filter bank 40 a downsampling of the impulse responses c B * / 0 (-t), c B * / 1 (-t), c B * / 2 (-t) and c B * / 3 (-t) to x B * / 0,2 (-t), x B * / 1,2 (-t), x B * / 2,2 (-t) and x B * / 3,2 (-t). In contrast to the first embodiment, there is now no splitting into first and second subpaths within the filter banks. In the first filter bank 30 becomes at the step 34 the temporally derived first symbol / A / 0.2 of the 0th first filter at time k = 2. Analog takes place in the other paths of the first filter bank 30 a time derivative of x A / 1,2, x A / 2,2 and x A / 3,2. In the second filter bank 40 becomes at the step 45 the conjugate complex second symbol x B * / 0.2 of the 0th second filter at the time κ = 2. Analog is in the other paths of the second filter bank 40 the conjugated complexes of x B / 1,2, x B / 2,2 and x B / 3,2 educated. The reference number 36 indicates the subsequent step in which the time-derived first symbol / A / 0.2 of the 0th first filter at time k = 2 with a weighting factor b A / 0.2 is multiplied. The reference number 47 analogously denotes the step in which the conjugate complex second symbol x B * / 0.2 of the 0th first filter at time κ = 2 with the conjugate complexing of a weighting factor b B / 0.2, so with b B * / 0.2, is multiplied. The weighting factor b A / 0.2 and the weighting factor b B / 0.2 can be different from each other. In the further paths of the first filter bank 30 done according to multiplications with the weighting factors b A / 1,2, b A / 2,2 and b A / 3,2. In the further paths of the second filter bank 40 done according to multiplications with the conjugate complexes of the weighting factors b B / 1,2, b B / 2,2 and b B / 3,2. The weighting factors b A / l, k and the weighting factors b B / λ, κ are available as results of the AMP decomposition. In particular, corresponding weighting factors can be the same, so it can apply b A / 0.2 = b B / 0.2 ≡ b _0.2 , b A / 1,2 = b B / 1,2 ≡ b _1,2 , b A / 2,2 = b B / 2,2 ≡ b _2,2 and b A / 3.2 = b B / 3.2 ≡ b _3.2 . Finally, the values of the paths of the first filter bank 30 added up (reference numeral 38 ). The result of the addition is u. ₂ . Also the values of the paths of the second filter bank 40 are added up (reference numeral 49 ). The result of the addition is here u * / 2. The first filter bank 30 has a first exit 30a and the second filter bank 40 a second exit 40b , It reaches u. ₂ over the first exit 30a the first filter bank 30 to the first entrance 20a of the cross multiplier 20 and u * / 2 over the second exit 40b the second filter bank 40 to the second entrance 20b of the cross multiplier 20 , For further steps, please refer to the description 2 directed.

Also in 3 illustrated second embodiment does not include a coherent Teilkorrelator. However, the second embodiment can also be modified accordingly by each a sum of values of u. _k at a certain number of successive times k and a sum of values of u * / κ is formed in each case at the same successive times κ. These sums are then stored in the cross-multiplier and multiplied together.

4 shows a schematic representation of the through the cross multiplier 20 performed multiplications. For the simplified representation, the line memory and the column memory each have three values, namely one value for each of the times k = 0, k = 1 and k = 2 2 and 3 illustrated embodiments are in the line memory the values u. ₀ , u. ₁ and u. ₂ and in the column memory the values u * / 0, u * / 1 and u * / 2 saved. In the in 4 The selected representation becomes the values belonging to the different times k = 0, k = 1 and k = 2 using delay elements 51 tapped. At each of the nine crossing points 52 a multiplication of the respective values is performed. The products obtained at the nine multiplications are tapped and then - as to the 2 and 3 explained - whose real parts added up.

LIST OF REFERENCE NUMBERS

1

antenna

2

demodulator

3

coarse synchronization of the carrier wave frequency

4

frame synchronization

5

interpolator

6

TED

7

loop filter

8th

NCO

10, 30

first filter bank

10a, 30a

first exit

10b, 40b

second exit

11, 31, 41

impulse response

12, 32, 42

downsampling

13, 29

breakdown

14, 34

time derivation

15, 45

complex conjugation

16, 17, 36, 47

Multiplication with weighting factor

18, 38

Add up the values of the first paths

19, 49

Add up the values of the second paths

20

cross multiplier

20a

first entrance

20b

second entrance

21

Formation of the real parts of the products

22

Add up the real parts of the products

40

second filter bank

51

delay elements

52

crossing points

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited non-patent literature

• "Timing recovery based on the PAM representation of CPM" by E. Perrins, S. Bose and M. Wylie-Green, published in the IEEE Military Communications Conference, November 2008, pages 1 to 8 [0002]
• "Timing recovery based on the PAM representation of CPM" by E. Perrins, S. Bose and M. Wylie-Green, published in the IEEE Military Communications Conference, November 2008, pages 1 to 8 [0016]

Claims (10)

Method for correcting a symbol clock frequency and phase of a data signal modulated on a carrier wave by means of a continuous phase shift keying in the case of wireless data transmission between a transmitter and a receiver by means of an incoherent block detector comprising the following steps: Receiving the data signal formed by a plurality of successive first data blocks by the receiver, Generating an identical to the data signal further data signal, which is formed from a plurality of identical to the first data blocks second data blocks, Filtering the data signal by means of a first filter, and filtering the further data signal by means of a second filter different from the first filter, Selecting a filtered first data block and a filtered second data block corresponding thereto, wherein the filtered first data block is formed from a multiplicity of first symbols and the filtered second data block is formed from a multiplicity of second symbols, Forming time-derived first symbols, or forming first sums of time-derived first symbols, Formation conjugates complex second symbols, or formation of second sums of complex second conjugated symbols, Multiplying each time-derived first symbol, or each first sum, each with a weighting factor, whereby weighted time-derived first symbols, or weighted first sums are formed, Multiplying each conjugate complex second symbol, or every other sum, each with a weighting factor, whereby weighted conjugate complex second symbols, or weighted second sums are formed, Multiplying each weighted time-derived first symbol, or each weighted first sum, by each weighted conjugate complex second symbol, or by each weighted second sum, thereby forming products, Summing each of the real part of the products, a resulting third sum being a measure of the symbol clock phase error e, and continuous correction of the symbol clock phase using the symbol clock phase error e and concomitant correction of the symbol clock frequency. Method for correcting a symbol clock frequency and phase of a data signal modulated on a carrier wave by means of a continuous phase shift keying in the case of wireless data transmission between a transmitter and a receiver by means of an incoherent block detector comprising the following steps: Receiving the data signal formed by a plurality of successive first data blocks by the receiver, Filtering the data signal by means of a first filter, Generating a further filtered data signal which is identical to the filtered data signal and which is formed from a multiplicity of filtered second data blocks identical to the filtered first data blocks, Selecting a filtered first data block and a filtered second data block corresponding thereto, wherein the filtered first data block is formed from a multiplicity of first symbols and the filtered second data block is formed from a multiplicity of second symbols, Forming time-derived first symbols, or forming first sums of time-derived first symbols, Formation conjugates complex second symbols, or formation of second sums of complex second conjugated symbols, Multiplying each time-derived first symbol, or each first sum, each with a weighting factor, whereby weighted time-derived first symbols, or weighted first sums are formed, Multiplying each conjugate complex second symbol, or every other sum, each with a weighting factor, whereby weighted conjugate complex second symbols, or weighted second sums are formed, Multiplying each weighted time-derived first symbol, or each weighted first sum, by each weighted conjugate complex second symbol, or by each weighted second sum, thereby forming products, Summing each of the real part of the products, a resulting third sum being a measure of the symbol clock phase error e, and continuous correction of the symbol clock phase using the symbol clock phase error e and concomitant correction of the symbol clock frequency. Method according to claim 1 or 2, wherein the first filter is part of a first filter bank ( 10 . 30 ), and / or wherein the second filter is part of a second filter bank ( 40 ). Method according to one of the preceding claims, wherein the respective impulse responses of the first filters are defined by first pulses of an AMP decomposition of the continuous-phase shift keying and / or the respective impulse responses of the second filters are determined by second pulses of an AMP decomposition of the continuous filters. phase shift keying. Method according to one of the preceding claims, wherein several or all weighting factors are equal. A method according to claim 4 or 5, wherein

where the index k passes through the first symbols in their chronological order and assigns a discrete time to each first symbol, the index κ traverses the second symbols in their chronological order and assigns a discrete time to each second symbol, the index l the first filters of the first symbol first filter bank ( 10 . 30 ), the index λ passes through the second filters of the second filter bank ( 40 ), L denotes the partial response length, L ₀ denotes the block length in symbols, R denotes both the number of first filters in the first filter bank ( 10 . 30 ) as well as the number of second filters in the second filter bank ( 40 ) designated, b * / l, k and b _{λ, κ} denote the respective weighting factors, ẋ _{l, k} the time-derived first symbols of the l-th first filter of the first filter bank ( 10 . 30 ) at time k, and x _{λ, κ} denotes the second symbols of the λ-th second filter of the second filter bank ( 40 ) at time κ. The method of claim 4 or 5, wherein each first pulse is decomposed into a plurality of first pulse portions and each second pulse into a plurality of second pulse portions, wherein

where the index k passes through the first symbols in their chronological order and assigns a discrete time to each first symbol, the index κ traverses the second symbols in their chronological order and assigns a discrete time to each second symbol, the index l the first filters of the first symbol first filter bank ( 10 . 30 ), the index λ passes through the second filters of the second filter bank ( 40 ), the index m passes through the first pulse sections, the index μ passes through the second pulse sections, L ₀ denotes the block length in symbols, R both the number of first filters in the first filter bank ( 10 . 30 ) as well as the number of second filters in the second filter bank ( 40 P denotes both the number of first pulse sections and the number of second pulse sections, b * / l, k + m and b _{λ, κ + μ} denote the respective weighting factors, ν. _{l, m, k are} the time-derived first symbols of the l-th first filter of the first filter bank ( 10 . 30 ) for the m-th first pulse section at time k, and ν * / λ, μ, κ the conjugate complex second symbols of the λ-th second filter of the second filter bank ( 40 ) for the μ-th second pulse section at time κ. The method of claim 6 or 7, wherein the weighting factors are obtained from the AMP decomposition. Method according to one of the preceding claims, wherein each real part of products is multiplied before the summation each with a forgetting factor which is exponentially dependent on the time interval of the mutually multiplied symbols. Method according to one of the preceding claims, wherein the data signal is transmitted by means of the transmitter using a spectral spreading method

Images