DE102013222178B4 - Control Method for Signal Timing and Frequency Deviation Compensation of the Orthogonal Frequency Division Multiplexing System - Google Patents

Abstract

A control method for signal timing and frequency deviation compensation of the orthogonal frequency division multiplexing system, characterized in that the subscriber signal occupies M subcarriers, and comprising the steps of: (101) that the signal receiving side of the orthogonal frequency division multiplexing system receives the time domain signal y (n) and then a separated time domain pilot frequency signal y (n) at the receiving end depending on the received time domain signal y (n), for which the equation n = 1, ..., N holds, where N is the frequency domain point number for the Fourier fast transform, the frequency range of the Fourier fast transform 1~N, (102) that the time-domain pilot-frequency signal y (n) at the receiving side is converted to a complementary Fourier fast transform frequency range Y (k) = FFT (y (n)) by the Fourier fast transformation, for which the equations n = 1 , ..., N and k = 1, ..., N, (103) that the frequency domain signal l Y (k) at the receiving side in conjunction with the pilot frequency signal P (k) to be transmitted is given the following formula to obtain corresponding estimate H (k) after elimination of the pilot frequency effect: (104) that a rough timing related estimate the estimated estimate H (k) is made according to the setting of the subcarrier pitch .DELTA.m to obtain a corresponding timing deviation estimation value .DELTA.t (105) that timing compensation is performed on the pilot frequency depending on the magnitude of the timing deviation estimation value .DELTA.t by the following formula; (106) that the sub-carrier pitch Δm is lengthened and the above steps (104) and (105) are repeated, the repetition number being determined according to the bandwidth of the orthogonal frequency-division multiplexing system and occupancy state of the subcarriers of the subscriber signal Subcarrier step size Δm after a certain Accuracy request is lengthened and the above steps (104) and (105) are repeated to make a fine timing related estimate on the detected estimate H (k); and (108) timing recovery is performed on the time domain pilot frequency signal y (n) at the receiving side becomes.

Description

Field of the invention

The invention relates to the radio traffic engineering, in particular to the orthogonal frequency division multiplex communication technique, and relates to a control method for signal timing and frequency deviation compensation of the orthogonal frequency division multiplexing system.

Background of the invention

Orthogonal Frequency Division Multiplexing (OFDM) has many features such as fast data transmission, high spectral efficiency and anti-multipath. For the past few years, the OFDM technique has been successfully used in the field of mobile and / or fixed data transmission, e.g. Asymmetric Digital Subscriber Line (ADSL), Digital Video Broadcasting (DVB-T and DVB-H), and Wireless LAN (WLAN), etc. At present, OFDM is considered the most competitive transmission technology for fourth generation mobile communications.

Because the OFDM can divide a channel in the frequency domain into a plurality of orthogonal sub-channels, with the carrier waves of the different sub-channels remaining orthogonal with each other and the spectra being overlapped with each other, it is possible to reduce the interference between the sub-channels and the Increase the degree of utilization of the spectra. The OFDM system is therefore very sensitive to the frequency offset. Even a very small frequency offset can jeopardize the orthogonality between the subcarriers, leading to interference (ICI) between the carrier waves, and thus to a poor reduction in system performance.

Another difficulty in implementing the OFDM system is that the system is subject to sophisticated synchronization because the OFDM system is more sensitive to timing variations compared to single carrier waves. Symbolic timing deviation means the lead time or delay of the FFT window when demodulating the OFDM symbol. If the symbolic timing deviation is too large and the cumulation of the timing offset with the extension length of the maximum time delay is thus less than the length of the cycle prefix, the interference (ISI) is caused between the characters, resulting in destroying the OFDM symbol and lowering the system behavior , It is obvious that correct symbolic timing deviation estimation is critical to the realization of the OFDM system.

It is very important for the planning of the OFDM receiver how to increase the timing tracking as well as the efficiency and accuracy of the carrier wave frequency deviation estimation.

Out EP 0 683 576 B1 (please refer 4 in connection with Sections [0001] and [0052] to [0090] of the description), a receiver of an OFDM transmission system and the associated method of operation is known, wherein frequency and phase errors based on the FFT-transformed input signal are estimated with pilot signals and stored on the Receiving side corrected, so be compensated.

Out US 2005/0036564 A1 (please refer 1 in connection with Sections [0001], [0006] and [0084] to [0090] of the description), a method is known for processing an OFDM signal for improving the synchronization performance of an OFDM receiver, in which one based on the FFT transformed metric is calculated iteratively to determine an integer frequency and / or time offset and compensate.

Out JP 2009-171266 A (please refer 3 In connection with the sections [0001], [0003] and [0022] to [0038], a receiving unit of a radio communication apparatus and its operation is known in which an automatic frequency control unit controls an oscillator on the receiving side based on an FFT-transformed input signal ,

Also known in the art is a combination synchronizing method related to OFDM timing and carrier wave frequency deviation. For most algorithms, the timing and carrier wave frequency deviation estimates result from the calculations based on the received signals. In June 2011 has Tao Wang, Hong Guo, Ou Li, and Luokun Liu disclosed a "FFT-based Multiple Frequency Multiple Deviation" OFDM symbol timing synchronous algorithm (see "Journal of Information Industry University" Vol. 12, No. 3 , June 2011 ). In particular, it relates to a combination estimation method suitable for the OFDM system, which is based on the time / frequency-technical two-dimensional symbolic timing synchronization of the parallel FFT arrangement and the integer multiple carrier wave frequency deviation. This algorithm uses the parallel FFT arrangement such that both a complementary calculation of the received pilot frequency codes of the OFDM signal with the local pilot frequency codes as well as an integer multiple frequency deviation compensation occur. Based on the assignment of the pilot frequency codes, a symbol timing of the OFDM signal can be obtained, which allocation is sensitive to the integer multiple carrier wave frequency offset. The parallel FFT unit can therefore compensate for the frequency deviation of the received signal, eliminate the effect of the integer multiple frequency deviation on the mapping behavior, and at the same time provide an integer multiple frequency deviation estimate.

It should be noted that the aforementioned method is not able to provide very accurate timing because, on the one hand, a large amount of FFT operation is required, which is why the products using the method are very difficult or expensive to implement On the other hand, an increase in the FFT score or massive increase in the amount of computation to be processed is required, since a highly accurate timing requires a multiple or dozens of times signal sampling.

Content of the invention

Based on the disadvantages of the conventional embodiment, there is provided a control method for signal timing and frequency deviation compensation of the orthogonal frequency division multiplexing system which can accurately calculate the timing during the resulting integer multiple frequency deviation estimation from the frequency domain processing, so that high accuracy of the timing and frequency deviation estimation can be achieved only by means of one small calculation size is available and thus low application costs and easy implementations are possible.

• - dass das Teilnehmer-Signal M- Unterträger belegt, (101) dass die Signalempfangsseite des orthogonalen Frequenzmultiplexsystems das Zeitbereichssignal y(n) empfängt und anschließend ein separiertes Zeitbereichspilotfrequenzsignal y_p(n) an der Empfangsseite je nach empfangenem Zeitbereichssignal y(n) erzielt, wofür die Gleichung n = 1,...,N gilt, wobei N als die Frequenzbereichspunktzahl für die Fourier-Schnelltransformation gilt, wobei der Frequenzbereich der Fourier-Schnelltransformation 1~N beträgt,
• - (102) dass das Zeitbereichspilotfrequenzsignal y_p(n) an der Empfangsseite in einen komplementären Fourier-Schnelltransformationsfrequenzbereich Y_p(k) = FFT(y_p(n)) durch die Fourier-Schnelltransformation konvertiert wird, wofür die Gleichungen n = 1,...,N und k = 1,...,N gelten,
• - (103) dass das Frequenzbereichssignal Y_p(k) an der Empfangsseite in Verbindung mit dem zu sendenden Pilotfrequenzsignal P(k) anhand der folgenden Formel gebracht wird, um entsprechenden Schätzwert H(k) nach dem Beseitigen der Pilotfrequenzauswirkung zu ermitteln: $$H\left(k\right)=Y_p\left(k\right)\times conj\left(P\left(k\right)\right)\text{,wobei }k=\mathrm{1,...,}N\text{,}$$
  
• - (104) dass eine grobe zeitgebungsbezogene Abschätzung an dem ermittelten Schätzwert H(k) je nach Einstellung der Unterträger-Schrittweite Δm vorgenommen wird, um einen entsprechenden Zeitgebungsabweichungsschätzwert Δt zu ermitteln,
• - (105) dass eine Zeitgebungskompensation an der Pilotfrequenz in Abhängigkeit von der Größe des Zeitgebungsabweichungsschätzwerts Δt anhand der folgenden Formel vorgenommen wird, $$H\text{\_}comp\left(k\right)=H\left(k\right)\times e^{-j\times 2\pi \times \Delta t\times k\text{/}N}\text{,wobei }k\text{=1}\mathrm{,...,}N\text{,}$$
  
• - (106) dass die Unterträger-Schrittweite Δm verlängert wird und die obigen Schritte (104) und (105) wiederholt werden, wobei die Wiederholungszahl je nach Bandbreite des orthogonalen Frequenzmultiplexsystems sowie Belegungszustand der Unterträger des Teilnehmer-Signals bestimmt wird,
• - (107) dass die Unterträger-Schrittweite Δm nach einer bestimmten Genauigkeitsanforderung verlängert wird und die obigen Schritte (104) und (105) wiederholt werden, um eine feine zeitgebungsbezogene Abschätzung an dem ermittelten Schätzwert H(k) vorzunehmen, und
• - (108) dass eine Zeitgebungswiederherstellung an dem Zeitbereichspilotfrequenzsignal y_p(n) an der Empfangsseite vorgenommen wird.

In order to achieve the object, the invention proposes a control method for signal timing and frequency deviation compensation of the orthogonal frequency division multiplex system, in which it is provided that

• (101) that the signal receiving side of the orthogonal frequency division multiplexing system receives the time domain signal y (n) and then obtains a separated time domain pilot frequency signal y _p (n) at the receiving side according to the received time domain signal y (n), for which the equation n = 1, ..., N holds, where N is the frequency domain point number for the Fourier fast transformation, the frequency range of the Fourier fast transformation being 1 ~ N,
• (102) that the time domain pilot frequency signal y _p (n) at the receiving side is converted into a complementary Fourier fast transformation frequency range Y _p (k) = FFT (y _p (n)) by the Fourier fast transformation, for which the equations n = 1, ..., N and k = 1, ..., N apply,
• - (103) that the frequency domain signal Y _p (k) at the receiving side in conjunction with the pilot frequency signal P (k) to be transmitted is given the following formula to obtain corresponding estimated value H (k) after elimination of the pilot frequency effect: $$H\left(k\right)=Y_p\left(k\right)\times cOnj\left(P\left(k\right)\right)\text{,in which}k=\mathrm{1,...,}N\text{.}$$
  
• (104) a coarse timing-related estimate is made on the determined estimate H (k) depending on the setting of subcarrier step size Δm to determine a corresponding timing deviation estimate Δt,
• - (105) that timing compensation is performed on the pilot frequency depending on the magnitude of the timing deviation estimation value Δt by the following formula. $$H\text{\_}cOmp\left(k\right)=H\left(k\right)\times e^{-j\times 2\pi \times \Delta t\times k\text{/}N}\text{,in which}k\text{= 1}\mathrm{,\; ...,}N\text{.}$$
  
• - (106) incrementing the sub-carrier pitch Δm and repeating the above steps (104) and (105), the repetition number being determined according to the bandwidth of the orthogonal frequency-division multiplexing system and occupancy state of the subcarriers of the subscriber signal,
• - (107) incrementing the sub-carrier pitch Δm after a certain accuracy requirement and repeating the above steps (104) and (105) to make a fine timing-related estimate on the determined estimate H (k), and
• - (108) that timing recovery is performed on the time domain pilot frequency signal y _p (n) at the receiving side.

• (104-1) Ermitteln von tg(k) laut der folgenden Formel durch die Division der Schätzwerte H der beiden Frequenzpunkte, die voneinander einen Abstand von Δm aufweisen und innerhalb des Fourier-Schnelltransformationsfrequenzbereichs durch das Teilnehmer-Signal belegt sind: $$tg\left(k\right)={\frac{H\left(k+\Delta m\right)}{H\left(k\right)}}_{,}$$
  
  dabei gelten k und k + Δm jeweils als der innerhalb des Fourier-Schnelltransformationsfrequenzbereichs durch das Teilnehmer-Signal belegte Frequenzpunkt,
• (104-2) Ermitteln des Mittelwertes aus allen tg(k) zum Ermitteln von tg_ave anhand der folgenden Formel: $$tg\text{\_}ave=\overline{tg\left(k\right)}\text{,}$$
  
• (104-3) Ermitteln des Zeitgebungsabweichungsschätzwertes Δt anhand der folgenden Formel: $$\Delta t=\text{atan}\left(tg\text{\_}ave\right)\times N\text{/}\left(2\pi \times \Delta m\right)\text{,}$$
  
  dabei gilt atan(tg_ave) für das Berechnen der Phasenlage.

It is expedient if step (104) comprises the following sub-steps:

• (104-1) Determining tg (k) according to the following formula by dividing the estimated values H of the two frequency points spaced from each other by Δm and occupied by the subscriber signal within the Fourier fast transformation frequency range: $$tG\left(k\right)={\frac{H\left(k+\Delta m\right)}{H\left(k\right)}}_{.}$$
  
  where k and k + Δm are each considered to be the frequency point occupied by the subscriber signal within the Fourier fast transform frequency range,
• (104-2) Find the mean of all tg (k) to find tg_ave using the following formula: $$tG\text{\_}ave=\overline{tG\left(k\right)}\text{.}$$
  
• (104-3) determining the timing deviation estimation value Δt by the following formula: $$\Delta t=\text{atan}\left(tG\text{\_}ave\right)\times N\text{/}\left(2\pi \times \Delta m\right)\text{.}$$
  
  atan (tg_ave) applies to the calculation of the phase position.

• - erneutes Einstellen der Unterträger-Schrittweite Δm je nach Belegungszustand der Teilnehmer in dem positiven und negativen Frequenzbereich: $$\Delta m=n_{neg}-n_{pos}\text{,}$$
  
  dabei gilt, dass n_pos eine Frequenzpunktnummer ist, welche in dem Nullfrequenzbereich und dem positiven Frequenzbereich innerhalb des Fourier-Schnelltransformationsfrequenzbereichs durch das Teilnehmer-Signal belegt ist, während n_neg eine Frequenzpunktnummer ist, welche in dem negativen Frequenzbereich innerhalb des Fourier-Schnelltransformationsfrequenzbereichs durch das Teilnehmer-Signal belegt ist.

In a further advantageous embodiment of the control method, it can be provided that the renewed setting of the subcarrier pitch Δm in the step (107) takes place as follows:

• - Re-setting the subcarrier step size Δm depending on the occupancy state of the participants in the positive and negative frequency range: $$\Delta m=n_{neG}-n_{pOs}\text{.}$$
  
  where n _{pos is} a frequency point number occupied by the subscriber signal in the zero frequency range and the positive frequency range within the Fourier fast transform frequency range, while n _{neg is} a frequency point number which is in the negative frequency range within the Fourier fast transform frequency range Subscriber signal is busy.

In this control method, n _{pos is} the first frequency point number occupied in the zero frequency range and the positive frequency range within the Fourier fast transform frequency range by the subscriber signal, while n _{neg is} the first frequency point number in the negative frequency range within the Fourier fast transform frequency range the subscriber signal is busy.

The positive frequency range is considered to be the positive frequency range in translating the subscriber signal to baseband for processing, and refers to all frequency points above the center of the system bandwidth with respect to the system bandwidth occupied by the subscribers. The negative frequency range refers to all frequency points below the center of the system bandwidth. The system bandwidth center frequency point is generally considered to be the frequency zero during processing of the baseband signal. In the Fourier fast transform frequency range, the numbers 1~N / 2 + 1 as the zero frequency range and the positive frequency range points, respectively, and the numbers N / 2 + 2~N as the negative frequency range points.

• (108-1) Umwandeln des Zeitbereichspilotfrequenzsignals y_p(n) an der Empfangsseite in das Frequenzbereichssignal Y(k) durch die Fourier-Schnelltransformation anhand der folgenden Formel: $$Y\left(k\right)=FFT\left(y\left(n\right)\right)\text{,wobei }n=\mathrm{1,...,}N\text{, }k=\mathrm{1,...,}N\text{,}$$
  
• (108-2) zeitgebungsbozogenes Kompensieren an dem Frequenzbereichssignal Y(k) anhand der folgenden Formel, $$Y\_comp\left(k\right)=Y\left(k\right)\times e^{-j\times 2\pi \times \Delta t\times k\text{/}N}\text{,wobei }k=\mathrm{1,...,}N\text{,}$$
  
• (108-3) Umwandeln des kompensierten Frequenzbereichssignals Y_comp(k) in das Zeitbereichssignal, um eine weitere Signalverarbeitung zu ermöglichen.

In the control method according to the invention, the step (108) can expediently comprise the following sub-steps:

• (108-1) converting the time-domain pilot-frequency signal y _p (n) at the receiving side into the frequency-domain signal Y (k) by the Fourier fast-transformation using the following formula: $$Y\left(k\right)=FFT\left(y\left(n\right)\right)\text{,in which}n=\mathrm{1,...,}N\text{.}k=\mathrm{1,...,}N\text{.}$$
  
• (108-2) timing-compensated on the frequency-domain signal Y (k) by the following formula, $$Y\_cOmp\left(k\right)=Y\left(k\right)\times e^{-j\times 2\pi \times \Delta t\times k\text{/}N}\text{,in which}k=\mathrm{1,...,}N\text{.}$$
  
• (108-3) converting the compensated frequency domain signal Y_comp (k) into the time domain signal to enable further signal processing.

• (101-1) Empfangen des Zeitbereichssignals y(n) durch die Signalempfangsseite des orthogonalen Frequenzmultiplexsystems,
• (101-2) Umsetzen des Mittenfrequenzpunktes von dem Zeitbereichssignal y(n) in den Nullfrequenzbereich, um das Signal y-shift(n) nach dem Spektralversetzen zu erfassen, falls das Zeitbereichssignal y(n) nur in dem positiven Frequenzbereich oder nur in dem negativen Frequenzbereich liegt, oder in dem positiven Frequenzbereich und dem negativen Frequenzbereich asymmetrisch verlegt ist,
• (101-3) Ermitteln des Zeitbereichspilotfrequenzsignals y_p(n) an der Empfangsseite gemäß dem spektralversetzten Signal y_shift(n).

Furthermore, it can be advantageously provided for the control method that the step (101) comprises the following sub-steps:

• (101-1) receiving the time domain signal y (n) through the signal receiving side of the orthogonal frequency division multiplexing system,
• (101-2) converting the center frequency point from the time domain signal y (n) to the zero frequency range to detect the signal y-shift (n) after the spectral shift if the time domain signal y (n) is only in the positive frequency range or only in the positive frequency range negative frequency range, or is asymmetrically shifted in the positive frequency range and the negative frequency range,
• (101-3) Determining the time domain pilot frequency signal y _p (n) at the receiving side in accordance with the spectrally offset signal y_shift (n).

dabei gelten f_offset = round ((m₁+m_M)/2)-M_all/2, round ((m₁+m_M)/2) als die Mittenfrequenzpunkte der M- Unterträger m₁∼m_M, m₁ als die kleinste Numerierung des in der Systembandbreite durch einen Teilnehmer belegten Frequenzpunktes, m_M als größte Numerierung des in der Systembandbreite durch einen Teilnehmer belegten Frequenzpunktes, M_all als die Anzahl der verfügbaren Unterträger in der Systembandbreite, Δf als die Intervalle der Unterträger in dem orthogonalen Frequenzmultiplexsystem.Furthermore, it may be expedient if the conversion of the center frequency point from the time domain signal y (n) into the zero frequency range for detecting the spectrally offset signal y_shift (n) takes place as follows: Determining the spectrally offset signal y_shift (n) using the following formula: $$y\text{\_}sHift\left(n\right)=y\left(n\right)\times e^{-j\times 2\pi \times f\text{\_}Offset\times \Delta f\text{/}N}\text{.}n\text{= 1}\text{.}\cdot \cdot \cdot \text{.}N\text{.}$$

where f_offset = round ((m ₁ + m _M ) / 2) -M _all / 2, round ((m ₁ + m _M ) / 2) as the center frequency points of the M subcarriers m ₁ ~m _M , m ₁ as the smallest numbering of the frequency point occupied by a subscriber in the system bandwidth, m _M as the largest numbering of the frequency point occupied by a subscriber in the system bandwidth, M _all as the number of available subcarriers in the system bandwidth, Δf as the intervals of the subcarriers in the orthogonal frequency division multiplex system ,

• (203) Feststellen eines ganzzahlig mehrfachen Frequenzversatzes ΔF je nach Spektralschwerpunkt,
• (204) Erneutes Berechnen der durch das Teilnehmer-Signal belegten M-Unterträger m₁+ΔF'∼m_M +ΔF anhand des festgestellten ganzzahlig mehrfachen Frequenzversatzes ΔF, wobei die folgenden Sub-Schritte nach dem Schritt (108) vorgegeben sind:
• (207) Ganzzahlig mehrfaches frequenzabweichungsbezogenes Kompensieren an dem Zeitbereichspilotfrequenzsignal y_p(n) an der Empfangsseite.

Advantageously, it can be provided that the control method comprises the following sub-steps between the step (102) and the step (103) if the received signal has an integer multiple frequency deviation:

• (203) determining an integer multiple frequency offset ΔF according to the spectral centroid,
• (204) Re-calculating the M-subcarrier m ₁ + ΔF'm _M + ΔF occupied by the subscriber signal on the basis of the determined integer multiple frequency offset ΔF, the following sub-steps being predefined after step (108):
• (207) Integer multiple frequency deviation related compensation to the time domain pilot frequency signal y _p (n) at the receiving side.

• (203-1) Berechnen der Energie jedes Frequenzpunktes des Frequenzbereichssignals und Generieren eines Energiespektrums,
• (203-2) Rauschunterdrücken an dem generierten Energiespektrum,
• (203-3) Berechnen des sich aus der Rauschunterdrückung ergebenden Spektralschwerpunktes,
• (203-4) Feststellen eines ganzzahlig mehrfachen Frequenzversatzes ΔF gemäß dem sich ergebenden Spektralschwerpunkt.

In this control method, the determination of an integer multiple frequency offset ΔF can advantageously have the following steps, depending on the spectral centroid:

• (203-1) calculating the energy of each frequency point of the frequency domain signal and generating an energy spectrum,
• (203-2) noise suppression on the generated energy spectrum,
• (203-3) calculating the spectral centroid resulting from the noise suppression;
• (203-4) detecting an integer multiple frequency offset ΔF according to the resulting spectral centroid.

The present control method is for detecting various timing estimation ranges and accuracy deviation estimates using the subcarriers of different step widths for timing-related estimation depending on the phase deviation in the positive and negative signal frequency ranges, respectively, to effectively improve the timing and frequency deviation related estimation accuracy only by a small amount of computation and accurate timing estimation while obtaining an integer multiple frequency deviation estimate by processing the frequency range if a signal of an integer multiple frequency deviation exists. The so-designed control method for signal timing and frequency deviation compensation of the orthogonal frequency division multiplexing system is characterized by low application cost and ease of implementation.

list of figures

• 1 ein Schrittdiagramm des erfindungsgemäßen Steuerverfahrens,
• 2 die Ansicht des Fourier-Schnelltransformationsfrequenzbereichs des erfindungsgemäßen Steuerverfahrens,
• 3 eine schematische Darstellung der auf den Fourier-Schnelltransformationsfrequenzbereich gemappten orthogonalen Frequenzmultiplexsystembandbreite bei der erfindungsgemäßen Ausführungsform und
• 4 ein Schrittdiagramm einer bevorzugten Ausführungsform des erfindungsgemäßen Steuerverfahrens

zeigt.Other objects, advantages, features and applications of the present invention will become apparent from the following description of the embodiments with reference to the drawing, which

• 1 a step diagram of the control method according to the invention,
• 2 the view of the Fourier fast transformation frequency range of the control method according to the invention,
• 3 a schematic representation of the mapped to the Fourier fast transformation frequency range orthogonal frequency division multiplex system bandwidth in the embodiment according to the invention and
• 4 a step diagram of a preferred embodiment of the control method according to the invention

shows.

Description of the preferred embodiment

In 1 is a step diagram of the control method according to the invention shown.

• (101) dass die Signalempfangsseite des orthogonalen Frequenzmultiplexsystems das Zeitbereichssignal y(n) empfängt und anschließend ein separiertes Zeitbereichspilotfrequenzsignal y_p(n) an der Empfangsseite je nach empfangenem Zeitbereichssignal y(n) erzielt, wofür die Gleichung n=1,...,N gilt,
• (102) dass das Zeitbereichspilotfrequenzsignal y_p(n) an der Empfangsseite in einen komplementären Fourier-Schnelltransformationsfrequenzbereich Y_p(k)=FFT(y_p(n)) durch die Fourier-Schnelltransformation konvertiert wird, wofür die Gleichungen n=1,...,N und k=1,...,N gelten,
• (103) dass das Frequenzbereichssignal Y_p(k) an der Empfangsseite in Verbindung mit dem zu sendenden Pilotfrequenzsignal P(k) anhand der folgenden Formel gebracht wird, um entsprechenden Schätzwert H(k) nach dem Beseitigen der Pilotfrequenzauswirkung zu ermitteln: $$H\left(k\right)=Y_p\left(k\right)\times conj\left(P\left(k\right)\right)\text{,wobei }k\text{=1}\mathrm{,...,}N\text{, }$$
  
• (104) dass eine grobe zeitgebungsbezogene Abschätzung an dem ermittelten Schätzwert H(k) je nach Einstellung der Unterträger-Schrittweite Δm vorgenommen wird, um einen entsprechenden Zeitgebungsabweichungsschätzwert Δt zu ermitteln,
• (105) dass eine Zeitgebungskompensation an der Pilotfrequenz in Abhängigkeit von der Größe des Zeitgebungsabweichungsschätzwerts Δt anhand der folgenden Formel vorgenommen wird, $$H\text{\_}comp\left(k\right)=H\left(k\right)\times e^{-j\times 2\pi \times \Delta t\times k\text{/}N}\text{,wobei }k\text{=1}\mathrm{,...,}N\text{,}$$
  
• (106) dass die Unterträger-Schrittweite Δm verlängert wird und die obigen Schritte (104) und (105) wiederholt werden, wobei die Wiederholungszahl je nach Bandbreite des orthogonalen Frequenzmultiplexsystems sowie Belegungszustand der Unterträger des Teilnehmer-Signals bestimmt wird,
• (107) dass die Unterträger-Schrittweite Δm nach einer bestimmten Genauigkeitsanforderung verlängert wird und die obigen Schritte (104) und (105) wiederholt werden, um eine feine zeitgebungsbezogene Abschätzung an dem ermittelten Schätzwert H(k) vorzunehmen, und
• (108) dass eine Zeitgebungswiederherstellung an dem Zeitbereichspilotfrequenzsignal y_p(n) an der Empfangsseite vorgenommen wird.

Assuming that both the positive frequency range and the negative frequency range are occupied by the subscriber signals, the subscriber signals occupy M subcarriers m ₁ ~m _M in the positive / negative frequency range as well as in the zero frequency range, the minimum subcarrier m ₁ is in the negative frequency range while the maximum sub-carrier m _{M is} in the positive frequency range. Complementary with the subcarriers in the negative frequency range, the Fourier fast transform frequency range n ₁ ~N remains, while the subcarriers in the zero frequency range and the positive frequency range are complementary to the Fourier fast transform frequency range 1n _M , so the control method comprises the following steps:

• (101) that the signal receiving side of the orthogonal frequency division multiplexing system receives the time domain signal y (n) and then obtains a separated time domain pilot frequency signal y _p (n) at the receiving side according to the received time domain signal y (n), for which the equation n = 1, ..., N is valid,
• (102) that the time domain pilot frequency signal y _p (n) at the receiving side is converted to a complementary Fourier fast transformation frequency range Y _p (k) = FFT (y _p (n)) by the Fourier fast transformation, for which the equations n = 1,. .., N and k = 1, ..., N apply,
• (103) that the frequency domain signal Y _p (k) at the receiving side in conjunction with the pilot frequency signal P (k) to be transmitted is expressed by the following formula to obtain corresponding estimated value H (k) after eliminating the pilot frequency impact: $$H\left(k\right)=Y_p\left(k\right)\times cOnj\left(P\left(k\right)\right)\text{,in which}k\text{= 1}\mathrm{,\; ...,}N\text{.}$$
  
• (104) that a rough timing related estimate is made to the determined estimate H (k) depending on the setting of the subcarrier pitch Δm to obtain a corresponding timing deviation estimation value Δt,
• (105) that timing compensation is performed on the pilot frequency depending on the magnitude of the timing deviation estimation value Δt by the following formula. $$H\text{\_}cOmp\left(k\right)=H\left(k\right)\times e^{-j\times 2\pi \times \Delta t\times k\text{/}N}\text{,in which}k\text{= 1}\mathrm{,\; ...,}N\text{.}$$
  
• (106) that the subcarrier pitch Δm is lengthened and the above steps ( 104 ) and ( 105 ), the repetition number being determined according to the bandwidth of the orthogonal frequency division multiplex system and the occupancy state of the subcarriers of the subscriber signal,
• (107) that the subcarrier pitch Δm is extended after a certain accuracy requirement and the above steps ( 104 ) and ( 105 ) to make a fine timing-related estimate on the determined estimate H (k), and
• (108) that timing recovery is performed on the time-domain pilot frequency signal y _p (n) at the receiving side.

• (104-1) Ermitteln von tg(k) laut der folgenden Formel durch die Division der Schätzwerte H(k) der beiden Frequenzpunkte, die all im Abstand von Δm Frequenzpunkten innerhalb des Fourier-Schnelltransformationsfrequenzbereichs durch das Teilnehmer-Signal belegt sind: $$tg\left(k\right)=\frac{H\left(k+\Delta m\right)}{H\left(k\right)}\text{,}$$
  
  dabei, k=1,...,n_M-Δm, n₁, Δm :
• (104-2) Ermitteln des Mittelwertes aus allen tg(k) zum Ermitteln von tg_ave anhand der folgenden Formel: $$tg\text{\_}ave=\overline{tg\left(k\right)}\text{,}$$
  
• (104-3) Ermitteln des Zeitgebungsabweichungsschätzwertes Δt anhand der folgenden Formel: $$\Delta t=\text{atan}\left(tg\text{\_}ave\right)\times N\text{/}\left(2\pi \times \Delta m\right)\text{,}$$
  
  dabei gilt atan(tg_ave) für das Berechnen der Phasenlage.

In a preferred embodiment, the step ( 104 ) the following sub-steps:

• (104-1) Determining tg (k) according to the following formula by dividing the estimates H (k) of the two frequency points occupied by the subscriber signal all at intervals of Δm frequency points within the Fourier fast transformation frequency range: $$tG\left(k\right)=\frac{H\left(k+\Delta m\right)}{H\left(k\right)}\text{.}$$
  
  thereby, k = 1, ..., n _M -Δm, n ₁ , Δm:
• (104-2) Find the mean of all tg (k) to find tg_ave using the following formula: $$tG\text{\_}ave=\overline{tG\left(k\right)}\text{.}$$
  
• (104-3) determining the timing deviation estimation value Δt by the following formula: $$\Delta t=\text{atan}\left(tG\text{\_}ave\right)\times N\text{/}\left(2\pi \times \Delta m\right)\text{.}$$
  
  atan (tg_ave) applies to the calculation of the phase position.

In another preferred embodiment, resetting the subcarrier pitch Δm in the step (FIG. 107 The procedure is as follows: resetting the subcarrier increment .DELTA.m depending on the occupancy state of the subscribers in the positive and negative frequency ranges: $$\Delta m=n_{neG}-n_{pOs}\text{.}$$

In this case, n _pos is a frequency point number, for example No. 1 for the first frequency point, which is occupied by the subscriber signal in the zero frequency range and the positive frequency range within the Fourier fast transformation frequency range. n _neg is considered a frequency point number, eg No. n1 for the first frequency point, which is occupied by the subscriber signal in the negative frequency range within the Fourier fast transformation frequency range, namely Δm = n ₁ -1.

The positive frequency range is considered to be the positive frequency range in translating the subscriber signal to baseband for processing, and refers to all frequency points above the center of the system bandwidth with respect to the system bandwidth occupied by the subscribers. The negative frequency range refers to all frequency points below the center of the system bandwidth. Of the System bandwidth center frequency point is generally considered to be the frequency zero during processing of the baseband signal. In the Fourier fast transform frequency range, the number 1~N / 2 + 1 counts as the zero frequency range and the positive frequency range points, respectively, N / 2 + 2~N as the negative frequency range points.

Δm may be a different setpoint, provided that the two H-values to be divided with one another lie in the positive frequency range and the negative frequency range, respectively.

• (108-1) Umwandeln des Zeitbereichspilotfrequenzsignals y_p(n) an der Empfangsseite in das Frequenzbereichssignal Y(k) durch die Fourier-Schnelltransformation anhand der folgenden Formel: $$Y\left(k\right)=FFT\left(y\left(n\right)\right)\text{,wobei }n=\mathrm{1,...,}N\text{, }k=\mathrm{1,...,}N\text{,}$$
  
• (108-2) Zeitgebungsbozogenes Kompensieren an dem Frequenzbereichssignal Y(k) anhand der folgenden Formel, $$Y\_comp\left(k\right)=Y\left(k\right)\times e^{-j\times 2\pi \times \Delta t\times k\text{/}N}\text{,wobei }k=\mathrm{1,...,}N\text{,}$$
  
• (108-3) Umwandeln des kompensierten Frequenzbereichssignals Y_comp(k) in das Zeitbereichssignal, um eine weitere Signalverarbeitung zu ermöglichen.

In another preferred embodiment, the step comprises 108 ) the following sub-steps:

• (108-1) converting the time-domain pilot-frequency signal y _p (n) at the receiving side into the frequency-domain signal Y (k) by the Fourier fast-transformation using the following formula: $$Y\left(k\right)=FFT\left(y\left(n\right)\right)\text{,in which}n=\mathrm{1,...,}N\text{.}k=\mathrm{1,...,}N\text{.}$$
  
• (108-2) Timing Boost Compensating on the frequency domain signal Y (k) by the following formula, $$Y\_cOmp\left(k\right)=Y\left(k\right)\times e^{-j\times 2\pi \times \Delta t\times k\text{/}N}\text{,in which}k=\mathrm{1,...,}N\text{.}$$
  
• (108-3) converting the compensated frequency domain signal Y_comp (k) into the time domain signal to enable further signal processing.

• (101-1) Empfangen des Zeitbereichssignals y(n) durch die Signalempfangsseite des orthogonalen Frequenzmultiplexsystems,
• (101-2) Umsetzen des Mittenfrequenzpunktes von dem Zeitbereichssignal y(n) in den Nullfrequenzbereich, um das Signal y_shift(n) nach dem Spektralversetzen zu erfassen, falls das Zeitbereichssignal y(n) nur in dem positiven Frequenzbereich oder nur in dem negativen Frequenzbereich liegt, oder in dem positiven Frequenzbereich und dem negativen Frequenzbereich asymmetrisch verlegt ist,
• (101-3) Ermitteln des Zeitbereichspilotfrequenzsignals y_p(n) an der Empfangsseite je nach Signal y_shift(n) nach dem Spektralversetzen.

In a further preferred embodiment, the step comprises 101 ) the following sub-steps:

• (101-1) receiving the time domain signal y (n) through the signal receiving side of the orthogonal frequency division multiplexing system,
• (101-2) converting the center frequency point from the time domain signal y (n) to the zero frequency range to detect the signal y_shift (n) after the spectral shift if the time domain signal y (n) is only in the positive frequency range or only in the negative frequency range or asymmetrically in the positive frequency range and the negative frequency range,
• (101-3) Determining the time domain pilot frequency signal y _p (n) at the receiving end according to the signal y_shift (n) after the spectral shift.

dabei gelten f_offset=round ((m₁+m_M)/2)-M_all/2, round ((m₁+m_M)/2) als die Mittenfrequenzpunkte der M- Unterträger m₁∼m_M, Δf als die Intervalle der Unterträger in dem orthogonalen Frequenzmultiplexsystem.That the conversion of the center frequency point from the time domain signal y (n) into the zero frequency range for detecting the signal y_shift (n) after the spectral shift is concrete as follows: Determining the signal y_shift (n) after the spectral offset using the following formula: $$y\text{\_}sHift\left(n\right)=y\left(n\right)\times e^{-j\times 2\pi \times f\text{\_}Offset\times \Delta f\text{/}N}\text{.}n=\text{1}\text{.}\cdot \cdot \cdot \text{.}N\text{.}$$

where f_offset = round ((m ₁ + m _M ) / 2) -M _all / 2, round ((m ₁ + m _M ) / 2) as the center frequency points of the M subcarriers m ₁ ~m _M , Δf as the Intervals of Subcarriers in the Orthogonal Frequency Division Multiplexing System.

• (203) Feststellen eines ganzzahlig mehrfachen Frequenzversatzes ΔF je nach Spektralschwerpunkt,
• (204) Erneutes Berechnen der durch das Teilnehmer-Signal belegten M-Unterträger m₁+ΔF∼m_M+ΔF anhand des festgestellten ganzzahlig mehrfachen Frequenzversatzes ΔF, wobei die folgenden Sub-Schritte nach dem Schritt (108) vorgesehen sind:
• (207) Ganzzahlig mehrfaches frequenzabweichungsbezogenes Kompensieren an dem Zeitbereichspilotfrequenzsignal y_p(n) an der Empfangsseite.

In a preferred embodiment, the control method comprises the following sub-steps between the steps ( 102 ) and ( 103 ), if the received signal has an integer multiple frequency deviation:

• (203) determining an integer multiple frequency offset ΔF according to the spectral centroid,
• (204) Re-calculating the M-subcarrier m ₁ + ΔF~ _{m M} + ΔF occupied by the subscriber signal based on the detected integer multiple frequency offset ΔF, the following sub-steps after the step ( 108 ) are provided:
• (207) Integer multiple frequency deviation related compensation to the time domain pilot frequency signal y _p (n) at the receiving side.

• (203-1) Berechnen der Energie jedes Frequenzpunktes des Frequenzbereichssignals und Generieren eines Energiespektrums,
• (203-2) Rauschunterdrücken an dem generierten Energiespektrum,
• (203-3) Berechnen des aus der Rauschunterdrückung ergebenen Spektralschwerpunktes,
• (203-4) Feststellen eines ganzzahlig mehrfachen Frequenzversatzes ΔF nach dem ergebenen Spektralschwerpunkt.

In another preferred embodiment it is provided that the determination of an integer multiple frequency offset ΔF according to the spectral center of gravity has the following steps:

• (203-1) calculating the energy of each frequency point of the frequency domain signal and generating an energy spectrum,
• (203-2) noise suppression on the generated energy spectrum,
• (203-3) calculating the spectral centroid resulting from the noise suppression,
• (203-4) Find an integer multiple frequency offset ΔF after the resulting spectral centroid.

In the orthogonal frequency multiplexing system, the timing-related evaluation is made according to the phase position of the frequency domain. Depending on the setting of the sub-carrier step size, various timing-related estimation ranges and accuracies can be determined. Further, highly accurate timing positions may be detected according to the estimation of the phase angle deviations in the positive and negative signal frequency ranges. In the case of the signal provided with an integer multiple frequency deviation, it is again possible to calculate, on the basis of the estimated integer multiple frequency deviation in the frequency range, on which subcarrier the signal is located. Thereupon, the timing related estimate results from the phase difference between these subcarriers.

The basic principle of the step size control method according to the invention is that the sequential time domain offsets remain complementary to the rotation of the phase positions in the frequency range. For the limited sequence x (n), 0≤n≤N-1, X (k) is considered the scattered Fourier transform related to x (n): X (k) = DFT [x (n)].

wobei $W_N^{-mk}=e^{j\times 2\pi \times m\times k\text{/}N}\text{.}$

In the formula x _m (n) = x ((n + m)) _N R _N (n), x ((n + m)) _N is the conversion of the interval continuation sequence relative to x _n , the multiplication by R _N (n ) means the selection of the main value sequence in the interval sequence after the conversion in the continuation sequence. Therefore, x _m (n) still holds as a restricted sequence with N points. For x _m (n) the scattered Fourier transformation X _m (k) holds, so that $X_m\left(k\right)=DFT\left[x_m\left(n\right)\right]=W_N^{-mk}X\left(k\right)\text{.}$

in which $W_N^{-mk}=e^{j\times 2\pi \times m\times k\text{/}N}\text{,}$

$$m{}_{\cdot }=\text{atan}\left(\Delta \right)\text{/}\left(2\pi \times \left(k_2-k_1\right)\text{/}N\right)$$

From the phase position difference between the frequency points k ₁ and k ₂ , the conversion m can be determined: $$\Delta =\frac{X_m\left(k_2\right)\text{/}X\left(k_2\right)}{X_m\left(k_1\right)\text{/}X\left(k_1\right)}=\frac{W_N^{-m{k}_2}}{W_N^{-\mathrm{<figure-callout\; id="1"\; label="m\; k"\; filenames="DE102013222178B4\_0046.png"\; state="\{\{state\}\}">m\; </figure-callout>}{k}_{}{}_1}}=\frac{e^{j\times 2\pi \times m\times {\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="DE102013222178B4\_0046.png"\; state="\{\{state\}\}">k</figure-callout>}}_2\text{/}N}}{e^{j\times 2\pi \times m\times {\mathrm{<figure-callout\; id="1"\; label="k"\; filenames="DE102013222178B4\_0046.png"\; state="\{\{state\}\}">k</figure-callout>}}_1\text{/}N}}=e^{j\times 2\pi \times m\times \left(k_2-k_1\right)}$$

$$m{}_{\cdot }=\text{atan}\left(\Delta \right)\text{/}\left(2\pi \times \left(k_2-k_1\right)\text{/}N\right)$$

Atan (•) applies to the calculation of the phase position.

Assume that the number of points in the FFT for signal processing is N (see 2 ) and the frequency range processed by FFT is 1 ~ N, where the zero frequency range is 1, 2~N / 2 + 1 is the positive frequency number, while N / 2 + 2~N is the negative frequency number ,

As in 3 is shown, the number of available sub-carriers within the system bandwidth M _all (M _all <N). The subcarrier numbers are 1~M _all , where 1-M _all / 2-1 are the subcarrier numbers of the negative frequency range, M _all / 2 is the frequency zero number, and M _all / 2 + 1~M is _all the subcarrier numbers. Numbers of the positive frequency range are.

The M _all subcarriers are assigned to the FFT frequency range (1~N):

The system bandwidth null frequency M _all / 2 is set to the FFT frequency range 1 mapped, the negative system bandwidth frequency range 1~M _all / 2-1 is mapped to the FFT frequency range _all NM / 2 + 2~N, the positive system bandwidth frequency range _all M / 2 + 1~M _all is applied to the FFT Frequency range 2 ~M _all / 2 + 1 mapped.

dabei liegt der obere Wert m_i in dem negativen Frequenzbereich, während der untere Wert m_i in dem Nullfrequenzbereich oder dem positiven Frequenzbereich liegt. Assuming that a particular subcarrier i has the subcarrier number m _i within the system bandwidth and the complementary FFT frequency domain number is n _i ,
then applies $$n_i=\{\begin{array}{l}N-\left(M_{all}\text{/ 2}-m_i+1\right)+2\\ m_i-M_{all}\text{/ 2}\mathrm{+}\text{1}\end{array}\text{.}$$

In this case, the upper value m _i lies in the negative frequency range, while the lower value m _i lies in the zero frequency range or the positive frequency range.

The first compensation procedure

Assume that the subscriber signal has occupied M subcarriers in both the positive / negative frequency range and the zero frequency range, and the numbering within the system bandwidth is m ₁ ~m _M , where m ₁ is the smallest number in the negative frequency range While m _M is the largest number in the positive frequency range, the complementary numbers of the FFT frequency range are n ₁ ~N with respect to the negative frequency range n ₁ , and with respect to the zero frequency range and the positive frequency range 1~n _M.

• Schritt 101: Empfangen des Signals y(n), aus dem die Pilotfrequenz daraufhin separiert werden soll, wobei die Formel $$y_p\left(n\right)\text{: }{y}_p\left(n\right)\text{,}n=\text{1}\text{,}\cdot \cdot \cdot \text{,}N$$
  
  gilt,
• Schritt 102: Umwandeln des Zeitbereichspilotfrequenzsignals an der Empfangsseite mithilfe von FFT in den Frequenzbereich: $$Y_p\left(k\right)=FFT\left(y_p\left(n\right)\right)\text{,wobei }n=\mathrm{1,...,}N\text{, }k=\mathrm{1,...,}N\text{,}$$
  
• Schritt 103: Zuordnen des empfangenen Frequenzbereichssignals mit dem gesendeten Pilotfrequenzsignals, und Beseitigen der Auswirkung der Pilotfrequenz: $$H\left(k\right)=Y_p\left(k\right)\times conj\left(P\left(k\right)\right)\text{,}$$
  
  wobei k =1,...,N .
• Schritt 104: Ausführen der groben zeitgebungsbezogenen Abschätzung.

y (n) is the received time domain signal, while P (k) is the transmitted pilot frequency domain signal. The control method according to the invention may comprise the following steps:

• step 101 Receiving the signal y (n) from which the pilot frequency is to be separated thereupon, where the formula $$y_p\left(n\right)\text{:}{y}_p\left(n\right)\text{.}n=\text{1}\text{.}\cdot \cdot \cdot \text{.}N$$
  
  applies,
• step 102 : Converting the time-domain pilot frequency signal on the receiving side to the frequency domain using FFT: $$Y_p\left(k\right)=FFT\left(y_p\left(n\right)\right)\text{,in which}n=\mathrm{1,...,}N\text{.}k=\mathrm{1,...,}N\text{.}$$
  
• step 103 Associating the received frequency domain signal with the transmitted pilot frequency signal, and eliminating the effect of the pilot frequency: $$H\left(k\right)=Y_p\left(k\right)\times cOnj\left(P\left(k\right)\right)\text{.}$$
  
  where k = 1, ..., N.
• step 104 : Perform the rough timing estimate.

Division of the H-values of the two frequency points spaced by Δm frequency points in the positive and negative frequency ranges (the two frequency points to be divided must be the frequency points occupied by the subscriber signals), namely: $$tG\left(k\right)=\frac{H\left(k+\Delta m\right)}{H\left(k\right)}\text{.}k=\text{1}\text{.}\cdot \cdot \cdot \text{.}{n}_M-\Delta m\text{, and}{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="DE102013222178B4\_0046.png"\; state="\{\{state\}\}">n</figure-callout>}}_1~N-\Delta m$$

Determining the mean value from tg (k): $$tG\text{\_}ave=\overline{tG\left(k\right)}\text{.}$$

Then, the timing estimate is determined by Formula 1 above: $$\Delta t=\text{atan}\left(tG\text{\_}ave\right)\times N\text{/}\left(2\pi \times \Delta m\right)\text{.}$$

step 105 : Timing Compensating at the Pilot Frequency Using the Frequency Range Compensation Compensation Rule $$H\_cOmp\left(k\right)=H\left(k\right)\times e^{j\times 2\pi \times \Delta t\times k\text{/}N}\text{.}wObei\text{.}k=\mathrm{1,}\cdot \cdot \cdot \text{.}N$$

step 106 : Repeating steps 104~105 after increasing the subcarrier pitch Δm,

To account for the estimation range and the accuracy, multiple evaluations are required (the sub-carrier pitch is always increased), because the larger the sub-carrier pitch, the higher the accuracy of the estimable timing and the smaller the estimation range.

Δm and the evaluation frequency are to be determined depending on the system bandwidth and occupancy state of the subcarriers of the subscriber signal.

step 107 Performing a fine timing estimation wherein the sub-carrier pitch used is limited in the positive frequency range or negative frequency range, and thus the estimation accuracy is also limited. To further increase the estimation accuracy, the phase positions of the positive frequency range and the negative frequency range can be considered together by setting Δm = n ₁ -1 to divide the H values of the two frequency points spaced by Δm frequency points in the positive frequency and negative frequency range are: $$tG\left(k\right)=\frac{H\left(k+\Delta m\right)}{H\left(k\right)}\text{.}k=\mathrm{1,}\cdot \cdot \cdot \text{,Max}\left(M_{all}\text{/ 2}\mathrm{-}{\mathrm{<figure-callout\; id="1"\; label="m"\; filenames="DE102013222178B4\_0046.png"\; state="\{\{state\}\}">m</figure-callout>}}_1\text{.}{m}_M-M_{all}\text{/}2\right)$$

The remaining steps are the same as steps 102~103.

Δm may be a different setpoint, provided that the two H-values to be divided lie in the positive frequency range and the negative frequency range, respectively.

step 108 : Scheduled recovery of the received signal. After each result of the timing-related estimation from the above-mentioned steps, either compensation is made to the completely received signal y (n) or the data of y (n), or the results can be offset from the timing-related estimates and then jointly compensated, where Perform the compensation method either in the time domain or in the frequency domain.

• Schritt 108.1: Umwandeln des Signals y(n) an der Empfangsseite mithilfe von FFT in den Frequenzbereich: $$Y\left(k\right)=FFT\left(y\left(n\right)\right)\text{,}wobei\text{, }n=\mathrm{1,}\cdot \cdot \cdot \text{,}N\text{,}k=\mathrm{1,}\cdot \cdot \cdot \text{,}N$$
  
• Schritt 108.2: Zeitgebungskompensation an dem Frequenzbereich: $$Y\text{\_}comp\left(k\right)=Y\left(k\right)\times e^{-j\times 2\pi \times \Delta t\times k\text{/}N}\text{,}wobei\text{,}k\text{=1}\text{,}\cdot \cdot \cdot \text{,}N$$
  
• Schritt 108.3: Umwandeln des kompensierten Frequenzbereichssignals in das Zeitbereichssignal, um eine weitere Signalverarbeitung zu ermöglichen.

Assuming that the timing deviation to be compensated is Δt, the frequency domain compensation methods are as follows:

• step 108.1 : Converting the signal y (n) on the receiving side by means of FFT in the frequency domain: $$Y\left(k\right)=FFT\left(y\left(n\right)\right)\text{.}wObei\text{.}n=\mathrm{1,}\cdot \cdot \cdot \text{.}N\text{.}k=\mathrm{1,}\cdot \cdot \cdot \text{.}N$$
  
• step 108.2 : Timing compensation at the frequency range: $$Y\text{\_}cOmp\left(k\right)=Y\left(k\right)\times e^{-j\times 2\pi \times \Delta t\times k\text{/}N}\text{.}wObei\text{.}k\text{= 1}\text{.}\cdot \cdot \cdot \text{.}N$$
  
• step 108.3 By converting the compensated frequency domain signal into the time domain signal to allow further signal processing.

The second compensation procedure

Due to the above step 105 For example, it can be seen that on the one hand Δm can be made very large for increasing the accuracy when the signals are in both the positive frequency range and the negative frequency range, on the other hand, almost all the H values can be used to calculate the phase deviation and to improve the noise suppression, respectively when the subcarrier number of the positive frequency range is substantially equal to that of the negative frequency range.

In cases where the signal is only in the positive frequency range or only in the negative frequency range, or asymmetrically in the positive frequency range and negative frequency range, the center frequency points of the signal can be converted into the zero frequency range by the spectral conversion, followed by a timing-related estimate should.

In this method, provided that, as before, the subscriber signals M subcarriers m ₁ ~m _M prove that the center frequency point is round ((m ₁ + m _M ) / 2) (if (m ₁ + m _M ) / 2 is not an integer, a rounding, rounding or rounding can be made according to the number (<5 or ≧ 5), which is indicated by round (•)) and the zero frequency range lies within the system bandwidth at M _all / 2. In order to convert the center frequency point of the signal into the zero frequency range, a frequency offset of
f_offset = round ((m ₁ + m _M ) / 2) -M _all / 2 required. The signal resulting from the spectral conversion corresponds
y_shift (n) = y (n) × e ^{-j × 2π × f_offset × Δf / N} , n = 1, ..., N, where Δf is the pitch of the subcarriers in the OFDM system (eg, Δf = 15 kHZ) in the LTE system).

Subsequently, the timing-related estimate is based on y_shift (n) or on the basis of the above-mentioned first method. This may result in the high accuracy timing deviation estimation for the asymmetric signals in the spectrum.

The third compensation procedure

• Schritt 201: gleich wie Schritt 101,
• Schritt 202: gleich wie Schritt 102,
• Schritt 203: Direkte Bewertung der ganzzahlig vielfachen Frequenzabweichung ΔF mithilfe des Spektrums. Es kann hier ein Schätzverfahren für die ganzzahlig vielfache Frequenzabweichung je nach Spektralschwerpunkt getroffen werden: Berechnen der Energie jedes Frequenzpunktes des Frequenzbereichssignals, Rauschunterdrücken an dem Energiespektrum, Ermitteln des Spektralschwerpunktes nach dem Rauschunterdrücken, Feststellen des ganzzahlig vielfachen Frequenzversatzes nach dem ermittelten Spektralschwerpunkt,
• Schritt 204: Erneutes Berechnen der durch das Signal belegten Unterträger je nach der ganzzahlig vielfachen Frequenzabweichung. Angenommen ebenfalls, dass die ursprünglichen Teilnehmer-Signale M-Unterträger belegen und deren Numerierung innerhalb der Systembandbreite m₁∼m_M ist, wandeln sich die Unterträger, wo sich die empfangenen Signale befinden, nach dem Auftreten der ganzzahlig vielfachen Frequenzabweichung auf m₁+ΔF∼m_M+ΔF um,
• Schritt 205: Ausführen der zeitgebungsbezogenen Abschätzung anhand des ersten oder des zweiten Verfahrens,
• Schritt 206: Terminierte Wiederherstellung des empfangenen Signals y(n) oder der Daten von y(n),
• Schritt 207: Ganzzahlig vielfache Frequenzabweichungskompensation an dem Zeitbereichssignal.

If the received signal is provided with an integer multiple frequency deviation, the timing-related estimate is generally performed only when the integer multiple frequency deviation estimate or compensation has already taken place. Here, a timing-related estimation method is proposed, which can be carried out without first correcting the integer multiple frequency deviation in advance, or with a reduced calculation quantity. This procedure (see 4 ) consists of the following steps:

• step 201 : same as step 101 .
• step 202 : same as step 102 .
• step 203 : Direct evaluation of the integer multiple frequency deviation ΔF by means of the spectrum. Here, an estimation method for the integer multiple frequency deviation depending on the spectral centroid can be made: calculating the energy of each frequency point of the frequency domain signal, noise suppression on the energy spectrum, determining the spectral centroid after the noise suppression, ascertaining the integer multiple frequency offset after the determined spectral centroid,
• step 204 Re-calculating the subcarrier occupied by the signal according to the integer multiple frequency deviation. Assuming also that the original subscriber signals occupy M subcarriers and their numbering is within the system bandwidth m ₁ ~m _M , the subcarriers where the received signals are located change to m ₁ + ΔF after the occurrence of the integer multiple frequency deviation ~ _{m M} + ΔF,
• step 205 : Performing the timing estimate using the first or the second method,
• step 206 : Scheduled recovery of the received signal y (n) or the data of y (n),
• step 207 : Integer multiple frequency deviation compensation on the time domain signal.

The control method of the present invention is for detecting various timing estimation and accuracy deviation estimates using the subcarriers of different step widths for timing-related estimation depending on the phase deviation in the positive and negative signal frequency ranges, respectively, to effectively improve the timing and frequency deviation related estimation accuracy only by a small amount of computation, and for accurate timing estimation while obtaining an integer multiple frequency deviation estimate by processing the frequency range, even if the signal has an integer multiple frequency deviation. The so-designed control method for signal timing and frequency deviation compensation of the orthogonal frequency division multiplexing system is characterized by low application cost and ease of implementation.

In the above sections the preferred embodiments are shown. It is to be noted, however, that these preferred embodiments do not limit the scope of the invention, but are only for the purpose of describing the invention. In this context, all new, disclosed in the description and / or drawing single and combination features are considered essential to the invention.

Claims (9)

A control method for signal timing and frequency deviation compensation of the orthogonal frequency division multiplexing system, characterized in that the subscriber signal occupies M subcarriers, and comprising the following steps: (101) that the signal reception side of the orthogonal frequency division multiplexing system receives the time domain signal y (n) and then a separated time domain pilot frequency signal y _p (n) is achieved at the receiving end according to the received time domain signal y (n), for which the equation n = 1, ..., N holds, where N is the frequency domain point number for the Fourier fast transformation, the frequency range of the Fourier Fast transformation 1~N, (102) that the time-domain pilot-frequency signal y _p (n) at the receiving side is converted to a complementary Fourier fast transformation frequency range Y _p (k) = FFT (y _p (n)) by the Fourier fast transformation, for which Equations n = 1, ..., N and k = 1, ..., N hold, (103) that the freq ence range signal Y _p (k) at the receiving side in conjunction with the pilot frequency signal P (k) to be transmitted is given by the following formula to obtain corresponding estimated value H (k) after elimination of the pilot frequency effect: $$H\left(k\right)=Y_p\left(k\right)\times cOnj\left(P\left(k\right)\right)\text{, in which}k=\mathrm{1,...,}N\text{.}$$

(104) that a rough timing related estimate is made on the determined estimate H (k) depending on the setting of the subcarrier pitch Δm to obtain a corresponding timing deviation estimation value Δt (105) that timing compensation at the pilot frequency depending on the magnitude of the Timing deviation estimate Δt is made by the following formula $$H\text{\_}cOmp\left(k\right)=H\left(k\right)\times e^{-j\times 2\pi \times \Delta t\times k\text{/}N}\text{,in which}k\text{= 1}\mathrm{,\; ...,}N\text{.}$$

(106) that the sub-carrier pitch Δm is lengthened and the above steps (104) and (105) are repeated, the repetition number being determined according to the bandwidth of the orthogonal frequency-division multiplexing system and occupancy state of the subcarriers of the subscriber signal Subcarrier step size Δm is extended after a certain accuracy requirement and the above steps (104) and (105) are repeated to make a fine timing related estimate on the determined estimate H (k), and (108) that a timing recovery on the time domain pilot frequency signal y _p (n) is made at the receiving side. Control method according to Claim 1 characterized in that step (104) consists of the following sub-steps: (104-1) determining tg (k) according to the following formula by dividing the estimates H of the two frequency points spaced from each other by Δm and occupied within the Fourier fast transform frequency range by the subscriber signal: $$tG\left(k\right)=\frac{H\left(k+\Delta m\right)}{H\left(k\right)}\text{.}$$

where k and k + Δm are each considered to be the frequency point occupied by the subscriber signal within the Fourier fast transform frequency range, (104-2) determining the average of all tg (k) to obtain tg_ave using the following formula: $$tG\text{\_}ave=\overline{tG\left(k\right)}\text{.}$$

(104-3) determining the timing deviation estimation value Δt by the following formula: $$\Delta t=\text{atan}\left(tG\text{\_}ave\right)\times N\text{/}\left(2\pi \times \Delta m\right)\text{.}$$

atan (tg_ave) applies to the calculation of the phase position. Control method according to Claim 1 , characterized in that the renewed setting of the subcarrier step size Δm in the step (107) is carried out as follows: - re-setting the subcarrier step size Δm depending on the occupancy state of the participants in the positive and negative frequency range: $$\Delta m=n_{neG}-n_{pOs}\text{.}$$

where n _{pos is} a frequency point number occupied in the zero frequency range and the positive frequency range within the Fourier fast transform frequency range by the subscriber signal, while n _{neg is} a frequency point number present in the negative frequency range within the Fourier fast transform frequency range by the subscriber signal is occupied. Control method according to Claim 3 characterized in that n _{pos is} the first frequency point number occupied by the subscriber signal in the zero frequency range and the positive frequency range within the Fourier fast transform frequency range, while n _{neg is} the first frequency point number present in the negative frequency range within the Fourier Fast transformation frequency range is occupied by the subscriber signal. Control method according to Claim 1 characterized in that the step (108) consists of the following sub-steps: (108-1) converting the time-domain pilot-frequency signal y _p (n) at the receiving side into the frequency-domain signal Y (k) by the Fourier fast-transformation using the following formula : $$Y\left(k\right)=FFT\left(y\left(n\right)\right)\text{,in which}n=\mathrm{1,...,}\text{N}\text{.}k=\text{1}\mathrm{,\; ...,}\text{N}\text{.}$$

(108-2) timing compensation on the frequency domain signal Y (k) by the following formula, $$Y\_cOmp\left(k\right)=Y\left(k\right)\times e^{-j\times 2\pi \times \Delta t\times k\text{/}N}\text{,in which}k=\mathrm{1,...,}N\text{.}$$

(108-3) converting the compensated frequency domain signal Y_comp (k) into the time domain signal to enable further signal processing. Control method according to Claim 1 characterized in that the step (101) consists of the following sub-steps: (101-1) receiving the time domain signal y (n) through the signal receiving side of the orthogonal frequency division multiplexing system, (101-2) converting the center frequency point from the time domain signal y ( n) in the null frequency range to detect the signal y_shift (n) after the spectral shift if the time domain signal y (n) is only in the positive frequency range or only in the negative frequency range, or asymmetrically in the positive frequency range and the negative frequency range is (101-3) Determining the time domain pilot frequency signal y _p (n) at the receiving side in accordance with the spectrally offset signal y_shift (n). Control method according to Claim 6 , characterized in that the conversion of the center frequency point of the time domain signal y (n) in the zero frequency range for detecting the spectrally offset signal y_shift (n) is concrete as follows: Determining the spectrally offset signal y_shift (n) using the following formula: $$y\_sHift\left(n\right)=y\left(n\right)\times e^{-j\times 2\pi \times f\text{\_}Offset\times \Delta f\text{/}N}\text{.}n=\mathrm{1,}\cdot \cdot \cdot \text{.}N\text{.}$$

where f_offset = round ((m ₁ + m _M ) / 2) -M _all / 2, round ((m ₁ + m _M ) / 2) as the center frequency points of the M subcarriers m ₁ ~m _M , m ₁ as the smallest numbering of the frequency point occupied by a subscriber in the system bandwidth, m _M as the largest numbering of the frequency point occupied by a subscriber in the system bandwidth, M _all as the number of available subcarriers in the system bandwidth, Δf as the intervals of the subcarriers in the orthogonal frequency division multiplex system , Control procedure according to one of Claims 1 to 7 , characterized in that said control method comprises the following sub-steps between step (102) and step (103) if the received signal has an integer multiple frequency deviation: (203) detecting an integer multiple frequency offset ΔF according to the spectral centroid, ( 204) Re-calculating the M sub-carrier occupied by the subscriber signal m ₁ + ΔF~ _{m M} + ΔFanhand of the detected integer multiple frequency offset ΔF, the following sub-steps being given after step (108): (207) Integer multiple frequency deviation related compensation at the time domain pilot frequency signal y _p (n) at the receiving side. Control method according to Claim 8 characterized in that determining an integer multiple frequency offset ΔF according to the spectral centroid comprises the following steps: (203-1) calculating the energy of each frequency point of the frequency domain signal and generating an energy spectrum, (203-2) noise suppression on the generated energy spectrum, (203-1) 3) calculating the spectral centroid resulting from the noise suppression; (203-4) detecting an integer multiple frequency offset ΔF according to the resulting spectral centroid.

Images