CN104821926B - The method and apparatus of unknown errors for estimating carrier frequency - Google Patents

Abstract

Frequency Error Estimation. A method for estimating an unknown error of a carrier frequency, the method comprising the steps of sampling a signal that is a modulated version of said carrier frequency to form a plurality of data blocks, each data block comprising a plurality of samples and each sample and into the phase error due to the carrier frequency error; for each data block: generate the product of each sample with another sample in the data block that is like the sample in the data block are located in the same position in the inverted version of the data block to form vectors of combined samples each incorporating the same phase error due to the carrier frequency error; and computing the average of the vectors of combined samples values to form a vector of average samples; and estimating the carrier frequency error in dependence on pitch frequencies in the vector of average samples generated from the plurality of data blocks.

Description

Method and apparatus for estimating unknown error of carrier frequency

technical field

The present invention relates to methods and devices for estimating errors in carrier frequencies.

Background technique

Narrowband communication systems are very sensitive to carrier frequency errors. This is because an error in the carrier frequency produces an unwanted phase rotation between symbols, which degrades receiver sensitivity. This degradation may be more severe when extensions are used. Typically, the receiver will first despread the received data and then sum the lengths of the spreading codes to obtain a good signal-to-noise ratio (SNR). However, due to the frequency error, the signals will no longer sum coherently. There is also a phase shift between symbols. The combination of these two effects results in a degradation of performance.

There are many techniques that standard receivers can use to estimate carrier frequency error. Many techniques rely on pilot or synchronization symbols, which are data sequences known in advance by the receiver. The receiver can estimate the carrier frequency error by comparing the phases of the received symbols with those it expects to receive. However, having to send a synchronization sequence reduces the achievable data rate. Decision guided techniques are another option. They estimate the phase error in the received symbols by looking at the phase difference between the received symbol and the closest point in the modulation constellation. This mechanism relies on decisions that are generally very error-prone at the SNR of interest. Angle manipulation is also very noise sensitive and can produce large errors. The additional noise degradation caused by imperfect coherent combining makes this even worse.

Therefore, there is a need for an improved mechanism for estimating carrier frequency error.

Contents of the invention

According to one embodiment, there is provided a method for estimating an unknown error of a carrier frequency, the method comprising the steps of: sampling a signal that is a modulated version of said carrier frequency to form a plurality of data blocks, each data block comprising a plurality of samples and each sample incorporates a phase error due to said carrier frequency error; for each data block: generating the product of each sample with another sample in said data block, said another sample being like said samples are located in the same position in the inverted version of the data block as in the data block to form combined samples each incorporating substantially the same phase error due to the carrier frequency error; and calculating the an average of the combined samples to form an average sample; and estimating the carrier frequency error relying on a plurality of average samples generated from the plurality of data blocks.

The method may include calculating an average of the combined samples by summing the combined samples and dividing by the number of samples in the data block.

The signal may be a version of the carrier frequency that has been modulated by spreading, and the method may include sampling the signal using a sampling rate equal to the chip rate used to spread the signal.

The signal may be a version of the carrier frequency that has been modulated by spreading, and the method may include forming the data blocks to each include a number of samples equal to a spreading factor.

The method may include estimating the carrier frequency error using a maximum likelihood estimator.

The method may include estimating the carrier frequency error by: performing a Fourier transform on the plurality of averaged samples; and selecting a frequency having a highest value in the transform as the carrier frequency error.

The method may include, for a plurality of antennas: generating a plurality of averaged samples for individual signals received by the plurality of antennas; for each set of averaged samples, performing a Fourier transform; and combining the obtained signals by squaring and summing. Obtain a number of said transformations.

The estimated carrier frequency error may incorporate an ambiguity factor, and the method may comprise a fine estimation phase in which a fine estimate of the carrier frequency error is obtained, and a coarse estimation phase in which stage resolves the ambiguity factors in the refined estimates.

The fine estimation stage may comprise the method according to claim 1 and generate a fine estimate by using a first sampling frequency to generate the data block; and the coarse estimation stage may comprise the method according to claim 1 A method is used to generate a rough estimate and resolve the ambiguity factor by generating the data block using a second sampling frequency that is lower than the first sampling frequency.

The coarse estimation stage may comprise generating the coarse estimate using a different technique for estimating the carrier frequency error than the technique used by the fine estimation stage.

The method may include, for the coarse estimation stage, generating a rough estimate of the carrier frequency error by mixing the plurality of averaged samples with two or more frequencies, each of the frequencies frequency is a multiple of the ambiguity factor; determining signal power included in each of the mixed signals; and selecting the frequency that generated the mixed signal including the highest signal power as a rough estimate of the carrier frequency error.

The method may comprise summing together the estimated carrier frequency error and the rough estimate of the carrier frequency error to form a complete estimate of the carrier frequency error in which the ambiguity factors are resolved.

The signal may be a version of the carrier frequency that has been modulated using phase shift keying.

According to a second embodiment, there is provided an apparatus configured to estimate an unknown error of a carrier frequency by implementing a method as claimed in any one of claims 1 to 12.

According to a third embodiment, there is provided an apparatus for estimating an unknown error of a carrier frequency, the apparatus comprising: a sampling unit configured to sample a signal that is a modulated version of the carrier frequency to form a multiple data blocks, each data block comprising a plurality of samples and each sample incorporating a phase error due to said carrier frequency error; a product unit configured to generate, for each data block, each sample and said data The product of another sample in the block that is in the same position in the inverted version of the data block as the sample is in the data block, to form combined samples of substantially the same phase error due to errors, and calculating an average of said combined samples to form an averaged sample; and a pitch estimator configured to use the A plurality of the averaged samples are used to estimate the carrier frequency error.

Description of drawings

The invention will now be described by way of example with reference to the accompanying drawings. In the attached picture:

Figure 1 shows an example of a method for estimating carrier frequency error;

Figure 2 shows an example of a method for implementing fine estimation and coarse estimation;

Figure 3 shows an example of a method for obtaining a coarse frequency estimate; and

Fig. 4 shows an example of an apparatus for estimating carrier frequency error.

Detailed ways

To avoid the problems of current techniques for estimating carrier frequency error, estimators are proposed that avoid the need for decisions, known preambles or pilots, and reduce the problem to the easier tone detection problem.

In this method, the unknown error of the carrier frequency is estimated by sampling a signal that is a modulated version of the carrier frequency to form multiple data blocks. Each data block includes a plurality of samples, and each sample incorporates a phase error due to a carrier frequency error.

Each chunk of data is then processed by generating a product of each sample in the chunk with another sample in the same chunk. Another sample is in the same position in the reversed version of the data block as the sample was in the data block. The result is a vector of combined samples each incorporating substantially the same phase error due to the carrier frequency error. An average of the combined samples is then calculated, and the number of averaged samples obtained by performing these steps across multiple data blocks is used to estimate the carrier frequency error.

In one example, the data is modulated using phase shift keying, specifically binary phase shift keying (BPSK). The techniques described herein may be particularly advantageous in communication systems employing a type of modulation in which data is encoded by means of phase modulation; it relies on the fact that the phase of the received signal is changed due to carrier frequency errors suffering from certain performance degradation of the phase modulation technique. However, this is only an example and it should be understood that other modulation methods may be used.

The techniques described herein may also be particularly advantageous in communication systems employing extensions. The performance degradation resulting from unwanted phase rotation due to carrier frequency errors can be particularly severe in systems employing spreading because the signals are no longer summed coherently. This shortcoming can be addressed by the techniques described herein, as they allow the unwanted phase rotation to be removed after the carrier frequency error has been estimated. Thus, although estimation techniques are described with reference to spreading signals, it should be understood that this is only to illustrate the advantages of these techniques.

An example of a method for estimating carrier frequency error is shown in FIG. 1 . The method starts by separating the received signal into data blocks (step 101). Each data block includes a plurality of samples. Data blocks can be the same size or different sizes. In one example, the data blocks each include a number of samples equal to the expansion factor _Ns . Data blocks of other sizes may also be used (as evident from the coarse frequency estimator described below).

In this example the signal has been spread such that the received chips are multiplied by the corresponding spreading value _sm (step 102). The element-wise product of the vector represented by the data block with the inverted version of itself is then computed (step 103). Each element of the resulting vector incorporates substantially the same unwanted phase rotation due to the carrier frequency error. Therefore, an average of the products is calculated (step 104), the elements summing predominantly coherently when averaged. For the sequence of each data block within the measurement period of L data symbols, it is judged whether each data block has been processed (step 105 ). The steps may be performed serially, such that individual data blocks are then processed, or may be efficiently processed in parallel. L need not be the length of the data sequence, and may be chosen as a tradeoff between accuracy and error variability. The resulting vector of length L is the sum of the aliased tone and noise at frequency 2f _e . The final step is to estimate the pitch frequencies present in this vector to obtain an estimate of the carrier frequency error (step 106).

Due to aliasing, the estimated carrier frequency error has an error equal to Multiples of the ambiguity factor (ambiguity). The coarse frequency estimation stage can resolve this ambiguity factor by performing a process similar to that shown in Figure 1 after fine frequency correction. This is described in more detail below.

The method shown in Figure 1 will now be described in more detail with reference to signals using binary phase shift keying (BPSK) and spread modulation. This is for example purposes only. As explained above, the invention is limited to a specific form of modulation or spreading.

Binary data b _m is spread by some binary sequence S _k and then mapped to BPSK constellation points d _k ∈ {−1,+1}. The data can also be encoded differently if desired. The transmitted signal is then filtered by some pulse (eg root raised cosine) p(t). The signal sent is thus

where _Tc is the chip time (ie the reciprocal of the 3dB bandwidth _Bw ) and _Ns is the spreading factor.

The received signal experiences frequency errors due to crystal errors, channel effects, and frequency errors of the transmitted signal. This frequency error is written as f _e and can vary with time. The following description assumes that the frequency error is constant over the burst used to estimate the error. However, the principles of the present invention can also be applied to systems encountering varying channel conditions, where f _e may vary (this is described in more detail below).

The received signal can be written as:

The signal is match-filtered and one-tenth the chip rate. A discrete signal can be written as:

This assumes that the receiver has knowledge of the timing due to timing alignment on the transmission of the burst or from some pre-processing stage.

Equation 3 ignores the term used to account for the Nyquist pulse structure being corrupted by frequency errors. This will cause intersymbol interference, but only at very high frequency errors. It can thus be ignored in the analysis.

The processing for signal channel symbol k is first described. The receive vector for this channel symbol is written as:

where φ _k is the initial phase of the kth symbol.

The extension is removed by multiplying by the extension sign s _m . Since these can be ±1, the resulting received sequence can be written as:

Then for this sequence and time-reversed version Perform element-wise multiplication:

For BPSK modulation, for all k This gives:

As shown in Equation 7, multiplying each sample by its corresponding amount in the inverted version of the data block produces combined samples that all have the same phase. (In practice, these phase components may not be exactly the same; there may be relatively small variations across data blocks due to fluctuating frequency errors and channels). The average of the combined samples is then calculated. Any form of averaging may be used, eg mean, median, mode. The preferred option will calculate the average:

Since the combined samples have substantially the same phase, they are summed mostly coherently when calculating the average.

The phase increases by 2πf _e (N _s -1) between consecutive symbols. Therefore, φ _k+1 = φ _k + 2πf _e (N _s -1). This implies:

Substituting θ ₀ as the initialization phase gives:

The averaged samples thus form a sequence in which a carrier frequency error causes a phase change from one sample to the next. It is clear from equation 10 that the vector z of data is thus the pitch frequency 2f _e .

The problem has now been reduced to a pitch frequency estimation problem. There are several methods that can be used to estimate the frequency of this pitch, and any pitch estimation algorithm can be applied to the vector z. Preferably, a maximum likelihood estimator is used to estimate pitch. A practical way to achieve this may be via a Fourier transform. For example, an L-point FFT can be performed on the vector z. The bin with the largest absolute value is identified. The frequency corresponding to this block is determined as pitch frequency 2f _e . This is close to a maximum likelihood estimator, where the only difference between this maximum likelihood estimator and the true maximum likelihood estimator is the discrete nature of the blocks.

In the case of multiple receive antennas, the sum of squares of the results of the FFTs for each of the receive antennas can be summed to combine the results. This is only one example of how the results from multiple receive antennas might be combined, and it should be understood that the invention encompasses the use of any suitable method.

A varying carrier frequency error can also be assumed. The frequency error is preferably constant over the length of the spreading symbols. The maximum permissible frequency error deviation from the mean is mean can be range any value within . The method described herein estimates the average over the length of the spread symbol burst. After correcting for this, any off-frequency error should be small and able to be tracked by known tracking schemes (eg by using Costas loop phase tracking).

The frequency capture range depends on the sampling frequency. In the example described above, this means that the frequency capture range depends on the expansion factor N _s . This introduces aliasing with the Nyquist (or folding) frequency which is half the sampling frequency. The carrier frequency error estimate output by the procedure described above is in the range Inside. The extra factor of two on the denominator is because the tones in vector z have frequency 2f _e . Especially for larger expansion factors, the frequency capture range may be limiting.

The limited capture range means that the estimated error can include ambiguity factors. This can result in corrections being made to remove unwanted phase rotation in erroneous received data. Thus a coarse frequency estimation stage can be applied. This is shown in Figure 2 where a coarse phase (step 202) is applied after the fine phase (step 201). The frequency ambiguity factor output by the coarse estimation stage is added to the fine estimate to generate a complete estimate of the carrier frequency error (step 203).

The blur factor can only take certain values. The range of these values for l = -1, 0, +1 etc. is The number of valid l values is a design choice based on the expected range of frequency errors to be handled. For illustration purposes only, a method for resolving this ambiguity factor is described below with reference to l=-1,0,+1.

The rationale behind the coarse phase is to create a new tone signal using the same technique outlined for fine frequency estimation. This is illustrated in FIG. 3 . Fine frequency error compensation is first applied to the data to remove the fine estimate of error (step 301). Coarse frequency estimation is likely to use data blocks of different lengths from fine frequency estimation to obtain a larger frequency capture range. Steps 302 and 303 are thus essentially a repetition of steps 101 and 102 but for different sampling rates. In one example, the _Ns chips for each channel symbol are divided into a pair of samples. The rough estimate then continues with the same processing steps as in Figure 1 (steps 304 to 306) but for a different sampling frequency to obtain a new pitch signal In the example where each channel symbol is divided into a pair of samples, this essentially results in repeating the process arranged above in Equations 4 to 10 but with N _s =2. tone signal Noisier than z, but we already have where we expect the pitch to be (i.e., for l=-1,0,+1 at nearby) knowledge. If there is a small error in the fine frequency errors that have been estimated, it is unlikely to be exactly at any of those frequencies.

The coarse frequency estimation stage deviates from the fine frequency estimation stage when it starts estimating pitch. tone signal Most likely noisy, but the frequency of the tone needs to be identified only to the extent necessary to identify which of the valid l values is the correct value. Therefore, different estimation techniques are likely to be applied from the fine estimation stage to the coarse estimation stage. In the example in Figure 3, for the chosen value of l, first pass mixed up (step 307). The resulting signal is then passed through a low pass filter (step 308). The bandwidth of this filter can be set based on the expected residual frequency error and Doppler spread of the channel. For example, a simple moving average filter can be used. The power contained in the resulting signal is calculated to obtain a power value _PI (step 309). Steps 306 to 308 are repeated for each selected value of 1 (step 310). Finally, the rough frequency is estimated by picking the value of _l that gives the maximum power value PI (step 311). Other selection methods may be used. However, this method is both simple to implement and close to a maximum likelihood estimator.

If there are multiple receive antennas, the respective _PI values for each value of 1 can be generated by summing the respective _PI values from the respective receive antennas for a given value of 1.

The means for implementing the invention are likely to be implemented as part of a receiver of a wireless communication system. Such a receiver may typically include RF circuitry well known for implementing wireless communications, including, for example, an antenna system, an RF transceiver, one or more amplifiers, a tuner, one or more oscillators, a digital signal processor, memory etc.

An example of a device is shown in FIG. 4 . The apparatus shown generally at 401 is configured to receive a signal r (402). Although the "received" signal may be the signal directly from the antenna, it may also have been subjected to some pre-processing before being input to the device.

The apparatus comprises a first sampling unit 402 for forming a data block. The device also comprises a product unit 403 for forming the vector z of the mean values. Also these are referred to as step I in FIG. 4 corresponding to steps 101 to 105 in FIG. 1 . The product unit 403 is followed by a pitch detector 404, which is configured to calculate the fine frequency error This is referenced as step II in FIG. 4 and corresponds to step 106 in FIG. 1 . The fine frequency error is output to summing unit 405 .

The device may also be configured to perform coarse frequency estimation, but this functionality may be disabled if the frequency acquisition range is deemed sufficient such that coarse frequency estimation is not necessary.

means comprising a fine frequency error configured to align the received signal with the estimated Mixer 406 for mixing. The apparatus also includes a second sampling unit and a second multiplying unit (407 and 408) configured to output another vector of average values to a set of parallel processing paths for estimating blur factors in the refined estimate. Each path includes a blur blending unit 409 . These are connected to a low pass filter 410 and a power calculation unit 411 . The final arrangement comprises a maximum value identification unit configured to receive the power values output by the power calculation unit and to identify the frequency associated with the maximum of those power values as a coarse frequency estimate. The maximum value identification unit is configured to output the rough frequency estimate to the summation unit 405 .

It is possible to implement the parallel processing paths illustrated in FIG. 4 sequentially as well. Similarly, it should be understood that other components may perform some of all their processing in parallel (eg, product calculations performed across multiple data blocks by product unit 403, second product unit 408).

The structure shown in Figure 4 is intended to correspond to a number of functional blocks. This is for illustrative purposes only. FIG. 4 is not intended to define a strict division between different parts of hardware on a chip or between different programs, procedures or functions in software. In some implementations, some or all of the algorithms described herein may be implemented in whole or in part in hardware. In many implementations, at least a portion of the communication means may be implemented by a processor acting under software control. Any such software is preferably stored on a non-transitory computer readable medium, such as memory (RAM, cache, hard disk, etc.) or other storage (USB stick, CD, diskette, etc.).

The method described herein allows accurate carrier frequency offset estimation even in very low signal-to-noise environments, especially for systems employing extended BPSK. These methods are applicable both to systems experiencing flat fading and to multipath channels as well. The frequency capture range is also increased over the current estimation algorithm to approach chip frequencies rather than symbol frequencies. These methods can also be applied to multiple antenna systems.

The methods described herein can be applied to communication networks constructed for Internet of Things (IoT) communication. Examples would include networks configured to operate according to the Weightless ^™ protocol (although the methods described herein could be readily implemented by networks configured to operate according to different protocols (eg LTE, Bluetooth, WiFi, VoIP)). Typically, a network will include a number of communication devices (eg, base stations) each configured to communicate with a large number of geographically spaced terminals. The communication means described herein can be realized only by such communication devices or terminals. The network may be a cellular network, where each communication device is responsible for over-the-air communication with terminals located in the respective cell. The communication device suitably communicates with the core network via a wired or wireless interface, and may function at least in part under the control of the core network.

In one example, the communication device described herein may be configured to operate according to the Weightless ^™ IoT specification. Weightless ^TM uses a cellular WAN architecture where the protocol is optimized for the requirements of IoT systems (low terminal cost, low terminal duty cycle and thus low power consumption, and scalability to very low data rates). It was originally designed to operate in the TV white space from 470MHz to 790MHz, but the PHY was promoted to operate in licensed bands of varying bandwidth, shared licensed access bands and license exempted bands.

The applicant therefore discloses each individual feature described herein in isolation, as well as any combination of two or more such features, to the extent that such feature or combination can generally be based on the present invention in light of the common knowledge of a person skilled in the art. The description is carried out regardless of whether such a feature or a combination of these features solves any problem disclosed herein, and is not limited to the scope of the claims. The applicant indicates that aspects of the invention may comprise any such feature alone or in combination. In view of the foregoing description it will be apparent to a person skilled in the art that various modifications may be made within the scope of the invention.

Claims (15)

1. a kind of method of unknown errors for estimating carrier frequency, this approach includes the following steps：

The signal of modulated version as the carrier frequency is sampled to form multiple data blocks, each data block includes Multiple samples and each sample incorporates the phase error caused by the carrier frequency error；

For each data block：

The product of each sample and the data another sample in the block is generated, sample is described described in another described sample picture Equally it is located in the same position of the data block reversed in version in data block, is respectively incorporated due to the load with being formed The vector of the combined sample of identical phase error caused by wave frequency rate error；And

The average value of vector of the combined sample is calculated to form the vector of average sample；And

Estimate the carrier frequency by from the pitch frequency in the vector for multiple average samples that the multiple data block generates Rate error.

2. according to the method described in claim 1, the method includes by the combined sample summation and divided by it is described The quantity of data sample in the block calculates the average value of the combined sample.

3. method according to claim 1 or 2, wherein the signal is the carrier wave modulated by extension The version of frequency, the method includes using the sampling rate for being equal to the spreading rate for being used for extending the signal to come to the letter It number is sampled.

4. method according to claim 1 or 2, wherein the signal is the carrier wave modulated by extension The version of frequency, the method includes being formed as including respectively many samples equal to spreading factor by the data block.

5. method according to claim 1 or 2, the method includes using maximum likelihood estimator module to estimate the carrier wave Frequency error.

6. method according to claim 1 or 2, the method includes estimating that the carrier frequency is missed by following steps Difference：

Fourier transformation is executed to the multiple average sample；And

Select have the frequency of peak as the carrier frequency error in the transformation.

7. according to the method described in claim 6, the method includes for mutiple antennas：

For each signal received by the multiple antenna, multiple average samples are generated；

To every group of average sample, Fourier transformation is executed；And

The multiple transformation obtained are combined by quadratic sum is summed.

8. method according to claim 1 or 2, wherein estimated carrier frequency error incorporates fuzzy factor, described Method includes：Fine estimation stages obtain the fine valuation of the carrier frequency error in the fine estimation stages, and The rough estimate stage parses the fuzzy factor in the fine valuation in the rough estimate stage.

9. according to the method described in claim 8, wherein：

The fine estimation stages include being generated according to the method for claim 1 and by using the first sample frequency The data block and generate fine valuation；And

The rough estimate stage include generated by according to the method for claim 1 rough valuation and by using Second sample frequency parses the fuzzy factor to generate the data block, and second sample frequency is adopted less than described first Sample frequency.

10. making with by the fine estimation stages according to the method described in claim 9, the rough estimate stage includes use Technology it is different for estimating the technology of the carrier frequency error to generate the rough valuation.

11. according to the method described in claim 8, the method includes for the rough estimate stage by following steps come Generate the rough valuation of the carrier frequency error：

Make the multiple average sample and two or more frequency compoundings, each frequency in the frequency is described fuzzy The multiple of the factor；

Determine the signal power that each signal in mixed signal includes；And

Select to generate rough valuation of the frequency as the carrier frequency error of the mixed signal including highest signal power.

12. method according to claim 1 or 2, the method includes to estimated carrier frequency error and the load The rough valuation of wave frequency rate error is summed, and the complete of the carrier frequency error of fuzzy factor has wherein been parsed with formation and has been estimated Value.

13. method according to claim 1 or 2, wherein the signal is the carrier wave for having used phase-shift keying (PSK) to modulate The version of frequency.

14. a kind of device of unknown errors for estimating carrier frequency, which is configured to realize and be wanted according to right The method described in any one of 1 to 12 is asked to carry out the unknown errors in estimating carrier frequency.

15. a kind of device of unknown errors for estimating carrier frequency, the device include：

Sampling unit, the sampling unit are configured to sample with shape the signal of the modulated version as the carrier frequency At multiple data blocks, each data block includes multiple samples and each sample is incorporated due to the carrier frequency error and led The phase error of cause；

Product unit, the product unit are configured to, for each data block, it is in the block another with the data to generate each sample The product of one sample, another sample sample as described in are located at the reversing version of the data block in the data block In same position in this, respectively the identical phase error caused by the carrier frequency error is incorporated to be formed The vector of combined sample, and the vectorial average value of the combined sample is calculated to form the vector of average sample；And

Pitch estimator, the pitch estimator are configured to using the multiple average samples generated from the multiple data block Vector in pitch frequency estimate the carrier frequency error.