A frequency offset estimation and compensation method for ultra-narrow band communication
By using the method of cross-correlation and summation of ZCM sequences and differential conjugate product matrices, the problem of insufficient frequency offset estimation accuracy under low signal-to-noise ratio environment is solved, and the frequency offset compensation performance is improved.
Patent Information
- Application Number
- CN202311017713.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-08-14
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2043-08-14
AI Technical Summary
Existing frequency offset estimation methods lack accuracy in low signal-to-noise ratio environments, resulting in poor frequency offset compensation performance.
Frequency offset estimation is performed using a newly constructed ZCM sequence. Combining fractional and integer multiple frequency offset estimation, a trade-off between performance and computational complexity is achieved by cross-correlation and summation of differential conjugate product matrices, thereby improving the frequency offset compensation performance under low signal-to-noise ratio.
It significantly improves the accuracy and compensation effect of frequency offset estimation in low signal-to-noise ratio environments, ensuring communication stability.
Smart Images

Figure CN117040991B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of communication technology, and more specifically, to a method for frequency offset estimation and compensation in ultra-narrowband communication. Background Technology
[0002] In recent years, with the rapid development of wireless communication services, available spectrum resources have become increasingly scarce. Therefore, how to efficiently utilize limited spectrum has become a research focus. Ultra-narrowband communication (UNB) is a technology that can provide extremely high spectrum utilization. It employs advanced coding techniques to minimize the waveform differences between modulation symbols, thereby achieving high spectrum utilization. In UNB communication, signals have the characteristics of long transmission distance and strong penetration. However, due to oscillator mismatch or Doppler shift, carrier frequency offset problems arise. Therefore, to ensure communication stability, research on frequency offset estimation and compensation techniques is necessary.
[0003] Numerous studies have been conducted on frequency offset estimation methods, which can be broadly categorized into two types based on whether or not they require special sequence assistance: frequency offset estimation methods based on special sequence assistance and blind estimation methods based on non-special sequence assistance. Blind estimation methods based on non-special sequence assistance mainly include those using cyclic prefixes and those utilizing virtual subcarriers. The main characteristic of these algorithms is high spectral efficiency, but their computational complexity is relatively high. Frequency offset estimation methods based on special sequence assistance are characterized by the use of special sequences, such as pilots, preambles, or training sequences. The main principle of these algorithms is to design a special sequence at the transmitting end, and then, after receiving the signal, the receiving end extracts this sequence and performs correlation operations with the transmitting end's sequence to obtain the estimated frequency offset value of the received signal. The advantages of blind frequency offset estimation methods based on special sequence assistance are fast acquisition speed and high accuracy, making them very suitable for frequency offset estimation and compensation in environments with low signal-to-noise ratios.
[0004] However, existing frequency offset estimation algorithms still have shortcomings, the most important of which is how to improve the accuracy of frequency offset estimation at the receiver in low signal-to-noise ratio environments, thereby improving the performance of frequency offset compensation. Summary of the Invention
[0005] This invention addresses the carrier frequency offset problem in ultra-narrowband communication. Based on a newly constructed ZCM sequence, leveraging its constant amplitude characteristic and good correlation, the receiver can effectively determine the preamble index number without prior information and perform frequency offset estimation accordingly. Furthermore, it provides an integer multiple frequency offset estimation function based on the difference conjugate product matrix of the local frequency domain sequence and the difference conjugate product matrix of the received frequency domain sequence. The balance between estimation performance and computational complexity can be achieved by adjusting the number of rows in the cross-correlation summation of the difference conjugate product matrix. By combining fractional and integer multiple frequency offset estimation, the frequency offset compensation performance in low signal-to-noise ratio environments is improved.
[0006] To achieve the above objectives, the technical solution of the present invention is as follows:
[0007] A frequency offset compensation algorithm for ultra-narrowband communication includes the following steps:
[0008] S1: Input the received signal after timing synchronization and demodulation. Extract the preamble sequence from it. ;
[0009] S2: Determine the preamble sequence The index number is used to obtain the preamble sequence corresponding to the local index number. ;
[0010] S3: Perform fractional octave frequency offset estimation to obtain the fractional octave frequency offset estimate. ;
[0011] S4: Perform fractional-fold frequency offset compensation on the received signal;
[0012] S5: Perform integer multiple frequency offset estimation to obtain the integer multiple frequency offset estimate. ;
[0013] S6: Perform integer multiple frequency offset compensation on the received signal;
[0014] S7: Frequency offset compensation ends, output results.
[0015] Preferably, the design of the preamble in step S1 is as follows:
[0016] Using M-sequence Scrambled ZC sequences The obtained ZCM sequence is expressed as follows:
[0017]
[0018] in, Here is the ZC sequence expression, where N is the sequence length. This is the root sequence index. The sequence is generated by a linear feedback shift register, and the generation expression is: The initial state is set as .
[0019] Preferably, in step S2, the preamble sequence is determined. The index number is as follows:
[0020] S2.1: All local sequences One-time and received sequence After performing relevant operations, the power time delay spectrum is obtained, and its expression is: ;
[0021] S2.2: The peak value of the power delay spectrum is denoted as... ,like If the index number of the local sequence is incorrect, the next local sequence correlation operation is performed until a correlation peak is found. The index number corresponding to this local sequence This is the preamble index number of the signal currently received by the receiver;
[0022] S2.3: Using the index number Preamble found .
[0023] Preferably, the threshold in step S2.2 The specific settings are as follows:
[0024]
[0025] In the formula, The threshold constant has a range of [value missing]. In the detection of the preamble index number, traverse from smallest to largest. This continues until the minimum threshold constant that satisfies the set detection miss rate is obtained.
[0026] Preferably, the method for estimating the fractional octet frequency offset in step S3 is as follows:
[0027] S3.1: The preamble sequence in the received signal Preamble corresponding to the local index number Perform conjugate multiplication to calculate the relevant sequence, i.e. ;
[0028] S3.2: Take the first half of the relevant sequence and the second half Perform conjugate correlation to obtain Its expression is:
[0029] S3.3: Calculate the fractional octet frequency offset estimate .
[0030] Preferably, the expression for compensating for the fractional frequency offset of the received signal in step S4 is:
[0031]
[0032] Preferably, the method for estimating integer multiples of frequency offset in step S5 is as follows:
[0033] S5.1: Extract the preamble sequence from the received signal after fractional-fold frequency offset compensation. And perform DFT on it, i.e. The frequency domain sequence is obtained. ;
[0034] S5.2: For Perform a cyclic shift to obtain a cyclically shifted receive sequence. ,use Construct the receive differential conjugate product matrix ;
[0035] S5.3: Similarly, for local time-domain preamble sequences Perform DFT, i.e. The local frequency domain sequence is obtained. ,use Construct the difference conjugate product matrix ;
[0036] S5.4: Using the local difference conjugate product matrix and receive differential conjugate product matrix The former Cross-correlation is performed on the row elements, and their absolute values are summed to obtain the cross-correlation absolute sum function. Its expression is ,in It is the number of rows in the summation of cross-correlation values. ;
[0037] S5.5: Obtain the integer frequency offset estimation function ,in The normalization function is expressed as follows: ;
[0038] S5.6: Estimation function for integer multiples of frequency offset Perform peak search to obtain integer multiples of frequency offset estimates. The expression is: .
[0039] Preferably, in step S5.2, the frequency domain received sequence is... The circular shift expression is:
[0040]
[0041] Circular shift sequence is Receive differential conjugate product matrix The construction is shown in the following formula:
[0042]
[0043] Preferably, the difference conjugate product matrix in step S5.3 The expression is:
[0044]
[0045] Preferably, the expression for compensating the received signal for integer multiples of frequency offset in step S6 is:
[0046] Attached Figure Description
[0047] Figure 1 This is a schematic diagram of the method flow of the present invention.
[0048] Figure 2 This is a schematic diagram comparing the frequency offset estimation performance of different algorithms provided in the example.
[0049] Figure 3 This is a schematic diagram comparing the frequency offset estimation performance under different sequence lengths provided in the example. Detailed Implementation
[0050] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent.
[0051] To better illustrate this embodiment, some parts in the accompanying drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions.
[0052] It will be understood by those skilled in the art that certain well-known structures and their descriptions may be omitted in the accompanying drawings.
[0053] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.
[0054] Example 1
[0055] This example provides a frequency offset estimation and compensation method based on ultra-narrowband communication, such as... Figure 1 As shown, it includes the following steps:
[0056] S1: Input the received signal after timing synchronization and demodulation. Extract the preamble sequence from it. ;
[0057] S2: Determine the preamble sequence The index number is used to obtain the preamble sequence corresponding to the local index number. ;
[0058] S3: Perform fractional octave frequency offset estimation to obtain the fractional octave frequency offset estimate. ;
[0059] S4: Perform fractional-fold frequency offset compensation on the received signal;
[0060] S5: Perform integer multiple frequency offset estimation to obtain the integer multiple frequency offset estimate. ;
[0061] S6: Perform integer multiple frequency offset compensation on the received signal;
[0062] S7: Frequency offset compensation ends, output results.
[0063] Example 2
[0064] This embodiment, based on Embodiment 1, continues to disclose the following content:
[0065] The design of the preamble in step S1 is as follows:
[0066] Using M-sequence Scrambled ZC sequences The obtained ZCM sequence is expressed as follows:
[0067]
[0068] in, Here is the ZC sequence expression, where N is the sequence length. This is the root sequence index. The sequence is generated by a linear feedback shift register, and the generation expression is: The initial state is set as .
[0069] In step S2, the preamble sequence is determined. The index number is as follows:
[0070] S2.1: All local sequences One-time and received sequence After performing relevant operations, the power time delay spectrum is obtained, and its expression is: ;
[0071] S2.2: The peak value of the power delay spectrum is denoted as... ,like If the index number of the local sequence is incorrect, the next local sequence correlation operation is performed until a correlation peak is found. The index number corresponding to this local sequence This is the preamble index number of the signal currently received by the receiver;
[0072] S2.3: Using the index number Preamble found .
[0073] Threshold in step S2.2 The specific settings are as follows:
[0074]
[0075] In the formula, The threshold constant has a range of [value missing]. In the detection of the preamble index number, traverse from smallest to largest. This continues until the minimum threshold constant that satisfies the set detection miss rate is obtained.
[0076] The method for estimating the fractional frequency offset in step S3 is as follows:
[0077] S3.1: The preamble sequence in the received signal Preamble corresponding to the local index number Perform conjugate multiplication to calculate the relevant sequence, i.e. ;
[0078] S3.2: Take the first half of the relevant sequence and the second half Perform conjugate correlation to obtain Its expression is:
[0079] S3.3: Calculate the fractional octet frequency offset estimate .
[0080] The expression for compensating the fractional frequency offset of the received signal in step S4 is as follows:
[0081]
[0082] The method for estimating integer frequency offsets in step S5 is as follows:
[0083] S5.1: Extract the preamble sequence from the received signal after fractional-fold frequency offset compensation. And perform DFT on it, i.e. The frequency domain sequence is obtained. ;
[0084] S5.2: For Perform a cyclic shift to obtain a cyclically shifted receive sequence. ,use Construct the receive differential conjugate product matrix ;
[0085] S5.3: Similarly, for local time-domain preamble sequences Perform DFT, i.e. The local frequency domain sequence is obtained. ,use Construct the difference conjugate product matrix ;
[0086] S5.4: Using the local difference conjugate product matrix and receive differential conjugate product matrix The former Cross-correlation is performed on the row elements, and their absolute values are summed to obtain the cross-correlation absolute sum function. Its expression is ,in It is the number of rows in the summation of cross-correlation values. ;
[0087] S5.5: Obtain the integer frequency offset estimation function ,in The normalization function is expressed as follows: ;
[0088] S5.6: Estimation function for integer multiples of frequency offset Perform peak search to obtain integer multiples of frequency offset estimates. The expression is: .
[0089] In step S5.2, the frequency domain received sequence The circular shift expression is:
[0090]
[0091] Circular shift sequence is Receive differential conjugate product matrix The construction is shown in the following formula:
[0092]
[0093] The difference conjugate product matrix in step S5.3 The expression is:
[0094]
[0095] The expression for compensating the received signal for integer multiples of frequency offset in step S6 is as follows:
[0096]
[0097] Example 3
[0098] Based on Embodiments 1 and 2, this embodiment provides the following specific embodiments:
[0099] The simulation parameters are set as follows: the channel used is additive white Gaussian noise (AWGN), the signal-to-noise ratio (SNR) is set to -20dB to -10dB, the bandwidth is 100Hz, the integer multiple frequency offset is 300Hz, the fractional multiple frequency offset is 50Hz, and the number of Monte Carlo iterations is 10,000.
[0100] The frequency offset estimation performance of the novel sequence proposed in this invention is compared with that of the preamble sequence based on the ZC sequence. Meanwhile, the mean square error (MSE) of the estimation results is calculated using the unestimated frequency offset as a reference. The details are as follows: Figure 2 As shown, the ZC sequence curve corresponds to the frequency offset estimation method using the preamble based on the ZC sequence; the ZCM sequence curve corresponds to the frequency offset estimation method using the new sequence proposed in this invention; the curve without estimation corresponds to the method without frequency offset estimation as a reference. The MSE of both frequency offset estimation methods decreases with increasing signal-to-noise ratio (SNR), and is significantly lower than the MSE without frequency offset estimation within the SNR range, demonstrating the effectiveness of the frequency offset estimation. Furthermore, under low SNR conditions, the frequency offset estimation method designed in this paper is superior, with a lower MSE curve. However, when the SNR is greater than -15dB, the performance difference between the two frequency offset estimation methods is not significant. This is because the noise interference to the preamble is reduced, allowing both methods to almost completely estimate the frequency offset.
[0101] Based on this, while keeping the simulation parameters unchanged, the length of the preamble sequence in the method proposed in this invention is changed. MSE is used to quantitatively analyze the frequency offset compensation effect of preamble sequences with lengths of 64, 128, 256, and 1024, as detailed below: Figure 3 As shown, the 1024 curve corresponds to a sequence of length 1024; the 512 curve corresponds to a sequence of length 512; the 128 curve corresponds to a sequence of length 128; and the 64 curve corresponds to a sequence of length 64. The MSE of the frequency offset estimation for all four sequence lengths decreases with increasing signal-to-noise ratio, indicating more accurate estimation. Furthermore, the frequency offset estimation performance improves with increasing sequence length, especially at low signal-to-noise ratios.
[0102] The above examples demonstrate that in AWGN channels with significant frequency offsets and low signal-to-noise ratios of the received signals, the method described in this invention can achieve significantly better compensation results, and can also perform frequency offset compensation well even at high signal-to-noise ratios.
[0103] The same or similar labels correspond to the same or similar parts;
[0104] The positional relationships depicted in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this patent.
[0105] Obviously, the above examples of the present invention are merely illustrative of the invention and are not intended to limit the implementation of the invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.
Claims
1. A method for frequency offset estimation and compensation in ultra-narrowband communication, characterized in that, Includes the following steps: S1: Input the received signal after timing synchronization and demodulation. Extract the preamble sequence from it. , The sequence length is given; the preamble design uses an M-sequence. Scrambled ZC sequences The obtained ZCM sequence is expressed as follows: in, For ZC sequence expressions, For the root sequence index; The sequence is generated by a linear feedback shift register, and the generation expression is: The initial state is set as ; S2: Determine the preamble sequence based on the set peak detection threshold. The index number is used to obtain the preamble sequence corresponding to the ZCM index number. ; S3: Perform fractional octave frequency offset estimation to obtain the fractional octave frequency offset estimate. The method for fractional frequency offset estimation specifically involves taking the preamble sequence from the received signal... Preamble sequence corresponding to the index number of the ZCM sequence Perform conjugate multiplication to calculate the relevant sequence, i.e. ; the first half of the relevant sequence and the second half Perform conjugate correlation to obtain Its expression is: Calculate the fractional frequency offset estimate ; S4: Perform fractional-fold frequency offset compensation on the received signal; S5: Extract the preamble sequence from the received signal after fractional frequency offset compensation. And perform DFT on it, i.e. The frequency domain sequence is obtained. ;right Perform a cyclic shift to obtain a cyclically shifted receive sequence. ,use Construct the receive differential conjugate product matrix For local time-domain preamble sequences Perform DFT, i.e. The local frequency domain sequence is obtained. ,use Construct the difference conjugate product matrix Use the local difference conjugate product matrix and receive differential conjugate product matrix The former Cross-correlation is performed on the row elements, and their absolute values are summed to obtain the cross-correlation absolute sum function. Its expression is , It is the number of rows in the summation of cross-correlation values. ; Obtain the integer frequency offset estimation function ,in The normalization function is expressed as follows: For integer multiples of frequency offset estimation function Perform peak search to obtain integer multiples of frequency offset estimates. The expression is: ; S6: Perform integer multiple frequency offset compensation on the received signal; S7: Frequency offset compensation ends, output results.
2. The frequency offset estimation and compensation method for ultra-narrowband communication according to claim 1, characterized in that, In step S2, the preamble sequence is determined. The index number is as follows: S2.1: All ZCM sequences One with preamble sequence After performing relevant operations, the power time delay spectrum is obtained, and its expression is: ; S2.2: The peak value of the power delay spectrum is denoted as... ,like If the index number of the ZCM sequence is incorrect, the correlation operation of the next ZCM sequence is performed until a correlation peak is found. The index number corresponding to this ZCM sequence This is the preamble index number of the signal currently received by the receiver. For the threshold; S2.3: Using the index number Preamble sequence found .
3. The frequency offset estimation and compensation method for ultra-narrowband communication according to claim 2, characterized in that, Threshold in step S2.2 The specific settings are as follows: , In the formula, The threshold constant has a range of [value missing]. In the detection of the preamble index number, traverse from smallest to largest. This continues until the minimum threshold constant that satisfies the set detection miss rate is obtained.
4. The frequency offset estimation and compensation method for ultra-narrowband communication according to claim 3, characterized in that, The expression for compensating for the fractional frequency offset of the received signal in step S4 is as follows: 。 5. The frequency offset estimation and compensation method for ultra-narrowband communication according to claim 1, characterized in that, In step S5, the frequency domain received sequence The circular shift expression is: , The cyclic shift sequence is: Receive differential conjugate product matrix The construction is shown in the following formula: 。 6. The frequency offset estimation and compensation method for ultra-narrowband communication according to claim 1, characterized in that, The difference conjugate product matrix in step S5 The expression is: 。
Citation Information
Patent Citations
Maximum correlation estimation detection method based on narrowband Internet of Things random access channel
CN108924943A
Carrier synchronization method of low-earth-orbit satellite broadband OFDM communication system
CN111147123A
