Digital signal processing method based on channel response-maximum ratio combination

Abstract

The invention relates to the technical field of wireless channels, and particularly provides a digital signal processing method based on channel response-maximum ratio combination, which comprises the following steps of: inserting a training symbol into each frame of a multi-carrier signal, and performing channel estimation on each training signal to obtain a corresponding channel response estimated value; estimating the SNR value of each training symbol, and calculating a normalized weight coefficient based on the SNR values; and performing weighted average based on the normalized weight coefficient of each training symbol and the channel response estimation value to obtain an overall channel estimation result of the multi-carrier signal. According to the method, the time diversity characteristic contained in continuous training symbol transmission is fully utilized, the estimation noise is effectively suppressed, and the system robustness is improved on the premise that the calculation overhead does not need to be remarkably increased. Besides, the method does not need to acquire channel statistical characteristics in advance, so that the method is particularly suitable for a wireless communication environment with dynamic and rapid change, and good balance among accuracy, calculation efficiency and practicability is realized.

Description

Technical Field

[0001] The present invention relates to the field of wireless channel technology, and in particular to channel estimation and channel equalization technology for Orthogonal Frequency Division Multiplexing (OFDM) / Discrete Multi-Carrier (DMT) signals, Maximum Ratio Combining (MRC) technology, and a digital signal processing algorithm at the transceiver end of a communication system. Background Art

[0002] As wireless communication technology advances from 5G to 6G, the demand for long-distance, high-capacity data transmission is growing. The terahertz (THz) frequency band is seen as a disruptive solution for future networks. Its unprecedented bandwidth is expected to meet the growing demand for ultra-high-speed, wide-coverage communications [Zhu M, Zhang J, Hua B, et al Ultra-wideband fiber-THz-fiber seamless integration communication system toward 6G: architecture, key techniques, and testbed implementation. Sci China InfSci, 2023, 66:1-18.]. This frequency band, covering approximately 0.1–10 THz, has the potential to achieve data rates of hundreds of Gbps, enabling bandwidth-demanding applications such as augmented reality, holographic communications, and large-scale machine-type communications. However, to fully realize the potential of THz communications, numerous technical challenges in high-frequency signal propagation must be overcome. Orthogonal frequency division multiplexing (OFDM) has become a key technology in emerging THz communication systems due to its robustness against frequency-selective fading and intersymbol interference. OFDM technology is mature and easily integrated with existing baseband processing architectures, making it widely considered an ideal choice for achieving high-data-rate THz communications. However, achieving stable and reliable systems in the THz band remains challenging, with accurate and efficient channel estimation being a key bottleneck limiting system performance. THz channels exhibit propagation characteristics significantly different from those of microwave and millimeter-wave channels, such as high propagation loss, molecular absorption, and severe phase noise. These factors significantly degrade the performance of traditional channel estimation algorithms. Traditional channel estimation methods, such as least squares (LS), are widely adopted due to their computational simplicity and suitability for real-time processing. However, under the low signal-to-noise ratio (SNR) conditions commonly found in THz communications, these methods are susceptible to significant noise amplification, severely reducing estimation accuracy. Furthermore, simply averaging LS estimates obtained from multiple training symbols offers limited performance improvement due to the large variations in SNR across training symbols. Therefore, there is an urgent need to develop new channel estimation methods that overcome these challenges while combining low complexity with high accuracy. Summary of the Invention

[0003] In view of this, the present invention proposes a digital signal processing method based on channel response-maximum ratio combining to solve the problem that the traditional LS algorithm is extremely susceptible to noise interference, resulting in a significant increase in estimation error. At the same time, it breaks through the bottleneck that the stability of LS estimation will be greatly reduced when the training symbol positions are unevenly distributed or the signal-to-noise ratios corresponding to different training symbols differ greatly.

[0004] To achieve the above object, the present invention proposes a digital signal processing method based on channel response-maximum ratio combining, which is characterized by comprising:

[0005] Inserting a training symbol into each frame of a multi-carrier signal, performing channel estimation on each of the training signals, and obtaining a corresponding channel response estimation value;

[0006] estimating an SNR value of each of the training symbols, and calculating a normalized weight coefficient based on the SNR value;

[0007] A weighted average is performed based on the normalized weight coefficient of each training symbol and the channel response estimation value to obtain an overall channel estimation result of the multi-carrier signal.

[0008] Furthermore, the multi-carrier signal is an OFDM signal or a DMT signal.

[0009] Furthermore, the number of inserted training symbols is determined according to the number of data symbols of the multi-carrier signal.

[0010] Furthermore, the channel estimation adopts LS algorithm.

[0011] Furthermore, the calculation method of the normalized weight coefficient is as follows:

[0012]

[0013] Among them, w i is the normalized weight coefficient of the i-th training symbol, SNR i is the SNR value of the i-th training symbol, and n is the total number of training symbols.

[0014] Furthermore, the calculation method of the overall channel estimation result is as follows:

[0015]

[0016] Among them, H MRC is the overall channel estimation result, H LS,i is the channel estimation value of the i-th training symbol, w iis the normalized weight coefficient of the i-th training symbol, and n is the total number of training symbols.

[0017] The present invention also proposes the application of the above method in a photon-assisted millimeter wave / terahertz wireless communication system, including: using the above method to perform digital signal processing on the oscilloscope's sampled data at the receiving end, restoring the original transmitted symbols, and evaluating the performance improvement of the system.

[0018] Compared with the prior art, the present invention has the following beneficial effects:

[0019] The present invention achieves the synergistic complementarity of the low complexity advantage of LS estimation and the noise resistance of diversity combining technology by performing SNR weighted averaging on the LS estimation results of multiple training symbols. The present invention fully utilizes the time diversity characteristics inherent in the transmission of continuous training symbols to effectively suppress estimation noise, thereby improving system robustness without significantly increasing computational overhead. In addition, the method does not require the pre-acquisition of channel statistical characteristics, making it particularly suitable for dynamic and rapidly changing wireless communication environments. Therefore, the CR-MRC method proposed in the present invention achieves a good balance between accuracy, computational efficiency and practicality, providing a feasible approach for efficient channel estimation in next-generation wireless communication systems.

[0020] The parameters of the CR-MRC algorithm proposed in the present invention, such as the number and distribution of training symbols, the channel estimation method for each training symbol, the method for merging the channel estimation results, etc., can be specifically selected according to different systems, different channel characteristics, different transmission indicators and other specific conditions, and have good adjustability and applicability.

[0021] The method of the present invention has good versatility and is applicable to various application scenarios such as photon-assisted millimeter-wave terahertz systems, all-solid-state circuit millimeter-wave terahertz systems, polarization-multiplexed coherent transmission systems, and multi-mode / multi-core optical fiber transmission systems. At the same time, the CR-MRC algorithm of the present invention can be used to improve the demodulation performance of both OFDM and DMT signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Various other advantages and benefits will become apparent to those skilled in the art by reading the detailed description of the preferred embodiment below. The accompanying drawings are only for the purpose of illustrating the preferred embodiment and are not to be considered as limiting the present invention. In the accompanying drawings:

[0023] Figure 1 Schematic diagram of the overall process of the method of the present invention;

[0024] Figure 2 This is a flowchart of OFDM modulation and demodulation using the method of the present invention in an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, unless there is a conflict, the embodiments of the present disclosure and the features described in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0026] This embodiment proposes a digital signal processing method based on channel response-maximum ratio combining, such as Figure 1 As shown, including:

[0027] Training symbols are extracted from the received frequency-domain OFDM signal and the demodulation signal-to-noise ratio (SNR) of each training symbol is calculated to evaluate the accuracy of the channel estimation. Based on the SNR, a weighted average of the LS channel estimation results for multiple training symbols is performed to obtain a more accurate channel estimation result. This allows channel equalization and demodulation of the OFDM data symbols to be performed, improving OFDM system performance.

[0028] The key part of the present invention is the CR-MRC algorithm, which works as follows:

[0029] In OFDM systems, accurate channel estimation is the key to achieving reliable signal recovery. The traditional least squares estimator is widely used due to its low computational complexity. Its basic form can be expressed as:

[0030]

[0031] Where Y represents the received training symbol, X is the known transmitted training symbol, H is the actual channel response, and N represents additive white Gaussian noise (AWGN). However, the LS estimator performs poorly under low signal-to-noise ratio conditions because it fails to utilize the statistical characteristics of noise, resulting in large channel estimation errors. In addition, the SNRs corresponding to different training symbols may vary greatly. If the LS estimates of multiple training symbols are directly averaged, it often only brings very limited performance improvements. The CR-MRC algorithm fully exploits the SNR diversity gain between multiple training symbols, thereby improving the accuracy of channel estimation and system robustness. The algorithm mainly includes the following three steps:

[0032] Step 1: LS channel estimation of multiple training symbols:

[0033] Multiple training symbols are inserted into each OFDM frame. The number of training symbols can be determined based on the number of data symbols. This improves the accuracy of channel estimation while ensuring low overhead. In an OFDM data frame, training symbols can be distributed at equal intervals or inserted at unequal intervals in different ways, which provides high flexibility.

[0034] Perform independent LS estimation on each training symbol to obtain the corresponding channel response estimation value. That is, for the i-th training symbol, calculate its LS estimation value H LS,i , where i = 1, 2, ..., n, and n represents the number of inserted training symbols. This step ensures that the training data collected at different time positions can respectively reflect the channel characteristics at their corresponding moments.

[0035] Step 2: Calculate the weighting coefficient based on the demodulation SNR:

[0036] In order to make full use of the channel information carried by each training symbol, it is necessary to weight it according to its SNR level. Specifically, the SNR value SNR corresponding to each training symbol is estimated first. i , and then calculate its normalized weight coefficient w i ,

[0037]

[0038] Among them, w i is the normalized weight coefficient of the i-th training symbol, SNR i is the SNR value of the i-th training symbol, and n is the total number of training symbols.

[0039] This weighting strategy can effectively enhance the contribution of high-quality estimation values to the final channel estimation result and suppress the noise interference caused by low SNR.

[0040] Step 3: Maximum ratio combining:

[0041] Combine all weighted LS estimates to get the final channel estimation result H MRC , whose expression is:

[0042]

[0043] Among them, H MRC is the overall channel estimation result, H LS,i is the channel estimation value of the i-th training symbol, w i is the normalized weight coefficient of the i-th training symbol, and n is the total number of training symbols.

[0044] This maximum ratio combining strategy effectively improves the robustness and accuracy of estimation while preserving the multipath diversity gain.

[0045] The CR-MRC algorithm exploits the SNR diversity between multiple training symbols to adaptively weight and fuse multiple channel estimation results under varying signal-to-noise ratios. Compared to direct averaging, this algorithm significantly improves estimation accuracy and system anti-interference capabilities in low SNR environments, making it particularly suitable for OFDM systems operating in complex channel environments.

[0046] The technical solution described in this embodiment is applicable to a variety of communication systems, such as photon-assisted millimeter-wave terahertz wireless communication systems, all-solid-state circuit millimeter-wave terahertz wireless communication systems, polarization-multiplexed coherent transmission systems, multimode / multi-core optical fiber transmission systems, and the like, and has excellent universality. For example, in a photon-assisted millimeter-wave / terahertz wireless communication system, an OFDM electrical signal undergoes electrical-to-optical conversion in an I / Q modulator, beats with another optical signal in a photodetector (PD) to generate a millimeter-wave / terahertz signal, and transmits it through an antenna and a lens in free space. After the antenna at the receiving end receives the high-frequency electrical signal, it is down-converted to an intermediate frequency through a low-noise amplifier and a mixer. The high-frequency signal is then sampled by an oscilloscope and subjected to receiving-end digital signal processing, including synchronization, CR-MRC channel estimation, and back-end equalization, to restore the original transmitted symbols for evaluating system performance improvements.

[0047] As a preferred embodiment, the parameters of the CR-MRC algorithm used in this embodiment, such as the number and distribution of training symbols, the channel estimation method for each training symbol, the merging method of each channel estimation result, etc., can be specifically selected according to different systems, different channel characteristics, different transmission indicators and other specific conditions, and have high adjustability and applicability.

[0048] Both OFDM and DMT signals are applicable to the CR-MRC scheme proposed in this embodiment, and OFDM is used as an example to illustrate the scheme. Figure 2 As shown, the input bit stream is mapped into complex symbols. These symbols are then serial-to-parallel converted and processed using an M-point discrete Fourier transform spread spectrum (DFT-s) to reduce the peak-to-average power ratio, thereby improving power amplifier efficiency. Each OFDM frame contains T payload symbols, with S training symbols inserted (training symbols can be equally spaced or unequally spaced using various methods) for accurate channel estimation and equalization. Next, an N-point (N > M) inverse fast Fourier transform (IFFT) is performed, and a cyclic prefix is added to mitigate intersymbol interference and inter-subcarrier interference. After parallel-to-serial conversion, the OFDM signal is resampled to match the sampling rate of the arbitrary waveform generator. Finally, pilot tones are inserted at the edge of the signal spectrum to assist with frequency synchronization and signal recovery at the receiver.

[0049] In the digital signal processing flow at the receiver, the intermediate frequency (IF) signal is first down-converted to baseband through digital mixing, followed by frequency offset estimation and carrier phase recovery using pilot signals. Subsequent steps include resampling, frame synchronization, and cyclic prefix removal. Frequency-domain OFDM symbols are then recovered using an N-point fast Fourier transform (FFT). Embedded training symbols are extracted for LS channel response estimation, and these initial channel responses are further optimized using the CR-MRC algorithm proposed in this paper. To further suppress frequency-domain noise, intra-symbol frequency-domain averaging is employed to smooth the estimated results. The equalized symbols are then spread-spectrum processed using an M-point inverse discrete Fourier transform to restore the original symbols. These symbols are then input into a decision-directed least mean square equalizer with adaptive taps to compensate for any residual linear distortion. Finally, the system's end-to-end link performance is evaluated by calculating the bit error rate and signal-to-noise ratio.

[0050] Compared to the channel response estimated from a single training symbol and the arithmetic average of the channel responses, this embodiment fully utilizes the time diversity and SNR diversity inherent in the transmission of continuous training symbols, effectively suppressing the estimation noise and improving the system robustness without significantly increasing the computational overhead. In addition, this framework does not require the pre-acquisition of channel statistics, making it particularly suitable for dynamic and rapidly changing wireless communication environments. Therefore, the CR-MRC method proposed in this embodiment achieves a good balance between accuracy, computational efficiency, and practicality, providing a feasible approach for efficient channel estimation in next-generation OFDM communication systems.

[0051] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A digital signal processing method based on channel response-maximum ratio combining, characterized in that: include: Inserting a training symbol into each frame of a multi-carrier signal, performing channel estimation on each of the training signals, and obtaining a corresponding channel response estimation value; estimating an SNR value of each of the training symbols, and calculating a normalized weight coefficient based on the SNR value; A weighted average is performed based on the normalized weight coefficient of each training symbol and the channel response estimation value to obtain an overall channel estimation result of the multi-carrier signal.

2. The method according to claim 1, characterized in that The multi-carrier signal is an OFDM signal or a DMT signal.

3. The method according to claim 1, characterized in that The number of inserted training symbols is determined according to the number of data symbols of the multi-carrier signal.

4. The method according to claim 1, wherein The channel estimation adopts LS algorithm.

5. The method according to claim 1, characterized in that The calculation method of the normalized weight coefficient is as follows: Among them, w i is the normalized weight coefficient of the i-th training symbol, SNR i is the SNR value of the i-th training symbol, and n is the total number of training symbols.

6. The method according to claim 1, characterized in that The calculation method of the overall channel estimation result is as follows: Among them, H MRC is the overall channel estimation result, H LS,i is the channel estimation value of the i-th training symbol, w i is the normalized weight coefficient of the i-th training symbol, and n is the total number of training symbols.

7. An application of the method according to any one of claims 1 to 6 in a photon-assisted millimeter wave / terahertz wireless communication system, characterized in that: include: The method according to any one of claims 1 to 6 is used to perform receiving-end digital signal processing on the sampled data of the oscilloscope to restore the original transmitted symbols for evaluating the performance improvement of the system.

Images