Terahertz frequency band-oriented low-complexity wireless communication baseband system

Abstract

The invention relates to the technical field of wireless communication and digital signal processing, in particular to a terahertz frequency band-oriented low-complexity wireless communication baseband system. According to the technical scheme, the system comprises a transmitting end processing module, a wireless channel module and a receiving end processing module which are connected in sequence; the transmitting end processing module is used for carrying out coding, modulation, pilot frequency insertion, inverse Fourier transform and cyclic prefix addition processing on an original signal and generating a transmitting signal; the wireless channel module is used for simulating signal transmission characteristics of a terahertz frequency band; and the receiving end processing module is used for performing Fourier transform, cyclic prefix removal, channel estimation and interpolation, equalization, demodulation and decoding processing on the received signal so as to recover the original signal. According to the invention, hardware adaptation is realized through low-complexity design, the terahertz channel is adapted through anti-fading optimization, and stable, efficient and practical wireless communication under the terahertz frequency band is finally realized.

Description

Technical Field

[0001] This invention relates to the fields of wireless communication and digital signal processing technology, and in particular to a low-complexity wireless communication baseband system for the terahertz frequency band. Background Technology

[0002] As mobile communications gradually evolve towards sixth-generation (6G) and higher frequency bands, the terahertz band, with its advantages of large bandwidth, narrow beamwidth, and good directivity, has become an important candidate for the next generation of high-speed communication frequency bands. However, the characteristics of the terahertz band also bring a series of new challenges.

[0003] Terahertz signals are easily absorbed by atmospheric water molecules during propagation in space, resulting in severe attenuation. During transmission, these attenuate into multiple short-delay components with similar energies, leading to significant spatial crosstalk. Terahertz communication bandwidths can reach hundreds of GHz, with each OFDM symbol containing tens of thousands of subcarriers. Real-time processing in the digital baseband domain places excessive demands on hardware processing power. Furthermore, existing millimeter-wave technology does not adequately consider the unique sparse multipath and high-speed processing requirements of the terahertz band. Therefore, a baseband system solution is needed that addresses the characteristics of the terahertz band while maintaining high bit error rate performance and reducing hardware implementation complexity.

[0004] In summary, this application proposes a low-complexity wireless communication baseband system for the terahertz band. Summary of the Invention

[0005] The purpose of this invention is to address the problem in the background technology that terahertz communication bandwidth can reach hundreds of GHz, each OFDM symbol contains tens of thousands of subcarriers, and various digital baseband domain processing needs to be completed in real time, which places excessive demands on hardware processing capabilities. The invention proposes a low-complexity wireless communication baseband system for the terahertz frequency band.

[0006] The technical solution of the present invention: a low-complexity wireless communication baseband system for the terahertz band, comprising a transmitting end processing module, a wireless channel module and a receiving end processing module connected in sequence;

[0007] The transmitting end processing module is used to encode, modulate, insert pilots, perform inverse Fourier transform and cyclic prefix addition on the original signal and generate the transmitted signal.

[0008] The wireless channel module is used to simulate the signal transmission characteristics of the terahertz frequency band.

[0009] The receiving end processing module is used to perform Fourier transform, cyclic prefix removal, channel estimation and interpolation, equalization, demodulation and decoding on the received signal to recover the original signal.

[0010] Optionally, the transmitting end processing module includes a convolutional code encoding and interleaving module, a constellation mapping module, a first serial-to-parallel conversion module, a pilot insertion module, an IFFT operation module, a cyclic prefix addition module, and a first parallel-to-serial conversion module connected in sequence.

[0011] The convolutional code encoding and interleaving module is used to perform convolutional code encoding and block interleaving on the original bitstream;

[0012] The constellation mapping module is used to map the interleaved bitstream into complex modulation symbols;

[0013] The first serial-to-parallel conversion module is used to convert serial modulation symbols into a parallel data matrix;

[0014] The pilot insertion module is used to insert pilot columns into the parallel data matrix;

[0015] The IFFT operation module is used to perform IFFT operations on the data matrix with inserted pilots.

[0016] The cyclic prefix addition module is used to add cyclic prefixes to time-domain OFDM symbols;

[0017] The first parallel-to-serial conversion module is used to convert the parallel signal with the added cyclic prefix into a serial transmission signal.

[0018] Optionally, in the convolutional code encoding and interleaving module, the convolutional code encoder has a constraint length of 7 and an encoding efficiency of 1 / 2. The polynomials of the two output paths are as follows: First path: G1(X) = x 6 +x 5 +x 3 +x and the second path: G2(x)=x 6 +x 3 +x 2 +x 1 Where G1(X) and G2(X) are two output paths, x y This represents performing y delay operations; block interleaving uses a size of The matrix is ​​used to fill the encoded bitstream column by column and then output it row by row.

[0019] Optionally, the mapping process of the constellation mapping module includes:

[0020] The interleaved bitstream {b0,b1,…,b L-1 The result is a two-dimensional bit matrix obtained by grouping bits into groups of log2(M), where M is the number of constellation points, {b0, b1, ..., b...}. L-1} represents the output bit signal;

[0021] Treat each row of the two-dimensional bit matrix as a binary number and convert it into a decimal symbol sequence s. i {0,1,…,M-1}, where s i This represents the decimal symbol sequence number after conversion, with a data length from 0 to the constellation point number M minus one;

[0022] The decimal symbol sequence is mapped to complex modulation symbols using 16-QAM constellation and Gray code encoding.

[0023] Optionally, the pilot insertion module adopts a block-shaped pilot insertion structure, and the specific steps include:

[0024] (1) Construct a two-dimensional data matrix by performing serial-to-parallel conversion on the modulation symbols with column widths equal to the IFFT size:

[0025] D data =reshape(S i N iffi [L / N] iffi ]),

[0026] Among them, D data S represents the constructed two-dimensional data matrix; i Input modulation symbol sequence; N iffi represents the size of the inverse fast Fourier transform; L represents the length of the modulation symbol sequence; reshape represents the operation of converting a one-dimensional sequence into a two-dimensional matrix.

[0027] (2) Set the pilot spacing to P In The pilot insertion position is P. In ={1,1+P In ,1+2P In Each pilot column is a fixed complex number sequence:

[0028]

[0029] Among them, P In Represents the pilot sequence; C represents the complex field; N represents the pilot sequence. polit Indicates the number of pilot columns.

[0030] (3) Insert pilot columns into the original data matrix according to the pilot positions. Where N polit N represents the number of pilot columns; col P represents the number of columns in the original data matrix. In This indicates the pilot interval, and all data columns are shifted to the right to form a complete data frame matrix.

[0031] Optionally, the wireless channel module is based on a cascaded channel structure consisting of a multipath Rayleigh fading channel and Gaussian white noise superimposed on it, and the total impulse response of the channel is:

[0032]

[0033] Where L is the number of paths, α l (t) represents the time-correlated fading coefficient of the l-th path, and τ represents the delay of that path; h(t,τ) represents the total impulse response of the channel; α l (t) represents the time-correlated fading coefficient of the l-th path; δ represents the impulse characteristic; τ and η are the time variables and the delay of the l-th path, respectively, with each path delay ranging from tens to hundreds of nanoseconds, and the channel symbol sampling rate f. s =120MHz, with a maximum Doppler shift of 0.

[0034] Optionally, the receiving end processing module includes a second serial-to-parallel conversion module, a cyclic prefix removal module, an FFT operation module, a channel estimation and linear interpolation module, a channel equalization module, a second parallel-to-serial conversion module, a constellation demodulation module, and a channel decoding and deinterleaving module connected in sequence.

[0035] The second serial-to-parallel conversion module is used to convert the received serial signal into a parallel signal;

[0036] The cyclic prefix removal module is used to remove cyclic prefixes from parallel signals;

[0037] The FFT operation module is used to perform FFT operation on the signal after removing the cyclic prefix;

[0038] The channel estimation and linear interpolation module is used to perform channel estimation based on pilot signals and obtain the full-band channel response through linear interpolation;

[0039] The channel equalization module is used to perform frequency domain equalization based on the channel response;

[0040] The second parallel-to-serial conversion module is used to convert the equalized parallel signal into a serial symbol;

[0041] The constellation demodulation module is used to demodulate serial symbols into a bit stream;

[0042] The channel decoding and deinterleaving module is used to deinterleave and decode the convolutional codes of the demodulated bit stream.

[0043] Optionally, the processing procedure of the channel estimation and linear interpolation module includes:

[0044] Extract the received value at the pilot subcarrier from the frequency domain received signal matrix:

[0045] R p (k)=R k (m), m∈P,

[0046] Where P represents the set of pilot subcarrier indices, R p (k) is the received value extracted at the pilot subcarrier, which is the correlation quantity of the received signal corresponding to the pilot position extracted from the frequency domain received signal matrix; R k (m) is the received signal value at the k-th data block and m-th subcarrier position in the frequency domain received signal matrix.

[0047] Using the least squares method Estimate the channel response at the pilot position, where H(m) is the estimated channel response at pilot position m, reflecting the channel transmission characteristics at the pilot subcarrier; R p (m) is the received value at pilot subcarrier m; P(m) is the transmitted pilot symbol at pilot subcarrier m; N p It is the number of pilot subcarriers.

[0048] The channel response of the non-pilot subcarrier is estimated by piecewise linear interpolation, specifically as follows:

[0049]

[0050] Where H(m) is the channel response obtained by linear interpolation at the non-pilot subcarrier, and m is the index of the non-pilot subcarrier to be interpolated. Finally, the full-band channel response is obtained.

[0051] Optionally, the channel equalization module employs a zero-forcing equalization algorithm, dividing the received symbol of each data subcarrier by the corresponding channel estimate, i.e. Where X(m) is the data symbol after channel equalization, i.e., the original transmitted data recovered after channel equalization processing; R k H(m) is the received signal value at data subcarrier m; H(m) is the channel estimate at subcarrier m obtained by interpolation, reflecting the channel characteristics of the subcarrier.

[0052] Optionally, in the channel decoding and deinterleaving module, deinterleaving is the inverse operation of interleaving, using an R-row C-column matrix to write rows and then read columns; Viterbi decoding uses a convolutional code with a constraint length of 7 and a coding rate of 1 / 2, with the generator polynomial being G1(D) = 1 + D + D. 2 +D 3 +D 6 and G2(D)=1+D+D 2 +D 5 +D 6 G1(D) and G2(D) are two output paths, D x This represents performing x delay operations. A state metric is calculated on the received hard-decision two-bit symbol sequence. After reaching the backtracking depth, the system backtracks along the minimum metric path and outputs the original bit stream.

[0053] Compared with the prior art, this application includes at least one of the following beneficial technical effects:

[0054] A channel estimation method employing a block-shaped pilot structure and piecewise linear interpolation reduces pilot overhead and computational load, adapting to the large-scale subcarrier processing requirements of the terahertz band. A zero-forcing equalization algorithm is selected, achieving channel compensation through simple frequency domain division. Compared to complex equalization algorithms, this reduces real-time computational resource consumption and is easier to deploy on hardware platforms such as FPGAs. Each module utilizes mature, low-complexity algorithms (such as convolutional code encoding and 16-QAM mapping), and the modular design simplifies the integration process and reduces system implementation difficulty.

[0055] By combining convolutional codes and block interleaving, convolutional codes enhance error resilience, while block interleaving mitigates burst errors, specifically addressing the challenges of strong frequency selectivity and significant multipath fading in the terahertz band. OFDM technology, combined with a cyclic prefix, transforms the frequency-selective channel into a flat sub-channel. The cyclic prefix effectively resists inter-symbol interference caused by multipath interference, adapting to the complex terahertz fading environment. A cascaded channel model based on multipath Rayleigh fading and Gaussian white noise accurately simulates terahertz channel characteristics, ensuring the system design closely matches real-world transmission scenarios.

[0056] Supporting real-time processing speeds exceeding 5Gb / s on a single link via 2048-point IFFT / FFT, it adapts to the high-speed transmission demands brought by the large bandwidth of terahertz frequencies and can be applied to short-range high-speed communication scenarios. While reducing complexity, it ensures low bit error rate performance through Gray code mapping, pilot-assisted estimation, and equalization designs, achieving a balance between performance and resource consumption. Optimized parameters (such as path delay and sampling rate) for point-to-point scenarios such as indoor environments demonstrate strong scenario adaptability and high practical application value.

[0057] This invention achieves hardware adaptation through low-complexity design and adapts to terahertz channels through anti-fading optimization, ultimately realizing stable, efficient, and practical wireless communication in the terahertz band. Attached Figure Description

[0058] Figure 1 This is a block diagram of a low-complexity wireless communication baseband system for the terahertz band.

[0059] Figure 2 Performance simulation for this invention Figure 1 ;

[0060] Figure 3 Performance simulation for this invention Figure 2 (I need to be more specific.) Detailed Implementation

[0061] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0062] Example

[0063] like Figure 1 As shown, this invention proposes a low-complexity wireless communication baseband system for the terahertz band, comprising a transmitter processing module, a wireless channel module, and a receiver processing module connected in sequence. The transmitter processing module is used to encode, modulate, insert pilots, perform inverse Fourier transform, and add cyclic prefixes to the original signal to generate a transmitted signal. The wireless channel module is used to simulate the signal transmission characteristics of the terahertz band. The receiver processing module is used to perform Fourier transform, cyclic prefix removal, channel estimation and interpolation, equalization, demodulation, and decoding on the received signal to recover the original signal. The transmitter processing module, wireless channel module, and receiver processing module are described in detail below.

[0064] I. Sending end processing module

[0065] In this embodiment, the transmitting end processing module includes a convolutional code encoding and interleaving module, a constellation mapping module, a first serial-to-parallel conversion module, a pilot insertion module, an IFFT operation module, a cyclic prefix addition module, and a first parallel-to-serial conversion module connected in sequence.

[0066] The convolutional coding and interleaving module performs convolutional coding and block interleaving on the original bitstream. This module uses a convolutional encoder with a constraint length of 7 and a coding efficiency of 1 / 2 to encode the original bitstream signal. This coding method improves the channel's error resilience and is particularly suitable for terahertz channels with strong frequency selectivity. The encoder's two output path polynomials are as follows: G1(X) = x, where x represents the first path of the time delay operation. 6 +x 5 +x 3 +x Second path: G2(x)=x 6 +x 3 +x 2 +x 1 Where G1(X) and G2(X) are two output paths, x y This represents performing y delay operations.

[0067] In addition, the convolutional code encoding and interleaving module uses a block interleaving structure to mitigate burst errors that may occur in the convolutional code output sequence. The encoded bitstream has a length of L and uses 16-QAM modulation. The L bits are divided into the following sizes: The matrix is ​​first filled column-wise, then output row-wise. M is the QAM modulation order, which is 16 in this system. The interleaved bitstream is directly used for constellation mapping.

[0068] Furthermore, the constellation mapping module is used to map the interleaved bit stream into complex modulation symbols; the first serial-to-parallel conversion module is used to convert the serial modulation symbols into a parallel data matrix; further, the first serial-to-parallel conversion module and the constellation mapping module are: (1) to map the interleaved bit stream {b0,b1,…,b L-1}where b L The output bit signal is obtained by grouping the bits into log2(M) bits to form a two-dimensional bit matrix. (2) Each row of the two-dimensional bit matrix is ​​treated as a binary number and converted into a decimal symbol sequence s. i {0,1,…,M-1} where s i The decimal symbol sequence after conversion has a data length of 0 to the constellation point number M minus one. (3) The decimal symbol sequence is mapped to the constellation using Gray code 16-QAM to obtain the complex form of the modulation symbol.

[0069] To achieve efficient estimation and compensation of terahertz channels, a pilot insertion module is used to insert pilot columns into the parallel data matrix; a block-shaped pilot insertion structure is adopted. This structure is suitable for pilot design under frequency-selective fast fading channels and is beneficial for low-complexity equalization and interpolation in the frequency domain. (1) The modulation symbols {s1,s2,…,s L}where s L The modulated symbol sequence is converted from serial to parallel using IFFT with column width to construct a two-dimensional data matrix D. data =reshape(S i N iffi [L / N] iffi ]), where D data S represents the constructed two-dimensional data matrix; i Input modulation symbol sequence; N iffi (1) Indicates the magnitude of the inverse fast Fourier transform; L represents the length of the modulation symbol sequence; reshape represents the operation of converting a one-dimensional sequence into a two-dimensional matrix; (2) Set the pilot interval to P In Then the pilot insertion position is: P In ={1,1+P In ,1+2P In Each pilot column is a fixed complex number sequence: (3) Insert pilot columns into the original data matrix according to the pilot positions, and shift all data columns to the right to obtain the complete data frame matrix. The pilot columns are: Where N polit N represents the number of pilot columns; col P represents the number of columns in the original data matrix. In The pilot interval is indicated, and all data columns are shifted to the right to form a complete data frame matrix. Finally, the pilots are inserted into the data sequence using the method described above, forming a data frame matrix containing the pilots, which is then output for subsequent OFDM modulation.

[0070] It is worth noting that the IFFT operation module is used to perform IFFT operations on the data matrix with inserted pilots; the cyclic prefix addition module is used to add cyclic prefixes to the time-domain OFDM symbols; and the first parallel-to-serial conversion module is used to convert the parallel signal with added cyclic prefixes into a serial transmission signal. After performing IFFT operations on the data frame matrix with inserted pilots to obtain a time-domain complex sequence, a cyclic prefix is ​​introduced to improve the anti-multipath capability and adapt to the complex fading channel characteristics under the terahertz band. (1) For each column of frequency domain data D full Perform N ifft Point IFFT operation:

[0071]

[0072] Where d k [n] represents the nth sampling point in the time domain after the k-th OFDM symbol has undergone IFFT transformation; N ifft The number of IFFT operation points determines the length of the time-domain sequence and the number of subcarriers; m and n are the time-domain sampling point index and the frequency-domain subcarrier index, respectively. The resulting time-domain OFDM symbol is: D. ofdm =[d1,d2,..,d n ], where D ofdm To obtain the complete time-domain OFDM symbol sequence, d n (2) For each time-domain symbol d, the n sampling points are represented; k Copy the last N from its front end cp Given a set of symbols, we obtain the symbols after inserting the cyclic prefix: in This indicates that the length of the complete symbol sequence after adding a cyclic prefix to the k-th OFDM symbol is N. ifft +N cp N cp The cyclic prefix length represents the number of sampling points copied from the end of the time-domain symbol to the front end. (3) Finally, the two-dimensional OFDM symbol is converted into the baseband transmit signal output by the transmitter through parallel-to-serial conversion. S txThis indicates that the final baseband transmit signal output by the transmitter is the result of serial-to-parallel conversion of multiple OFDM symbols, which is then used for channel transmission; N sym This represents the total number of symbols used in the serial-to-parallel conversion.

[0073] In this embodiment, convolutional codes with a constraint length of 7 and a coding efficiency of 1 / 2 effectively improve error resistance. Combined with a block interleaving structure, burst errors can be dispersed, specifically addressing the channel characteristics of the terahertz band, which features strong frequency selectivity and significant multipath fading, thus reducing signal distortion caused by channel fluctuations. Adding a cyclic prefix effectively resists inter-symbol interference caused by terahertz multipath propagation; and OFDM-based multicarrier modulation technology transforms the frequency-selective channel into a flat fading sub-channel, further reducing the impact of channel fading on signal transmission.

[0074] II. Wireless Channel Module

[0075] In this embodiment, the wireless channel module is based on a cascaded channel structure consisting of a multipath Rayleigh fading channel and Gaussian white noise superimposed on it, and the total impulse response of the channel is:

[0076]

[0077] Where L is the number of paths, α l (t) represents the time-correlated fading coefficient of the l-th path, and τ represents the delay of that path; h(t,τ) represents the total impulse response of the channel; α l (t) represents the time-correlated fading coefficient of the l-th path; δ represents the impulse characteristic; τ and η are the time variables and the delay of the l-th path, respectively, with each path delay ranging from tens to hundreds of nanoseconds, and the channel symbol sampling rate f. s =120MHz, with a maximum Doppler shift of 0.

[0078] In the wireless channel module of this embodiment, the system includes a modeling module for terahertz point-to-point channel simulation. The modeling module is based on a cascaded channel structure composed of a multipath Rayleigh fading channel and Gaussian white noise superposition, used to simulate terahertz band communication under point-to-point communication. Specifically, it includes:

[0079] (1) This system uses a Gaussian white noise channel combined with a Rayleigh multipath channel to simulate the real channel performance. The Gaussian white noise channel is used to simulate the thermal noise and quantization noise at both ends of the receiver when the system is operating in the terahertz band. The Rayleigh multipath channel is used to simulate the multipath fading propagation characteristics in the terahertz band.

[0080] (2) Let the total impulse response of the channel be: Where L is the number of paths α l(t) represents the time-correlated fading coefficient of the l-th path, and τ represents the path delay. The path delay τ is set with reference to the delay spread characteristics of the terahertz band in a typical indoor scenario, ranging from tens to hundreds of nanoseconds. In this example, τ = [0, 50, 110, 170] ns. Considering that the actual use is approximately a static scenario, the maximum Doppler shift is 0, and the channel symbol key sampling rate f... s =120MHz meets the high-speed adoption requirements of terahertz systems.

[0081] (3) The simulated terahertz channel is synthesized as y(t)=h(t,τ)*x(t)+n(t), where x(t) is the transmitted signal, n(t) is zero-mean complex Gaussian white noise, and h(t,τ) is the channel response. A cascaded channel model of multipath Rayleigh fading and Gaussian white noise is adopted to accurately simulate the multipath fading characteristics and noise environment of the terahertz band, making the system design more in line with the actual application scenario and ensuring the stability of communication.

[0082] III. Receiver Processing Module

[0083] In this embodiment, the receiver processing module includes a second serial-to-parallel conversion module, a cyclic prefix removal module, an FFT operation module, a channel estimation and linear interpolation module, a channel equalization module, a second parallel-to-serial conversion module, a constellation demodulation module, and a channel decoding and deinterleaving module connected in sequence.

[0084] The second serial-to-parallel conversion module is used to convert the received serial signal into a parallel signal, the cyclic prefix removal module is used to remove the cyclic prefix from the parallel signal, and the FFT operation module is used to perform FFT operation on the signal after removing the cyclic prefix; after passing through the terahertz channel, the signal is down-converted into a baseband signal and received by the baseband signal processing system at the receiving end. It includes the second serial-to-parallel conversion module, the cyclic prefix removal module, and the FFT operation module. (1) The received serial signal is reassembled into a two-dimensional array according to the OFDM symbol length. (2) The N prefix at the beginning of each OFDM symbol is removed. cp Each symbol retains only the effective subcarriers (3). A fast discrete Fourier transform is performed on each OFDM symbol to recover the complex subcarrier information in the frequency domain, i.e.: R k This represents the frequency domain complex subcarrier information recovered at the m-th subcarrier position after the k-th OFDM symbol undergoes FFT operation. k (n) represents the received signal value at the nth sampling point in the time domain after removing the cyclic prefix of the k-th OFDM symbol.

[0085] The channel estimation and linear interpolation module is used to perform channel estimation based on pilots and obtain the full-band channel response through linear interpolation; the channel equalization module is used to perform frequency domain equalization based on the channel response; the baseband system receiver signal enters the channel estimation and equalization module after passing through the signal reshaping module, which is used to perform pilot-assisted channel response estimation on the received frequency domain signal and complete the frequency domain equalization recovery of the received signal based on the estimation result. (1) Pilot extraction: extract the received value at the pilot subcarrier from the frequency domain received signal matrix: R p (k)=R k (m), m∈P, where P represents the set of pilot subcarrier indices, R p (k) is the received value extracted at the pilot subcarrier, which is the correlation quantity of the received signal corresponding to the pilot position extracted from the frequency domain received signal matrix; R k (m) is the received signal value at the k-th data block and m-th subcarrier position in the frequency domain received signal matrix. (2) The channel response is estimated based on the extracted pilot signals using the least squares method: Where H(m) is the estimated channel response at pilot position m, reflecting the channel transmission characteristics at the pilot subcarrier; R p (m) is the received value at pilot subcarrier m; P(m) is the transmitted pilot symbol at pilot subcarrier m; N p This is the number of pilot subcarriers. (3) Since the pilot subcarriers only occupy sparse points in the frequency band, we use piecewise linear interpolation to estimate the channel response of the entire subcarrier. The specific steps are as follows: (4) The zero-forcing channel equalization of each data subcarrier is performed using the channel response obtained by linear interpolation. The result is: Where X(m) is the data symbol after channel equalization, that is, the original transmitted data recovered after channel equalization processing; R k H(m) is the received signal value at data subcarrier m; H(m) is the channel estimate at subcarrier m obtained by interpolation, reflecting the channel characteristics of the subcarrier.

[0086] The second parallel-to-serial conversion module converts the equalized parallel signal into a serial symbol; the constellation demodulation module demodulates the serial symbol into a bit stream; the equalized data passes through the second parallel-to-serial conversion module and the constellation demodulation module, specifically including:

[0087] (1) For the equalized frequency domain data matrix D = [D m,n Expanding all m and n sequentially yields a result of length N. sc N sym vector Where r represents the serial symbol vector obtained by expanding the frequency domain data matrix after equalization. If r contains a tail virtual carrier due to zero padding, then the first L valid symbols r are truncated.i =[r0,r1,...,r L-1 ], r i This represents the vector of the first L valid serial symbols retained after removing the tail virtual carrier.

[0088] (2) For each received symbol r i In the 16QAM constellation point set S = {s0, s1, ..., s} M-1 Select the nearest neighbor in} The final output bit sequence is concatenated in sign order as follows: Where k i Represents the i-th received symbol r i The corresponding decimal index is calculated after finding the nearest zero point in the constellation point set; b i , j represents the value of the i-th symbol and the j-th bit; M represents the number of constellation points in the constellation diagram corresponding to 16QAM, and in this embodiment M = 16.

[0089] It should be noted that the channel decoding and deinterleaving module is used to deinterleave and decode the convolutional code of the demodulated bit stream. The demodulated data passes through the deinterleaving and channel decoding module, which includes: (1) a bit deinterleaving module that performs the reverse operation of the transmitting end interleaving module, where the R-row C-column matrix is ​​written row by row and then read column by column. Its deinterleaving operation satisfies: b' m,n =b n+mC ,m=0,...,R-1,n=0,...,C-1,where b` m，n This represents the bit value located in the m-th row and n-th column after deinterleaving. (2) Viterbi decoding uses a convolutional code with a constraint length W = 7 and a coding rate of 1 / 2, generating the polynomial: G1(D) = 1 + D + D 2 +D 3 +D 6 G2(D) = 1 + D + D 2 +D 5 +D 6 G1(D) and G2(D) are two output paths, D x This represents performing x delay operations; for the received hard-decision two-bit symbol sequence {r i The metric for state s at step i: Where M s (i) represents the state s at step i. i The corresponding path metric reflects the cumulative distance of the path to that state; r i Let be the i-th received symbol; c(s`, s) represents the symbol output by the convolutional encoder during the two state transitions. When i reaches the backtracking depth L... tb Then select the value with the smallest metric, and backtrack L from the minimum value along the path of the minimum metric. tbThe process is repeated step by step, and the corresponding final processed bit sequence is output.

[0090] This invention provides a low-complexity wireless communication baseband system and its implementation method for the terahertz band. It combines classic OFDM modulation with low-complexity algorithm design, significantly improving the system's feasibility and reliability in high-speed transmission scenarios. The system incorporates a channel estimation method based on block pilot structures and linear interpolation, a simplified frequency domain equalization algorithm, and resource-friendly convolutional coding and Viterbi decoding techniques, such as... Figure 2-3 The diagram illustrates low-complexity resource consumption and stable transmission in a simulated terahertz channel. Figure 2-3 It can be seen that this baseband system achieves 10 when the signal-to-noise ratio is less than 25 dB. -5 With a low bit error rate and no frequency offset or constellation point disorder in the final receiver constellation diagram, this system has excellent communication error handling capabilities, can cope with the high fading channel communication characteristics of the terahertz band, and its low system complexity makes it more suitable for subsequent FPGA implementation.

[0091] The method employs a block-shaped pilot insertion structure combined with a piecewise linear interpolation channel estimation approach, avoiding complex adaptive pilot design and dense full-frequency-domain pilot deployment. This reduces pilot overhead and computational load while maintaining estimation accuracy, making it particularly suitable for the large-scale subcarrier processing demands arising from high-spectrum resources in the terahertz band. By using a zero-forcing equalization algorithm, channel compensation is achieved through simple frequency-domain division, significantly reducing real-time computational resource consumption compared to complex equalization algorithms such as Minimum Mean Square Error (MMSE), and making it easier to implement on hardware platforms such as FPGAs. Each module has a clearly defined function and simple interfaces. For example, convolutional code encoding and 16-QAM constellation mapping utilize mature and low-complexity classic algorithms, avoiding overly complex signal processing flows and reducing system integration difficulty.

[0092] This invention utilizes a 2048-point IFFT / FFT operation module to support real-time baseband processing speeds exceeding 5Gb / s on a single link, adapting to the high-speed transmission demands of the large bandwidth in the terahertz band. It provides a feasible solution for future short-range high-speed communication (such as data transmission between indoor devices, and millimeter-wave and terahertz integrated communication). While reducing complexity, it ensures the system's bit error rate performance through Gray code-encoded 16-QAM constellation mapping, pilot-assisted channel estimation, and equalization. Simulation results demonstrate stable transmission in terahertz channels, balancing communication quality and resource consumption. Specifically designed for terahertz point-to-point short-range communication scenarios, the path delay parameters (tens to hundreds of nanoseconds) and sampling rate (120MHz) are set with reference to typical indoor environments, making it directly applicable to practical scenarios such as indoor high-speed data transmission and device interconnection, demonstrating high practical value.

[0093] The above specific embodiments are merely several optional embodiments of the present invention. Based on the technical solutions of the present invention and the relevant teachings of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

Claims

1. A low-complexity wireless communication baseband system for the terahertz frequency band, characterized in that, It includes a transmitting end processing module, a wireless channel module, and a receiving end processing module connected in sequence; The transmitting end processing module is used to encode, modulate, insert pilots, perform inverse Fourier transform and cyclic prefix addition on the original signal and generate the transmitted signal. The wireless channel module is used to simulate the signal transmission characteristics of the terahertz frequency band. The receiving end processing module is used to perform Fourier transform, cyclic prefix removal, channel estimation and interpolation, equalization, demodulation and decoding on the received signal to recover the original signal.

2. The low-complexity wireless communication baseband system for the terahertz band according to claim 1, characterized in that, The transmitting end processing module includes a convolutional code encoding and interleaving module, a constellation mapping module, a first serial-to-parallel conversion module, a pilot insertion module, an IFFT operation module, a cyclic prefix addition module, and a first parallel-to-serial conversion module connected in sequence. The convolutional code encoding and interleaving module is used to perform convolutional code encoding and block interleaving on the original bitstream; The constellation mapping module is used to map the interleaved bitstream into complex modulation symbols; The first serial-to-parallel conversion module is used to convert serial modulation symbols into a parallel data matrix; The pilot insertion module is used to insert pilot columns into the parallel data matrix; The IFFT operation module is used to perform IFFT operations on the data matrix with inserted pilots. The cyclic prefix addition module is used to add cyclic prefixes to time-domain OFDM symbols; The first parallel-to-serial conversion module is used to convert the parallel signal with the added cyclic prefix into a serial transmission signal.

3. A low-complexity wireless communication baseband system for the terahertz band according to claim 2, characterized in that, In the convolutional code encoding and interleaving module, the convolutional code encoder has a constraint length of 7 and an encoding efficiency of 1 / 2. The polynomials of the two output paths are as follows: First path: G1(X) = x 6 +x 5 +x 3 +x and the second path: G2(x)=x 6 +x 3 +x 2 +x 1 Where G1(X) and G2(X) are two output paths, x y This represents performing y delay operations; Block interlacing uses a size of The matrix is ​​used to fill the encoded bitstream column by column and then output it row by row.

4. A low-complexity wireless communication baseband system for the terahertz band according to claim 2, characterized in that, The mapping process of the constellation mapping module includes: The interleaved bitstream {b0,b1,…,b L-1 The result is a two-dimensional bit matrix obtained by grouping bits into groups of log2(M), where M is the number of constellation points, {b0, b1, ..., b...}. L-1 } represents the output bit signal; Treat each row of the two-dimensional bit matrix as a binary number and convert it into a decimal symbol sequence s. i {0,1,…,M-1}, where s i This represents the decimal symbol sequence number after conversion, with a data length from 0 to the constellation point number M minus one; The decimal symbol sequence is mapped to complex modulation symbols using 16-QAM constellation and Gray code encoding.

5. A low-complexity wireless communication baseband system for the terahertz band according to claim 2, characterized in that, The pilot insertion module adopts a block-shaped pilot insertion structure, and the specific steps include: (1) Construct a two-dimensional data matrix by performing serial-to-parallel conversion on the modulation symbols with column widths equal to the IFFT size: D data =reshape(S i ,N iffi ,[L / N iffi ]), Among them, D data S represents the constructed two-dimensional data matrix; i Input modulation symbol sequence; N iffi represents the size of the inverse fast Fourier transform; L represents the length of the modulation symbol sequence; reshape represents the operation of converting a one-dimensional sequence into a two-dimensional matrix. (2) Set the pilot interval to PIn and the pilot insertion position to P. In ={1,1+P In ,1+2P In Each pilot column is a fixed complex number sequence: Among them, P In Represents the pilot sequence; C represents the complex field; N represents the pilot sequence. polit Indicates the number of pilot columns. (3) Insert pilot columns into the original data matrix according to the pilot positions. Where N polit N represents the number of pilot columns; col P represents the number of columns in the original data matrix. In This indicates the pilot interval, and all data columns are shifted to the right to form a complete data frame matrix.

6. A low-complexity wireless communication baseband system for the terahertz band according to claim 1, characterized in that, The wireless channel module is based on a cascaded channel structure consisting of a multipath Rayleigh fading channel and Gaussian white noise superimposed on it. The total impulse response of the channel is: Where L is the number of paths, α l (t) represents the time-correlated fading coefficient of the l-th path, and τ represents the delay of that path; h(t,τ) represents the total impulse response of the channel; α l (t) represents the time-correlated fading coefficient of the l-th path; δ represents the impulse characteristic; τ and η are the time variables and the delay of the l-th path, respectively, with each path delay ranging from tens to hundreds of nanoseconds, and the channel symbol sampling rate f. s =120MHz, with a maximum Doppler shift of 0.

7. A low-complexity wireless communication baseband system for the terahertz band according to claim 1, characterized in that, The receiving end processing module includes a second serial-to-parallel conversion module, a cyclic prefix removal module, an FFT operation module, a channel estimation and linear interpolation module, a channel equalization module, a second parallel-to-serial conversion module, a constellation demodulation module, and a channel decoding and deinterleaving module connected in sequence. The second serial-to-parallel conversion module is used to convert the received serial signal into a parallel signal; The cyclic prefix removal module is used to remove cyclic prefixes from parallel signals; The FFT operation module is used to perform FFT operation on the signal after removing the cyclic prefix; The channel estimation and linear interpolation module is used to perform channel estimation based on pilot signals and obtain the full-band channel response through linear interpolation; The channel equalization module is used to perform frequency domain equalization based on the channel response; The second parallel-to-serial conversion module is used to convert the equalized parallel signal into a serial symbol; The constellation demodulation module is used to demodulate serial symbols into a bit stream; The channel decoding and deinterleaving module is used to deinterleave and decode the convolutional codes of the demodulated bit stream.

8. A low-complexity wireless communication baseband system for the terahertz band according to claim 7, characterized in that, The processing steps of the channel estimation and linear interpolation module include: Extract the received value at the pilot subcarrier from the frequency domain received signal matrix: R p (k)=R k (m),m∈P, Where P represents the set of pilot subcarrier indices, R p (k) is the received value extracted at the pilot subcarrier, which is the correlation quantity of the received signal corresponding to the pilot position extracted from the frequency domain received signal matrix; R k (m) is the received signal value at the k-th data block and the m-th subcarrier position in the frequency domain received signal matrix. Using the least squares method i = 1, 2, ..., N p Estimate the channel response at the pilot position, where H(m) is the estimated channel response at pilot position m, reflecting the channel transmission characteristics at the pilot subcarrier; R p P(m) is the received value at pilot subcarrier m; P(m) is the transmitted pilot symbol at pilot subcarrier m; N p It is the number of pilot subcarriers. The channel response of the non-pilot subcarrier is estimated by piecewise linear interpolation, specifically as follows: Where H(m) is the channel response obtained by linear interpolation at the non-pilot subcarrier, and m is the index of the non-pilot subcarrier to be interpolated. Finally, the full-band channel response is obtained.

9. A low-complexity wireless communication baseband system for the terahertz band according to claim 8, characterized in that, The channel equalization module employs a zero-forcing equalization algorithm, dividing the received symbol of each data subcarrier by the corresponding channel estimate, i.e. m∈D, where X(m) is the data symbol after channel equalization, i.e., the original transmitted data recovered after channel equalization processing; R k H(m) is the received signal value at data subcarrier m; H(m) is the channel estimate at subcarrier m obtained by interpolation, reflecting the channel characteristics of the subcarrier.

10. A low-complexity wireless communication baseband system for the terahertz band according to claim 8, characterized in that, In the channel decoding and deinterleaving module, deinterleaving is the inverse operation of interleaving, using an R-row C-column matrix to write rows and then read columns; Viterbi decoding uses a convolutional code with a constraint length of 7 and a coding rate of 1 / 2, with the generator polynomial being G1(D) = 1 + D + D. 2 +D 3 +D 6 and G2(D)=1+D+D 2 +D 5 +D 6 G1(D) and G2(D) are two output paths, D x This represents performing x delay operations. A state metric is calculated on the received hard-decision two-bit symbol sequence. After reaching the backtracking depth, the system backtracks along the minimum metric path and outputs the original bit stream.

Images