Signal detection method, device, electronic device, and computer-readable storage medium

Abstract

The present disclosure provides a signal detection method, device, electronic equipment and computer readable storage medium. The signal detection method comprises: obtaining a target received signal; obtaining channel information corresponding to the target received signal; performing linear equalization processing on the target received signal and the target received signal through the channel information to obtain approximate channel information and an estimated value of a target transmitted signal corresponding to the target received signal; a residual calculation unit of a signal detection model processes the target received signal, the approximate channel information and the estimated value of the target transmitted signal to determine a signal residual, the signal detection model further comprising a signal detection unit; and the signal detection unit processes the estimated value of the target transmitted signal and the signal residual to complete signal detection on the target received signal and obtain a target signal. The technical solution provided by the embodiments of the present disclosure can simply and accurately complete the detection of the target received signal.

Description

Technical Field

[0001] The present disclosure relates to the field of computer and Internet technology, and in particular to a signal detection method and device, an electronic device, and a computer-readable storage medium. Background Art

[0002] OTFS (Orthogonal Time Frequency Space) is a new type of orthogonal time-frequency modulation technology, which is particularly suitable for systems with high Doppler, short packets and large antenna arrays.

[0003] OTFS still faces some challenges in practical applications. In OTFS systems, it is difficult to achieve ideal transmit and receive filters that satisfy time-frequency biorthogonality, so rectangular filters are usually used instead in practice. At this time, in high-speed mobile scenarios, OTFS signals will also be subject to inter-carrier interference and inter-symbol interference. In addition, because in OTFS systems, an OTFS symbol block is equivalent to several consecutive multi-carrier symbols, when the received signal is represented as the product of the transmitted signal and the channel matrix, the size of the channel matrix is ​​usually very large. Using traditional linear detection methods for interference elimination involves matrix inversion, which has high time and space complexity. When using nonlinear methods for detection, the complexity cannot be reduced, and the key parameters are usually manually set based on simulation results. Whether these parameters are optimal is worth discussing.

[0004] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure.

[0005] Public content

[0006] The purpose of the present disclosure is to provide a signal detection method, device, electronic device and computer-readable storage medium, which can simply and accurately complete the detection of the target received signal.

[0007] Other features and advantages of the present disclosure will become apparent from the following detailed description, or may be learned in part by practice of the present disclosure.

[0008] An embodiment of the present disclosure provides a signal detection method, including: obtaining a target received signal; obtaining channel information corresponding to the target received signal; performing linear equalization processing on the target received signal and the target received signal through the channel information to obtain approximate channel information and an estimated value of a target transmitted signal, wherein the target transmitted signal corresponds to the target received signal; a residual calculation unit of a signal detection model processes the target received signal, the approximate channel information and the estimated value of the target transmitted signal to determine a signal residual, and the signal detection model also includes a signal detection unit; the estimated value of the target transmitted signal and the signal residual are processed by the signal detection unit to complete signal detection for the target received signal and obtain the target signal.

[0009] In some embodiments, the target received signal is a signal in the time domain; wherein, the target received signal and the target received signal are linearly equalized through the channel information to obtain approximate channel information and an estimated value of the target transmitted signal, including: obtaining the Doppler domain input-output relationship of the target received signal, the channel information and the target transmitted signal in the delayed Doppler domain, the Doppler domain input-output relationship including a delay dimension parameter; replacing the delay dimension parameter in the Doppler domain input-output relationship with the mean of the delay dimension parameter to convert the Doppler domain input-output relationship into a Doppler domain approximate input-output relationship; determining the approximate channel information from the Doppler domain approximate input-output relationship; and determining the estimated value of the target transmitted signal through the Doppler domain approximate input-output relationship, the approximate channel information and the target received signal.

[0010] In some embodiments, obtaining the Doppler domain input-output relationship of the target received signal, the channel information, and the target transmitted signal in the delayed Doppler domain includes: determining the time domain input-output relationship of the target received signal, the channel information, and the target transmitted signal in the time domain; transforming the time domain input-output relationship through inverse Fourier transform and inverse sigmoid Fourier transform to determine the Doppler domain input-output relationship of the target received signal, the channel information, and the target transmitted signal in the delayed Doppler domain.

[0011] In some embodiments, the residual calculation unit of the signal detection model processes the target received signal, the approximate channel information, and the estimated value of the target transmitted signal to determine a signal residual, including: performing decision processing on the estimated value of the target transmitted signal to determine a decision estimate; processing the target received signal through the approximate channel information to determine a first transmitted signal; processing the decision estimate through the approximate channel information to determine a second transmitted signal; and determining the signal residual based on the first transmitted signal and the second transmitted signal.

[0012] In some embodiments, the estimated value of the target transmitted signal includes an estimated real part and an estimated imaginary part; wherein, the estimated value of the target transmitted signal is subjected to decision processing to determine a decision estimated value, including: performing decision processing on the estimated real part and the estimated imaginary part respectively to obtain a decision real part and a decision imaginary step; and determining the decision estimated value based on the decision real part and the decision imaginary step.

[0013] In some embodiments, the signal residual includes a real part of the residual and an imaginary part of the residual, the real part of the residual includes an absolute value residual real part and a positive and negative sign residual real part, and the imaginary part of the residual includes an absolute value residual imaginary part and a positive and negative sign residual imaginary part; wherein the estimated value of the target transmitted signal and the signal residual are processed by the signal detection unit to complete the signal detection for the target received signal and obtain the target signal, including: performing feature extraction processing on the absolute value residual real part by the signal detection unit to obtain an absolute value residual real part estimate; performing feature extraction processing on the absolute value residual real part by the signal detection unit to obtain an absolute value residual real part estimate The imaginary part of the residual is subjected to feature extraction processing to obtain an absolute value residual imaginary part estimate; the absolute value residual real part estimate and the absolute value residual imaginary part estimate are activated to obtain a real part activation absolute estimate and an imaginary part activation absolute estimate; the real part of the target signal is determined by the real part of the residual, the positive and negative sign residual real part and the real part activation absolute estimate; the imaginary part of the target signal is determined by the imaginary part of the residual, the positive and negative sign residual imaginary part and the imaginary part activation absolute estimate; the target signal is determined according to the real part of the target signal and the imaginary part of the target signal.

[0014] In some embodiments, the signal detection unit performs feature extraction processing on the absolute value residual real part to obtain an absolute value residual real part estimate, including: performing mean processing on the absolute value residual real part along the Doppler dimension to obtain the absolute value Doppler dimension mean real part; performing mean processing on the absolute value residual real part along the delay dimension to obtain the absolute value delay dimension mean real part; determining the absolute value residual real part estimate through the absolute value Doppler dimension mean real part, the absolute value delay dimension mean real part and the absolute value residual real part. wherein, the signal detection unit performs feature extraction processing on the absolute value residual imaginary part to obtain an estimated value of the absolute value residual imaginary part, including: performing mean processing on the absolute value residual imaginary part along Doppler to obtain an absolute value Doppler dimension mean imaginary part; performing mean processing on the absolute value residual imaginary part along the delay dimension to obtain an absolute value delay dimension mean imaginary part; and determining the estimated value of the absolute value residual imaginary part through the absolute value Doppler dimension mean imaginary part, the absolute value delay dimension mean imaginary part and the absolute value residual imaginary part.

[0015] An embodiment of the present disclosure provides a signal detection device, including: a signal receiving module, a channel information acquisition module, a linear equalization processing module, a residual calculation module, and a signal detection module.

[0016] Among them, the signal receiving module is used to obtain the target received signal; the channel information acquisition module can be used to obtain the channel information corresponding to the target received signal; the linear equalization processing module can be used to perform linear equalization processing on the target received signal and the target received signal through the channel information to obtain approximate channel information and an estimated value of the target transmitted signal, and the target transmitted signal corresponds to the target received signal; the residual calculation module can be used for the residual calculation unit of the signal detection model to process the target received signal, the approximate channel information and the estimated value of the target transmitted signal to determine a signal residual, and the signal detection model also includes a signal detection unit; the signal detection module can be used to process the estimated value of the target transmitted signal and the signal residual through the signal detection unit to complete signal detection for the target received signal and obtain the target signal.

[0017] An embodiment of the present disclosure proposes an electronic device, which includes: one or more processors; a storage device for storing one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement any of the signal detection methods described above.

[0018] An embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the signal detection method as described in any one of the above items is implemented.

[0019] The present disclosure provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the above-described signal detection method.

[0020] The signal detection method, device, electronic device and computer-readable storage medium provided by the embodiments of the present disclosure, on the one hand, perform signal detection on the target received signal through a signal detection model, which is simple and accurate; on the other hand, when using the signal detection model to perform signal detection, the signal residual is taken into consideration, thereby reducing the influence of the signal residual peak position on the signal detection and improving the accuracy of signal detection.

[0021] It should be understood that the foregoing general description and the following detailed description are exemplary only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification, are used to explain the principles of the present disclosure. Obviously, the drawings described below are only some embodiments of the present disclosure, and those skilled in the art can derive other drawings based on these drawings without inventive effort.

[0023] Figure 1 A schematic diagram showing an exemplary system architecture that can be applied to the signal detection method or signal detection device of the embodiments of the present disclosure is shown.

[0024] Figure 2 The figure is a flow chart showing a signal detection method according to an exemplary embodiment.

[0025] Figure 3 The figure shows a signal modulation and demodulation block diagram according to an exemplary embodiment.

[0026] Figure 4 The figure is a signal detection block diagram according to an exemplary embodiment.

[0027] Figure 5 The figure is a flow chart showing a method for linearly equalizing a target received signal according to an exemplary embodiment.

[0028] Figure 6 The figure is a flowchart of a method for determining a signal residual according to an exemplary embodiment.

[0029] Figure 7is a flow chart of a target signal detection method according to an example embodiment.

[0030] Figure 8 is a signal residual distribution diagram according to an example embodiment.

[0031] Figure 9 is a signal residual distribution diagram according to an example embodiment.

[0032] Figure 10 is a network structure diagram of a signal detection model according to an example embodiment.

[0033] Figure 11 is a flow chart of a signal detection method according to an example embodiment.

[0034] Figure 12 is a multiple signal detection method bit error rate comparison diagram according to an example embodiment.

[0035] Figure 13 is a block diagram of a signal detection device according to an example embodiment.

[0036] Figure 14 shows a structural schematic diagram of an electronic device suitable for implementing the embodiments of the present disclosure. DETAILED DESCRIPTION

[0037] Example embodiments now will be described more fully hereinafter with reference to the accompanying drawings. Example embodiments, however, can be implemented in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of example embodiments to those skilled in the art. Like reference numerals refer to like elements throughout the figures, and thus description of the same will be omitted.

[0038] The features, structures, or characteristics described in connection with the present disclosure can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the present disclosure. One skilled in the relevant art will recognize, however, that the techniques described herein can be practiced without one or more of the specific details, or with other methods, components, materials, and so forth. In other instances, well-known structures, devices, implementations, or operations are not shown or described in detail to avoid obscuring aspects of the present disclosure.

[0039] The accompanying drawings are merely schematic illustrations of the present disclosure. Identical reference numerals in the drawings denote identical or similar components, and thus their repeated descriptions will be omitted. Some of the block diagrams shown in the accompanying drawings do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0040] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all content and steps, nor must they be executed in the order described. For example, some steps may be decomposed, while others may be combined or partially combined. Therefore, the actual execution order may vary depending on the actual situation.

[0041] In this specification, the terms "a", "an", "the", "said" and "at least one" are used to indicate the presence of one or more elements / components / etc.; the terms "comprising", "including" and "having" are used to express open-ended inclusion and mean that additional elements / components / etc. may exist in addition to the listed elements / components / etc.; the terms "first", "second" and "third" etc. are used only as labels and are not intended to limit the quantity of their objects.

[0042] Hereinafter, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings.

[0043] Figure 1 A schematic diagram showing an exemplary system architecture that can be applied to the signal detection method or signal detection device of the embodiments of the present disclosure is shown.

[0044] like Figure 1 As shown, system architecture 100 may include terminal devices 101, 102, 103, a network 104, and a server 105. Network 104 is a medium for providing communication links between terminal devices 101, 102, 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables.

[0045] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Terminal devices 101, 102, and 103 can be various electronic devices with display screens and support web browsing, including but not limited to smartphones, tablet computers, laptop computers, desktop computers, wearable devices, virtual reality devices, smart homes, etc.

[0046] The server 105 may be a server that provides various services, such as a background management server that provides support for devices operated by users using the terminal devices 101, 102, and 103. The background management server may analyze and process received requests and other data, and feed back the processing results to the terminal device.

[0047] The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), as well as basic cloud computing services such as big data and artificial intelligence platforms, etc. This disclosure does not impose any restrictions on this.

[0048] The server 105 may, for example, obtain a target received signal; the server 105 may, for example, obtain channel information corresponding to the target received signal; the server 105 may, for example, perform linear equalization processing on the target received signal and the target received signal through the channel information to obtain approximate channel information and an estimated value of a target transmitted signal, and the target transmitted signal corresponds to the target received signal; the server 105 may, for example, process the target received signal, the approximate channel information and the estimated value of the target transmitted signal through a residual calculation unit of a signal detection model to determine a signal residual, and the signal detection model also includes a signal detection unit; the server 105 may, for example, process the estimated value of the target transmitted signal and the signal residual through the signal detection unit to complete signal detection for the target received signal and obtain the target signal.

[0049] It should be understood that Figure 1 The number of terminal devices, networks and servers is merely illustrative. The server 105 may be a single entity server or may be composed of multiple servers. It may have any number of terminal devices, networks and servers according to actual needs.

[0050] During high-speed mobility, wireless signals reflect off obstacles such as buildings and vehicles before arriving at the receiver. Because signals arriving at the receiver along different paths experience varying delays and fading, a phenomenon known as multipath, the channel exhibits a fluctuating envelope in the frequency domain, resulting in frequency selectivity. Furthermore, the relative motion of the transmitter and receiver causes each signal path to have its own Doppler shift, creating selectivity in the time domain, known as a time-varying channel. High-speed mobility channels are often both time- and frequency-selective. The primary communication modulation scheme used in 4G systems is Orthogonal Frequency Division Multiplexing (OFDM). OFDM is a multi-carrier modulation scheme that leverages the orthogonality between subcarriers to achieve high spectrum efficiency. However, in high-speed mobility, subcarrier orthogonality is compromised, and the maximum Doppler shift is proportional to velocity. Therefore, OFDM performance deteriorates with increasing velocity. In short, traditional OFDM cannot effectively meet the communication requirements in this scenario.

[0051] OTFS has been shown to be more suitable for communications in high-speed mobile scenarios than traditional modulation methods. OTFS performs signal modulation and demodulation in the Delay-Doppler domain. A dual-time-frequency selective channel can be represented in the Delay-Doppler domain using a few taps, resulting in a sparse channel in the Delay-Doppler domain. In this dual-time-frequency selective channel, the delay and Doppler frequency offset of each signal path are considered to be roughly constant over short time periods (e.g., a few milliseconds). The signal inserted in the Delay-Doppler domain can be transformed to the time-frequency domain using an inverse sigmoid Fourier transform (ISFFT), which is implemented by combining a fast Fourier transform (FFT) and an inverse fast Fourier transform (IFFT). The time-frequency domain signal can then be transformed back to the Delay-Doppler domain using a sigmoid Fourier transform (SFFT). After being transformed into a time-frequency signal, the Delay-Doppler domain signal can be viewed as a series of continuous multicarrier symbols in the time domain. Existing multicarrier modulation schemes, such as OFDM, can then be used for signal transmission, ensuring good compatibility with existing signal transmission systems. In the transmitted OTFS signal, the signal modulated in the delay-Doppler domain is transformed to cover the entire time-frequency domain, and can be considered to experience the same time-frequency fading, while also generating diversity gain. Therefore, OTFS has significant performance advantages over traditional methods in time-varying channels.

[0052] However, OTFS still faces some challenges in practical applications. In OTFS systems, it is difficult to achieve ideal transmit and receive filters that satisfy time-frequency biorthogonality, so rectangular filters are usually used instead in practice. At this time, in high-speed mobile scenarios, OTFS signals will also be subject to inter-carrier interference and inter-symbol interference. In addition, because in the OTFS system, an OTFS symbol block is equivalent to several consecutive multi-carrier symbols, when the received signal is represented as the product of the transmitted signal and the channel matrix, the size of the channel matrix is ​​usually very large. Using traditional linear detection methods for interference elimination involves matrix inversion, which has high time and space complexity. When using nonlinear methods for detection, the complexity cannot be reduced, and the key parameters are usually manually set based on simulation results. Whether these parameters are optimal is worth discussing.

[0053] The disclosed embodiment introduces a machine learning model to improve OTFS signal detection performance.

[0054] Deep learning is one of the fastest-growing approaches in machine learning. It utilizes a multi-layered neural network structure composed of complex connections, which is more expressive than a network with fewer layers. Deep learning consists of two phases: training and application. In the training phase, data is input into the neural network to generate output. The difference between the actual output and the ideal result is calculated, and the neural network weights are optimized using a backpropagation algorithm. This process is repeated repeatedly until the neural network reaches the set convergence criteria, resulting in the final optimized model. In the application phase, real data is directly input into the trained network model to generate the desired output, typically with very fast execution speed. In this context, deep learning reduces the time required to execute online tasks at the expense of offline training time.

[0055] In high-speed mobile scenarios, traditional modulation methods such as OFDM suffer from poor performance due to the effects of Doppler frequency offset. OTFS can effectively combat Doppler frequency offset interference, but its signal detection methods are complex and existing solutions are less practical. To address these issues, this paper proposes a general OTFS signal detection method based on deep learning to reduce signal detection complexity and improve detection performance, thereby enhancing the practicality of OTFS in high-speed mobile communication scenarios.

[0056] Figure 2 The method provided by the embodiment of the present disclosure can be executed by any electronic device with computing and processing capabilities. For example, the method can be executed by the above-mentioned Figure 1 The server or terminal device in the embodiment may be executed, or the server and the terminal device may be executed together. In the following embodiments, the server is used as the execution subject for example, but the present disclosure is not limited to this.

[0057] Reference Figure 2 The signal detection method provided by the embodiment of the present disclosure may include the following steps.

[0058] Step S202: Acquire a target received signal.

[0059] During the communication process, the signal sent by the signal transmitter can be called the target transmission signal. After the target transmission signal reaches the receiving end through the channel, it can be called the target reception signal. The signal obtained after signal detection of the target reception signal can be called the target signal.

[0060] In some embodiments, OTFS may be used as a communication modulation scheme to obtain a target received signal.

[0061] Among them, using OTFS as the communication modulation method, the received signal can include: Figure 3 The steps included in the OTFS modulation block diagram shown.

[0062] Before modulating the target transmitted signal, a two-dimensional delay-Doppler grid Λ = {(lΔτ,kΔv,l = 0, 1, 2, ..., M-1, k = 0, 1, 2, ..., N-1}) can be constructed at the transmitter. Its size is M × N, where the total number of points in the delay dimension is M and the total number of points in the Doppler dimension is N. Δτ is the delay dimension grid interval, and Δν is the Doppler dimension grid interval. The target transmitted signal is placed on the delay-Doppler grid point, denoted as x[k, l], in vector form x.

[0063] In some embodiments, an inverse sigmoid Fourier transform (ISFFT) may be used to transform the target transmitted signal x[k, l] on the delay-Doppler grid into the time-frequency domain, as shown in equation (1):

[0064]

[0065] Where m = 0, 1, 2, ..., M-1, n = 0, 1, 2, ..., N-1, X[m, n] is the symbol of the target signal in the time-frequency domain, and its vector form is denoted by X.

[0066]

[0067] in F M and F N is the normalized discrete Fourier transform matrix, (·) H Represents the conjugate transpose of a matrix.

[0068] At this point, the signal on the delay-Doppler two-dimensional grid is converted into a symbol on the time-frequency two-dimensional grid. The time-frequency grid is recorded as Γ = {(mΔf,nT), m = 0, 1, 2, ..., M-1, n = 0, 1, 2, ..., N-1}. Where Δf is the frequency domain grid interval and T is the time domain grid interval. The delay-Doppler grid and the time-frequency grid satisfy the relationship shown in Equation (3):

[0069]

[0070] Use Heisenberg transform to process the signal X[m,n]. As shown in formula (4):

[0071]

[0072] where g tx (t) is the transmit filter. The matrix representation of formula (4) is:

[0073]

[0074] Among them G tx =diag[g tx (0),g tx (T / M),...,g tx ((M-1)T / M)], diag() is used to construct a diagonal matrix, VEC() represents pulling a matrix into a column vector, S is the transmitted signal, expressed as a two-dimensional matrix, and s is the signal vector finally sent to the receiving end. When the transmit filter is a rectangular window, G tx =I M At this point, the Heisenberg transform is simplified to the Fourier transform IFFT, and equation (4) is simplified to:

[0075]

[0076] in represents the Kronecker product.

[0077] After passing through the transmit filter (e.g., a rectangular filter), the signal s(t) enters the wireless channel for propagation. Considering the high-speed mobile scenario, the channel can be expressed as:

[0078]

[0079] Where P represents the total number of multipaths, τ i is the delay of the i-th path, v i is the Doppler frequency deviation of the i-th path, h i is its channel gain, and δ(·) is the impulse function.

[0080]

[0081] wherein is a delay dimension index, is a Doppler dimension index.

[0082] At the receiving end, there are:

[0083] r(t) = ∫∫h(τ, v)s(t - τ)e j2πv(t-τ) dτdν+w(t) (9)

[0084] wherein r(t) represents a target received signal received by the receiving end, and a vector form of r(t) is denoted as r. Channel noise is denoted as w(t), which is a zero-mean additive Gaussian white noise.

[0085] In step S204, channel information corresponding to the target received signal is obtained.

[0086] In some embodiments, channel information h(τ, v) of the target received signal can be obtained.

[0087] In step S206, the target received signal and the target received signal are linearly equalized by using the channel information, so as to obtain approximate channel information and an estimated value of a target sending signal corresponding to the target received signal.

[0088] Linear equalization in communication refers to equalization of channel characteristics, that is, an equalizer at the receiving end generates a characteristic opposite to the channel, so as to offset the inter-symbol interference caused by the time-varying multipath propagation characteristics of the channel. In other words, the frequency and time selectivity of the channel is eliminated through the equalizer.

[0089] The received signal is linearly equalized by using the channel information. Direct use of linear equalization requires inversion of a matrix, which has too high complexity. Low-complexity linear equalization can be used to quickly detect the received signal, which is used to initialize the signal detection result. The linear equalization with better performance is mainly linear minimum mean square error (LMMSE) equalization. Then, the multiplication operation of the received signal and the channel matrix is simplified by using the idea of reducing the complexity of linear equalization, so as to solve the residual error next.

[0090] In step S207, a residual error of a signal detection model is determined by processing the target received signal, the approximate channel information, and the estimated value of the target sending signal, and the signal detection model further includes a signal detection unit.

[0091] In some embodiments, an estimated sending signal can be determined according to the approximate channel and the target received signal, and then a signal residual error is determined based on the estimated sending signal and the estimated value of the target sending signal. The method for determining the signal residual error is not limited in the present disclosure.

[0092] Step S210: Processing the estimated value of the target transmitted signal and the signal residual by the signal detection unit to complete signal detection for the target received signal and obtain a target signal.

[0093] like Figure 4 As shown, after receiving the target signal, the signal detection model completed by linear training can be used to perform signal detection processing on the target signal to obtain the target signal

[0094] The signal detection model can be obtained offline through a known transmitted signal, a received signal corresponding to the transmitted signal, and channel training.

[0095] like Figure 4 As shown, the present disclosure can adopt a fixed network structure model (signal detection model). The generated network model does not change due to changes in the wireless channel. The model can be trained offline and directly used for demodulation online, which has strong versatility. During online demodulation, only the channel matrix and the received signal need to be input. After the fixed model is operated, good signal detection results can be obtained, which not only reduces the complexity of signal detection but also improves the accuracy of signal detection.

[0096] This paper designs a network based on the residual characteristics in OTFS to extract residual information for iterative optimization. Under the same scale, fewer parameters need to be trained, the network converges faster, and the performance is better for the OTFS system. It also has low complexity and greater practicality.

[0097] Figure 5 The figure is a flow chart showing a method for linearly equalizing a target received signal according to an exemplary embodiment.

[0098] In some embodiments, the target received signal received by the receiving end is a signal in the time domain. Figure 5 The flowchart of the method for performing linear equalization on the target received signal may include the following steps.

[0099] Step S502: Acquire a Doppler domain input-output relationship of the target received signal, the channel information, and the target transmitted signal in the delay-Doppler domain, wherein the Doppler domain input-output relationship includes a delay dimension parameter.

[0100] First, it is necessary to determine the time domain input-output relationship among the target received signal, the channel information, and the target transmitted signal in the time domain.

[0101] In some embodiments, the relationship between the received signal r and the transmitted signal s in the time domain can be expressed as:

[0102] r=Hs+w (10)

[0103] Secondly, the time domain input-output relationship is transformed by inverse Fourier transform and inverse sigmoid Fourier transform to determine the Doppler domain input-output relationship of the target received signal, the channel information and the target transmitted signal in the delayed Doppler domain.

[0104] At the receiving end, the corresponding inverse transform of the received signal r can be performed to obtain the received delayed Doppler signal Y. First, the multi-carrier symbol r is transformed by FFT, and we have:

[0105] Y=F M (unvec(r)) (11)

[0106] Among them, unvec() is the inverse function of vec().

[0107] The time-frequency signal Y is transformed into the delay-Doppler domain after SFFT, as shown in equation (12):

[0108]

[0109] Formula (12) can be expressed as a matrix:

[0110]

[0111] Then the Doppler domain input-output relationship in the delay-Doppler domain can be expressed as:

[0112]

[0113] in,

[0114]

[0115] The matrix form of the input-output relationship in the Doppler domain is:

[0116]

[0117] It should be noted that the input-output relationship in the Doppler domain may include a delay dimension parameter l.

[0118] Step S504 : replacing the delay dimension parameter in the Doppler domain input-output relationship by the mean value of the delay dimension parameter, so as to convert the Doppler domain input-output relationship into a Doppler domain approximate input-output relationship.

[0119] This paper proposes an approximate linear detection method that uses channel information to generate an approximate time-frequency channel matrix and simultaneously performs approximate LMMSE detection. For ease of expression and analysis, wireless channel noise is not included in the formula. When inter-symbol interference is not considered, the input-output relationship of OTFS in the delay-Doppler domain can be expressed as:

[0120]

[0121] In order to facilitate rapid signal detection and neural network design, the input-output relationship is approximated. For the received signal:

[0122]

[0123] The input-output relationship in the Doppler domain mentioned above includes the delay dimension parameter l'.

[0124] After transforming the delayed Doppler domain signal into the time-frequency domain through ISFFT, the present disclosure adopts an approximate processing method to replace l′ in formula (20) with the expected M / 2 of l′, that is,

[0125]

[0126] make At this time, the received signal in the time-frequency domain can be expressed as the dot product of the transmitted signal and the channel matrix, as shown in formula (20):

[0127]

[0128] Step S505: Determine the approximate channel information from the Doppler domain approximate input-output relationship.

[0129] According to formula (20), the approximate channel information can be determined

[0130] Step S508 : determining an estimated value of the target transmitted signal by using the Doppler domain approximate input-output relationship, the approximate channel information, and the target received signal.

[0131] The result after LMMSE equalization is:

[0132]

[0133] in(·) * The detection results in the time-frequency domain can be transformed into the delay-Doppler domain signal through SFFT.

[0134] According to formula (21), the approximate channel information is known and can be calculated, Y[m,n] can be determined based on the target received signal, then the estimated value of the target sent signal is It is also possible to ask for it.

[0135] Compared with the traditional linear aggregation method, the linear equalization method provided by the embodiment of the present disclosure has lower complexity and obtains accurate results.

[0136] Figure 6FIG. 1 is a flow chart showing a method for determining a signal residual according to an exemplary embodiment. Figure 6 , the above-mentioned signal residual determination method may include the following steps.

[0137] Step S602: performing decision processing on the estimated value of the target transmitted signal to determine a decision estimated value.

[0138] In some embodiments, the estimated value of the target transmitted signal includes an estimated real part and an estimated imaginary part; then, the estimated value of the target transmitted signal is subjected to decision processing to determine a decision estimated value, including: performing decision processing on the estimated real part and the estimated imaginary part respectively to obtain a decision real part and a decision imaginary step; and determining the decision estimated value based on the decision real part and the decision imaginary step.

[0139] In some embodiments, QPSK (quadrature phase shift keying) may be used as the information symbol, for example, using the function in equation (22) for decision processing:

[0140]

[0141] ψ t (·) can convert all variables into data within [-1, +1]. t In (·), when x≤-|t|, the decision is -1, when x≥|t|, the decision is +1, and when -|t|<x<|t|, the output is x / |t|, which is similar to the simplified soft information output. In some embodiments, t=0.05 can be taken to and Make a certain degree of judgment while retaining some soft information to facilitate training. The real and imaginary parts of are passed through this function respectively, and the signs can be obtained after a certain degree of judgment:

[0142]

[0143] in, and The estimated values ​​of the signals sent to the targets are The real and imaginary parts of .

[0144] Step S604: Process the target received signal using the approximate channel information to determine a first transmitted signal.

[0145] In some embodiments, the first transmission signal can be calculated by first performing a dot multiplication in the time-frequency domain and then converting it into a delayed Doppler signal.

[0146]

[0147] where ⊙ represents the Hadamard product, is the approximate channel information.

[0148] Step S606: Process the decision estimate value using the approximate channel information to determine a second transmitted signal.

[0149] In some embodiments, the second transmission signal can be calculated by first performing a dot multiplication in the time-frequency domain and then converting it into a delayed Doppler signal. as follows:

[0150]

[0151] where ⊙ represents the Hadamard product, is the approximate channel information.

[0152] Step S607: Determine the signal residual according to the first transmitted signal and the second transmitted signal.

[0153] In some embodiments, the first transmit signal and the second transmit signal may be processed according to formula (26) to obtain a signal residual V.

[0154]

[0155] The disclosed embodiment performs decision processing on the estimated value of the target transmitted signal, thereby removing certain redundant information while retaining some soft information, thereby reducing the training complexity and improving the accuracy of signal detection.

[0156] Figure 7 The figure is a flow chart of a target signal detection method according to an exemplary embodiment.

[0157] Figure 8 The image shows the absolute value of the residual error when four symbol decisions are erroneously made in the OTFS system at a signal-to-noise ratio of 20 dB. When an erroneous decision occurs, the residual image exhibits four large peaks, corresponding to the locations of the errors. In addition to these large peaks, there are several smaller peaks, which are caused by channel multipath. Figure 9 The image shows the absolute value of the residual error when four symbols are incorrectly judged in a MIMO (multiple input, multiple output) system. In a MIMO system, larger dimensions reduce the power of small peak paths. However, in an OTFS system, the signal matrix dimension has little impact on the wireless channel, but it is still significantly affected by small peak paths. One way to reduce the impact of small peak paths is to input the statistical characteristics of the residual into the neural network training. To more effectively approach the target value during nonlinear iteration, the residual is processed.

[0158] In some embodiments, the signal residual includes the real part of the residual V R and the residual imaginary part V I , the real part of the residual V R It can include the absolute value of the real part of the residual V R,abs and the real part of the sign residual V R,sign , residual imaginary part V I Including the absolute value of the residual imaginary part V I,abs and the positive and negative sign of the residual imaginary part V I,sign .

[0159] Step S702: performing feature extraction processing on the real part of the absolute value residual by the signal detection unit to obtain an estimated value of the real part of the absolute value residual.

[0160] In some embodiments, [] M ,and[] ,N denote the mean along the first dimension (e.g., the Doppler dimension in the delay-Doppler domain) and the second dimension (e.g., the delay dimension in the delay-Doppler domain), respectively, then:

[0161]

[0162] In some embodiments, the real part of the absolute value residual V R,abs The mean value is calculated along the Doppler plane to obtain the real part of the absolute value Doppler dimension mean [V R,abs ] M ,[n]; the absolute value residual real part V R,abs The mean value is calculated along the delay dimension to obtain the real part of the absolute value delay dimension mean [V R,abs ], N [m]; The absolute value residual real part estimate is determined by the absolute value Doppler dimension mean real part, the absolute value delay dimension mean real part and the absolute value residual real part, for example, the absolute value residual real part estimate can be determined by the following formula.

[0163]

[0164] Where W R,1 is a trainable weight matrix of size 1×N, W R,2 is a trainable weight matrix of size M×1.

[0165] Step S704: performing feature extraction processing on the imaginary part of the absolute value residual by the signal detection unit to obtain an estimated value of the imaginary part of the absolute value residual.

[0166] In some embodiments, the imaginary part of the absolute value residual V I,abs The mean value is calculated along the Doppler plane to obtain the imaginary part of the absolute value Doppler dimension [V I,abs] M ,[n]; for the absolute value residual imaginary part V I,abs The mean value is calculated along the delay dimension to obtain the imaginary part of the absolute value delay dimension mean [V I,abs ] ,N [m]; The absolute value residual imaginary part estimate is determined by the absolute value Doppler dimension mean imaginary part, the absolute value delay dimension mean imaginary part and the absolute value residual imaginary part, for example, the absolute value residual imaginary part estimate can be determined by the following formula.

[0167]

[0168] Among them, W I,1 is a trainable weight matrix of size 1×N, W I,2 is a trainable weight matrix of size M×1.

[0169] Step S706 , performing activation processing on the real part estimated value of the absolute value residual and the imaginary part estimated value of the absolute value residual to obtain a real part activation absolute estimated value and an imaginary part activation absolute estimated value.

[0170] In some embodiments, an activation function may be used to estimate the real part of the absolute value residual. and the estimated imaginary part of the absolute value residual The activations are processed separately to obtain the absolute estimates of the real and imaginary activations.

[0171] For example, we can use the activation function ρ to estimate the real part of the absolute value residual and the estimated imaginary part of the absolute value residual Activation processing is performed separately to obtain the absolute estimate of the real part activation and the absolute estimate of the imaginary activation and

[0172] Step S707 : determining the real part of the target signal through the real part of the residual, the positive and negative signed real part of the residual, and the real part activation absolute estimate.

[0173] In some embodiments, the processed residual can be used to update the detection signal and design the signal detection network unit. For example, formula (30) can be used to update the decision symbol using the residual to determine the real part of the target signal

[0174]

[0175] Where W R,3 is the parameter value to be trained, and ρ(·) is the activation function.

[0176] Step S79: determine the imaginary part of the target signal through the residual imaginary part, the positive and negative signed residual imaginary part, and the imaginary part activation absolute estimation value.

[0177] For example, formula (31) can be used to update the decision symbol using the residual to determine the imaginary part of the target signal

[0178]

[0179] Among them, W I,3 is the parameter value to be trained, and ρ(·) is the activation function.

[0180] Step S711: Determine the target signal according to the real part of the target signal and the imaginary part of the target signal.

[0181]

[0182] The method used in the present disclosure is to add the mean of the absolute value of the signal residual along the delay axis and the mean along the Doppler axis to the neural network training in the OTFS system. The advantage of doing so is that it can reduce the impact of small peak diameters and noise to a certain extent, and the performance improvement effect is more significant when the noise power is large and the bit error rate is large. The idea of ​​designing a neural network in this example is to use the absolute value of the signal residual V to find the point where the previous iteration result may have errors, and then use the value of the corresponding position of V to update the possible erroneous symbols. The designed neural network intends to train the residual, solve the position where the large peak appears and correct the wrong symbol. In order to reduce the impact of small peaks, the mean of the residual along the delay dimension and the mean along the Doppler dimension are both added to the training.

[0183] Figure 10 The diagram shows a network structure of a signal detection model according to an exemplary embodiment.

[0184] like Figure 10 As shown, the signal detection model may include one signal detection unit (1-th Layer) or multiple signal detection units (1-th Layer to K-th Layer, where K represents the number of signal detection units), and the present disclosure does not impose any restrictions on this.

[0185] The embodiment of the present disclosure will take the i-th signal detection unit (i-th Layer) in the i-th signal detection model as an example to illustrate the signal detection method. The operation methods of other signal detection units can refer to this embodiment.

[0186] Combine Figure 11 The signal detection method that can be implemented by the i-th signal detection unit in the above figure may include the following steps.

[0187] Step S1101: Use OTFS as a modulation method to obtain a target received signal.

[0188] In some embodiments, a target received signal can be obtained. The target received signal can be the received received signal X, or the signal X after the signal received by the receiving end is processed by the i-1th signal detection unit. i-1 , this disclosure does not limit this.

[0189] Step S1102: Perform low-complexity linear equalization.

[0190] In some embodiments, the channel information corresponding to the target received signal can be obtained Then, low-complexity linear equalization processing is performed on the target received signal using the channel information to obtain approximate channel information and an estimated value of a target transmitted signal, where the target transmitted signal corresponds to the target received signal.

[0191] Step S1103: quickly calculate the signal residual.

[0192] In some embodiments, the function (f countV ())accomplish:

[0193] Perform decision processing on the estimated value of the target transmitted signal to determine a decision estimate; process the target received signal using the approximate channel information to determine a first transmitted signal; process the decision estimate using the approximate channel information to determine a second transmitted signal; and determine the signal residual based on the first transmitted signal and the second transmitted signal.

[0194] Step S1104: design a method to extract key information based on the residual characteristics.

[0195] In some embodiments, the signal detection unit can perform feature extraction processing on the real part of the absolute value residual to obtain an absolute value residual real part estimate; the signal detection unit can perform feature extraction processing on the imaginary part of the absolute value residual to obtain an absolute value residual imaginary part estimate; the absolute value residual real part estimate and the absolute value residual imaginary part estimate are activated to obtain a real part activation absolute estimate and an imaginary part activation absolute estimate; the real part of the target signal is determined by the real part of the residual, the positive and negative sign residual real part, and the real part activation absolute estimate; the imaginary part of the target signal is determined by the imaginary part of the residual, the positive and negative sign residual imaginary part, and the imaginary part activation absolute estimate; the target signal is determined based on the real part of the target signal and the imaginary part of the target signal.

[0196] Among them, performing feature extraction processing on the real part of the absolute value residual by the signal detection unit to obtain an estimated value of the real part of the absolute value residual can include: performing mean processing on the real part of the absolute value residual along the Doppler dimension to obtain the mean real part of the absolute value Doppler dimension; performing mean processing on the real part of the absolute value residual along the delay dimension to obtain the mean real part of the absolute value delay dimension; and determining the estimated value of the real part of the absolute value residual by the mean real part of the absolute value Doppler dimension, the mean real part of the absolute value delay dimension and the absolute value residual real part.

[0197] Among them, performing feature extraction processing on the absolute value residual imaginary part by the signal detection unit to obtain an estimated value of the absolute value residual imaginary part can include: performing mean processing on the absolute value residual imaginary part along Doppler to obtain an absolute value Doppler dimension mean imaginary part; performing mean processing on the absolute value residual imaginary part along the delay dimension to obtain an absolute value delay dimension mean imaginary part; determining the estimated value of the absolute value residual imaginary part by the absolute value Doppler dimension mean imaginary part, the absolute value delay dimension mean imaginary part and the absolute value residual imaginary part.

[0198] Step S1105: Update the signal detection result using the residual information to determine the target signal.

[0199] In some embodiments, formula (30) can be used to update the decision symbol using the residual to determine the real part of the target signal Then, using formula (31), the decision symbol is updated using the residual to determine the imaginary part of the target signal Finally, based on formula (32), the updated detection result is determined to obtain a detection signal X i , so that based on the detection signal X i Identify the target signal.

[0200] Step S1106: Offline training of the network to generate a model and online direct signal detection

[0201] The present disclosure utilizes a fixed network structure model (signal detection model). The generated network model remains unchanged by changes in the wireless channel. This model can be trained offline and then used directly for demodulation online, demonstrating its versatility. Online demodulation requires only the input of the channel matrix and the received signal. Following the fixed model, superior signal detection results are achieved, reducing the complexity of signal detection while improving its accuracy.

[0202] This paper designs a network based on the residual characteristics in OTFS to extract residual information for iterative optimization. Under the same scale, fewer parameters need to be trained, the network converges faster, and the performance is better for the OTFS system. It also has low complexity and greater practicality.

[0203] In the designed network, in order to reduce the operation, it is usually stored and calculated once.

[0204] and Then store it for later processing CountV The neural network input is initialized using the proposed approximate LMMSE algorithm, with the input information including the received signal, the initialized detection signal, and the channel matrix. Several identical neural network units are connected in series, each simulating a nonlinear iterative process. The input signal is first simply judged, then combined with the received signal to generate a residual, train the parameters, output the optimized result, update the detected signal, and continue the optimization process with the next network unit.

[0205] like Figure 10 As shown, the signal detection model for performing signal detection provided by the embodiment of the present disclosure may include one signal detection unit or multiple signal detection units, and the present disclosure does not impose any limitation on this.

[0206] If the signal detection model includes multiple signal detection units, it is necessary to select an appropriate loss function and optimization method to train and optimize the network when training the signal detection model. There are usually two methods to train this type of neural network. One is to combine the network outputs of all multiple signal detection units and use the weighted average of the MSE of the detection signal output at each level and the true signal as the loss function to train the entire signal detection network together. In this case, the loss function is:

[0207]

[0208] Another training method is to train each level of neural network unit separately, fix the parameters after convergence, and then train the next level of neural network unit. The loss function of each level of training is the MSE of the network output and the actual transmitted signal, that is:

[0209]

[0210] After the network model is trained, the desired output is obtained by inputting the corresponding data into the model. After hard decision, the original transmitted bit information is obtained. A sufficient amount of simulation data is generated using various existing channel models to train the network and generate a model, which is then used to detect OTFS signals online.

[0211] In order to verify the practical performance of the embodiments of the present disclosure, the author conducted multiple Monte Carlo simulation experiments. Figure 12This figure compares the bit error rates of various signal detection methods under the same conditions. The "Ideal LMMSE" uses an ideal filter, which is not achievable in practice, but does determine the lower bound of the LMMSE algorithm's bit error rate. The message passing algorithm is a classic nonlinear iterative algorithm in OTFS, DetNet is a MIMO signal detection network, and ScNet is an improved version of DetNet. The method proposed in this disclosure slightly outperforms the message passing algorithm when the signal-to-noise ratio is below 10dB, and performs even better when the signal-to-noise ratio is high.

[0212] The time consumption comparison between the message passing algorithm and the proposed method under the same conditions can obtain the following comparison results.

[0213] Table 1

[0214]

[0215] It can be seen that the proposed method takes much less time than the message passing algorithm, effectively reducing the complexity of signal detection. In terms of bit error rate performance, the proposed method outperforms the LMMSE algorithm and the compared MIMO signal detection network.

[0216] Table 2 compares the number of network connections between the proposed method and the signal detection network in MIMO. The complexity of the proposed method and ScNet in online demodulation is comparable. The number of network connections represents the number of parameters that need to be trained. The number of network connections in the proposed method is much smaller than that of ScNet, so the network training is faster.

[0217] Table 2

[0218]

[0219] The disclosed embodiments provide a deep learning-based OTFS signal detection method, characterized by using deep learning methods for signal detection in the OTFS system and implementing direct online demodulation using an offline training network to achieve communication in high-speed mobile scenarios. Furthermore, the linear equalization algorithm and fast residual calculation method used in the example simplify matrix multiplication and division operations to matrix dot multiplication and dot division by approximating key parameters. Finally, the key to neural network design lies in the processing of residuals. The residual characteristics are analyzed to design a neural network that reduces the influence of small peaks and extracts large peaks. In the example, the large peak information is extracted by training with the mean along the delay and Doppler dimensions.

[0220] Figure 13 A block diagram of a signal detection device according to an exemplary embodiment is shown. Figure 13 The signal detection device 1300 provided in the embodiment of the present disclosure may include: a signal receiving module 1301, a channel information acquisition module 1302, a linear equalization processing module 1303, a residual calculation module 1304 and a signal detection module 1305.

[0221] Among them, the signal receiving module 1301 can be used to obtain the target received signal; the channel information acquisition module 1302 can be used to obtain the channel information corresponding to the target received signal; the linear equalization processing module 1303 can be used to perform linear equalization processing on the target received signal and the target received signal through the channel information to obtain approximate channel information and an estimated value of the target transmitted signal, and the target transmitted signal corresponds to the target received signal; the residual calculation module 1304 can be used for the residual calculation unit of the signal detection model to process the target received signal, the approximate channel information and the estimated value of the target transmitted signal to determine a signal residual, and the signal detection model also includes a signal detection unit; the signal detection module 1305 can be used to process the estimated value of the target transmitted signal and the signal residual through the signal detection unit to complete signal detection for the target received signal and obtain the target signal.

[0222] In some embodiments, the target received signal is a signal in the time domain; wherein, the linear equalization processing module 1303 may include: a Doppler domain input-output relationship acquisition submodule, an approximate replacement submodule, an approximate channel information acquisition module, and an estimated value determination submodule.

[0223] Among them, the Doppler domain input-output relationship acquisition submodule can be used to obtain the Doppler domain input-output relationship of the target received signal, the channel information and the target transmitted signal in the delayed Doppler domain, and the Doppler domain input-output relationship includes a delay dimension parameter; the approximate replacement submodule can be used to replace the delay dimension parameter in the Doppler domain input-output relationship by the mean of the delay dimension parameter to convert the Doppler domain input-output relationship into a Doppler domain approximate input-output relationship; the approximate channel information acquisition module can be used to determine the approximate channel information from the Doppler domain approximate input-output relationship; the estimated value determination submodule can be used to determine the estimated value of the target transmitted signal through the Doppler domain approximate input-output relationship, the approximate channel information and the target received signal.

[0224] In some embodiments, the Doppler domain input-output relationship acquisition submodule may include: a time domain input-output relationship determination unit and a Doppler domain input-output relationship determination unit.

[0225] Among them, the time domain input-output relationship determination unit can be used to determine the time domain input-output relationship of the target received signal, the channel information and the target transmitted signal in the time domain; the Doppler domain input-output relationship determination unit can be used to transform the time domain input-output relationship through inverse Fourier transform and inverse sigmoid Fourier transform to determine the Doppler domain input-output relationship of the target received signal, the channel information and the target transmitted signal in the delayed Doppler domain.

[0226] In some embodiments, the residual calculation module 1304 may include: a decision estimate determination submodule, a first transmit signal determination submodule, a second transmit signal determination submodule, and a signal residual determination submodule.

[0227] Among them, the decision estimate value determination submodule can be used to perform decision processing on the estimated value of the target transmitted signal to determine the decision estimate value; the first transmitted signal determination submodule can be used to process the target received signal through the approximate channel information to determine the first transmitted signal; the second transmitted signal determination submodule can be used to process the decision estimate value through the approximate channel information to determine the second transmitted signal; the signal residual determination submodule can be used to determine the signal residual based on the first transmitted signal and the second transmitted signal.

[0228] In some embodiments, the estimated value of the target transmitted signal includes an estimated real part and an estimated imaginary part; wherein the decision estimate value determination submodule may include: a real part decision unit and a decision estimate value determination unit.

[0229] Among them, the real part decision unit can be used to perform decision processing on the estimated real part and the estimated imaginary part respectively to obtain the decision real part and the decision imaginary step; the decision estimate value determination unit can be used to determine the decision estimate value based on the decision real part and the decision imaginary step.

[0230] In some embodiments, the signal residual includes a real part of the residual and an imaginary part of the residual, the real part of the residual includes an absolute value real part of the residual and a positive and negative sign real part of the residual, and the imaginary part of the residual includes an absolute value imaginary part of the residual and a positive and negative sign imaginary part of the residual; wherein, the signal detection module 1305 may include: an absolute value real part estimation value acquisition submodule, an absolute value imaginary part estimation value acquisition submodule, an activation absolute estimation value acquisition submodule, a real part determination submodule of the target signal, an imaginary part determination submodule of the target signal, and a target signal determination submodule.

[0231] Among them, the absolute value residual real part estimation value acquisition submodule can be used to perform feature extraction processing on the absolute value residual real part through the signal detection unit to obtain the absolute value residual real part estimation value; the absolute value residual imaginary part estimation value acquisition submodule can be used to perform feature extraction processing on the absolute value residual imaginary part through the signal detection unit to obtain the absolute value residual imaginary part estimation value; the activation absolute estimation value acquisition submodule can be used to perform activation processing on the absolute value residual real part estimation value and the absolute value residual imaginary part estimation value to obtain the real part activation absolute estimation value and the imaginary part activation absolute estimation value; the real part determination submodule of the target signal can be used to determine the real part of the target signal through the residual real part, the positive and negative sign residual real part and the real part activation absolute estimation value; the imaginary part determination submodule of the target signal can be used to determine the imaginary part of the target signal through the residual imaginary part, the positive and negative sign residual imaginary part and the imaginary part activation absolute estimation value; the target signal determination submodule can be used to determine the target signal according to the real part of the target signal and the imaginary part of the target signal.

[0232] In some embodiments, the absolute value residual real part estimation value acquisition submodule may include: an absolute value Doppler dimension mean real part acquisition unit, an absolute value delay dimension mean real part acquisition unit, and an absolute value residual real part estimation determination unit.

[0233] Among them, the absolute value Doppler dimension mean real part acquisition unit can be used to perform mean processing on the absolute value residual real part along Doppler to obtain the absolute value Doppler dimension mean real part; the absolute value delay dimension mean real part acquisition unit can be used to perform mean processing on the absolute value residual real part along the delay dimension to obtain the absolute value delay dimension mean real part; the absolute value residual real part estimation determination unit can be used to determine the absolute value residual real part estimation value through the absolute value Doppler dimension mean real part, the absolute value delay dimension mean real part and the absolute value residual real part.

[0234] The absolute value residual imaginary part estimation value acquisition submodule may include: an absolute value Doppler dimension mean imaginary part determination unit, an absolute value delay dimension mean imaginary part determination unit and an absolute value residual imaginary part estimation value determination unit.

[0235] Among them, the absolute value Doppler dimension mean imaginary part determination unit can be used to perform mean processing on the absolute value residual imaginary part along Doppler to obtain the absolute value Doppler dimension mean imaginary part; the absolute value delay dimension mean imaginary part determination unit can be used to perform mean processing on the absolute value residual imaginary part along the delay dimension to obtain the absolute value delay dimension mean imaginary part; the absolute value residual imaginary part estimation value determination unit can be used to determine the absolute value residual imaginary part estimation value through the absolute value Doppler dimension mean imaginary part, the absolute value delay dimension mean imaginary part and the absolute value residual imaginary part.

[0236] Since the functions of the apparatus 1300 have been described in detail in the corresponding method embodiments, they will not be described in detail in this disclosure.

[0237] The modules (and / or submodules and / or) units involved in the embodiments described in the present application may be implemented in software or in hardware. The modules (and / or submodules and / or) units described may also be provided in a processor. The names of these modules (and / or submodules and / or) units do not, in certain circumstances, constitute limitations on the modules (and / or submodules and / or) units themselves.

[0238] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the above-mentioned module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of the boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0239] Furthermore, the figures above are merely illustrative of the processes included in the methods according to exemplary embodiments of the present disclosure and are not intended to be limiting. It is readily understood that the processes illustrated in the figures above do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0240] Figure 14 Schematic diagram of the structure of an electronic device suitable for implementing the embodiment of the present disclosure is shown. Figure 14The electronic device 1400 shown is only an example and should not limit the functions and scope of use of the embodiments of the present disclosure.

[0241] like Figure 14 As shown, electronic device 1400 includes a central processing unit (CPU) 1401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1402 or a program loaded from a storage portion 1408 into a random access memory (RAM) 1403. Various programs and data required for the operation of electronic device 1400 are also stored in RAM 1403. CPU 1401, ROM 1402, and RAM 1403 are connected to each other via a bus 1404. An input / output (I / O) interface 1405 is also connected to bus 1404.

[0242] The following components are connected to the I / O interface 1405: an input section 1406 including a keyboard, a mouse, and the like; an output section 1407 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and speakers; a storage section 1408 including devices such as a hard disk; and a communication section 1409 including a network interface card such as a LAN card or a modem. The communication section 1409 performs communication processing via a network such as the Internet. A drive 1410 is also connected to the I / O interface 1405 as needed. Removable media 1411, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 1410 as needed, so that computer programs read therefrom can be installed in the storage section 1408 as needed.

[0243] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 1409, and / or installed from a removable medium 1411. When the computer program is executed by the central processing unit (CPU) 1401, the above-mentioned functions defined in the system of the present application are executed.

[0244] It should be noted that the computer-readable storage medium shown in the present disclosure can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this application, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. Program code contained on a computer-readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wireline, optical cable, RF, or any suitable combination thereof.

[0245] As another aspect, the present application also provides a computer-readable storage medium, which may be included in the device described in the above embodiment; or it may exist independently and not be assembled into the device. The above-mentioned computer-readable storage medium carries one or more programs, and when the above-mentioned one or more programs are executed by a device, the device can implement functions including: obtaining a target received signal; obtaining channel information corresponding to the target received signal; performing linear equalization processing on the target received signal and the target received signal through the channel information to obtain approximate channel information and an estimated value of a target transmitted signal, the target transmitted signal corresponding to the target received signal; a residual calculation unit of a signal detection model processes the target received signal, the approximate channel information, and the estimated value of the target transmitted signal to determine a signal residual, the signal detection model also including a signal detection unit; the signal detection unit processes the estimated value of the target transmitted signal and the signal residual to complete signal detection for the target received signal and obtain the target signal.

[0246] According to one aspect of the present application, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the methods provided in various optional implementations of the above-described embodiments.

[0247] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described here can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution of the embodiment of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including a number of instructions for enabling a computing device (which can be a personal computer, a server, a mobile terminal, or a smart device, etc.) to execute the method according to the embodiment of the present disclosure, for example Figure 2 、 Figure 5 、 Figure 6 ,or Figure 7 One or more of the steps shown.

[0248] Other embodiments of the present disclosure will readily occur to those skilled in the art after considering the specification and practicing the disclosure herein. This disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art not claimed herein. The specification and examples are to be considered as exemplary only, with the true scope and spirit of the present disclosure being indicated by the claims.

[0249] It should be understood that the present disclosure is not limited to the detailed structures, drawings or implementations shown herein, but rather is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A signal detection method, characterized in that: include: Acquire target receiving signal; Acquiring channel information corresponding to the target received signal; performing linear equalization processing on the target received signal using the channel information to obtain approximate channel information and an estimated value of a target transmitted signal, the target transmitted signal corresponding to the target received signal; The residual calculation unit of the signal detection model processes the target received signal, the approximate channel information and the estimated value of the target transmitted signal to determine a signal residual, and the signal detection model further includes a signal detection unit; Processing the estimated value of the target transmitted signal and the signal residual by the signal detection unit to complete signal detection for the target received signal and obtain a target signal; The signal residual includes a real part of the residual and an imaginary part of the residual, the real part of the residual includes an absolute value real part of the residual and a positive and negative sign real part of the residual, and the imaginary part of the residual includes an absolute value imaginary part of the residual and a positive and negative sign imaginary part of the residual; wherein, the estimated value of the target transmitted signal and the signal residual are processed by the signal detection unit to complete signal detection for the target received signal to obtain the target signal, including: Performing feature extraction processing on the real part of the absolute value residual by the signal detection unit to obtain an estimated value of the real part of the absolute value residual; Performing feature extraction processing on the imaginary part of the absolute value residual by the signal detection unit to obtain an estimated value of the imaginary part of the absolute value residual; Performing activation processing on the real part estimated value of the absolute value residual and the imaginary part estimated value of the absolute value residual to obtain a real part activation absolute estimated value and an imaginary part activation absolute estimated value; Determine the real part of the target signal by using the real part of the residual, the positive and negative signed real part of the residual, and the real part activation absolute estimate; Determine the imaginary part of the target signal by using the imaginary part of the residual, the positive and negative sign imaginary part of the residual, and the imaginary part activation absolute estimate; The target signal is determined according to the real part of the target signal and the imaginary part of the target signal.

2. The method according to claim 1, characterized in that The target received signal is a signal in the time domain; wherein, performing linear equalization processing on the target received signal using the channel information to obtain approximate channel information and an estimated value of the target transmitted signal includes: Acquire a Doppler domain input-output relationship of the target received signal, the channel information, and the target transmitted signal in a delay-Doppler domain, wherein the Doppler domain input-output relationship includes a delay dimension parameter; Substituting the delay dimension parameter in the Doppler domain input-output relationship by the mean value of the delay dimension parameter to convert the Doppler domain input-output relationship into a Doppler domain approximate input-output relationship; determining the approximate channel information from the Doppler domain approximate input-output relationship; An estimated value of the target transmitted signal is determined based on the Doppler domain approximate input-output relationship, the approximate channel information, and the target received signal.

3. The method according to claim 2, characterized in that Obtaining a Doppler domain input-output relationship of the target received signal, the channel information, and the target transmitted signal in the delay-Doppler domain includes: Determining a time domain input-output relationship between the target received signal, the channel information, and the target transmitted signal in the time domain; The time domain input-output relationship is transformed by inverse Fourier transform and inverse sigmoid Fourier transform to determine the Doppler domain input-output relationship of the target received signal, the channel information and the target transmitted signal in the delayed Doppler domain.

4. The method according to claim 1, characterized in that The residual calculation unit of the signal detection model processes the target received signal, the approximate channel information, and the estimated value of the target transmitted signal to determine a signal residual, including: performing decision processing on the estimated value of the target transmitted signal to determine a decision estimated value; Processing the target received signal using the approximate channel information to determine a first transmitted signal; Processing the decision estimate using the approximate channel information to determine a second transmitted signal; The signal residual is determined based on the first transmit signal and the second transmit signal.

5. The method according to claim 4, characterized in that: The estimated value of the target transmitted signal includes an estimated real part and an estimated imaginary part; wherein, performing decision processing on the estimated value of the target transmitted signal to determine a decision estimated value includes: performing decision processing on the estimated real part and the estimated imaginary part respectively to obtain a decided real part and a decided imaginary part; The decision estimate is determined based on the decision real part and the decision imaginary step.

6. The method according to claim 1, characterized in that Performing feature extraction processing on the real part of the absolute value residual by the signal detection unit to obtain an estimated value of the real part of the absolute value residual includes: Performing a mean processing on the real part of the absolute value residual along the Doppler dimension to obtain the real part of the absolute value Doppler dimension mean; Performing mean processing on the real part of the absolute value residual along the delay dimension to obtain the real part of the absolute value delay dimension mean; Determine the absolute value residual real part estimate value by using the absolute value Doppler dimension mean real part, the absolute value delay dimension mean real part and the absolute value residual real part; The signal detection unit performs feature extraction processing on the imaginary part of the absolute value residual to obtain an estimated value of the imaginary part of the absolute value residual, including: Performing a mean processing on the imaginary part of the absolute value residual along Doppler to obtain the imaginary part of the absolute value Doppler mean; Performing mean processing on the imaginary part of the absolute value residual along the delay dimension to obtain the imaginary part of the absolute value delay dimension mean; The absolute value residual imaginary part estimation value is determined by the absolute value Doppler dimension mean imaginary part, the absolute value delay dimension mean imaginary part and the absolute value residual imaginary part.

7. A signal detection device, characterized in that: include: A signal receiving module, used to obtain a target receiving signal; A channel information acquisition module, configured to acquire channel information corresponding to the target received signal; a linear equalization processing module, configured to perform linear equalization processing on the target received signal and the target received signal using the channel information to obtain approximate channel information and an estimated value of a target transmitted signal, the target transmitted signal corresponding to the target received signal; a residual calculation module, wherein the residual calculation unit of the signal detection model processes the target received signal, the approximate channel information, and the estimated value of the target transmitted signal to determine a signal residual, and the signal detection model further includes a signal detection unit; a signal detection module, configured to process the estimated value of the target transmitted signal and the signal residual through the signal detection unit to complete signal detection for the target received signal and obtain a target signal; The signal residual includes a real part of the residual and an imaginary part of the residual, the real part of the residual includes an absolute value real part of the residual and a positive and negative sign real part of the residual, and the imaginary part of the residual includes an absolute value imaginary part of the residual and a positive and negative sign imaginary part of the residual; wherein, the estimated value of the target transmitted signal and the signal residual are processed by the signal detection unit to complete signal detection for the target received signal to obtain the target signal, including: Performing feature extraction processing on the real part of the absolute value residual by the signal detection unit to obtain an estimated value of the real part of the absolute value residual; Performing feature extraction processing on the imaginary part of the absolute value residual by the signal detection unit to obtain an estimated value of the imaginary part of the absolute value residual; Performing activation processing on the real part estimated value of the absolute value residual and the imaginary part estimated value of the absolute value residual to obtain a real part activation absolute estimated value and an imaginary part activation absolute estimated value; Determine the real part of the target signal by using the real part of the residual, the positive and negative signed real part of the residual, and the real part activation absolute estimate; Determine the imaginary part of the target signal by using the imaginary part of the residual, the positive and negative sign imaginary part of the residual, and the imaginary part activation absolute estimate; The target signal is determined according to the real part of the target signal and the imaginary part of the target signal.

8. An electronic device, characterized in that: include: Memory; as well as A processor coupled to the memory, wherein the processor is configured to execute the signal detection method according to any one of claims 1 to 6 based on instructions stored in the memory.

9. A computer-readable storage medium having a program stored thereon, wherein when the program is executed by a processor, the signal detection method according to any one of claims 1 to 6 is implemented.

Images