Multi-path wireless communication method based on AI model intelligent matching waveform

Abstract

The invention relates to the technical field of multipath wireless communication, and particularly discloses a multipath wireless communication method based on AI model intelligent matching waveform, comprising the following steps: S1, a transmitting end repeatedly transmits a test signal to a receiving end, and the position of the receiving end is moved each time to obtain multipath information of different space points; the receiving end analyzes the signal, extracts the multipath structure parameters of each multipath component, and synthesizes the multipath structure parameters to obtain the channel diffusion characteristics of the channel; s2, pre-constructing a waveform primitive library with different time-frequency characteristics; intelligently selecting a group of basic waveforms from the library based on channel diffusion characteristics; calculating emission parameters of each basic waveform in combination with the multipath structure parameters; and S3, according to the calculated emission parameters, generating a plurality of waveform signals, superposing and synthesizing the waveform signals into a composite emission signal, and emitting the composite emission signal, thereby realizing adaptive matching and optimized transmission of the multipath channel.

Description

Technical Field

[0001] This invention relates to the field of multipath wireless communication technology, and more specifically to a multipath wireless communication method based on AI model-based intelligent waveform matching. Background Technology

[0002] In current wireless communication, especially in beyond-line-of-sight tropospheric scattering communication, signal transmission heavily relies on randomly distributed and dynamically changing scatterers in the atmosphere, leading to significant multipath effects. Traditional techniques typically treat multipath as a harmful source of interference, primarily employing channel equalization, diversity reception, and adaptive modulation and coding to counteract and compensate for it. However, these methods are essentially passive, reactive strategies with limited adaptive capabilities. They usually adjust parameters based on preset thresholds or simplified channel models, making it difficult to accurately track and match the complex and rapidly changing time-frequency dual-selective fading characteristics of the channel.

[0003] Tropospheric scattering channels exhibit large time delay spread and Doppler spread, causing fixed or finitely adjustable waveforms to suffer severe inter-symbol interference and frequency-selective fading during transmission, resulting in reduced system capacity and increased bit error rate. Although existing adaptive techniques can switch between a limited number of modulation and coding schemes based on channel state information, they fail to perform in-depth optimization of the waveform's time-frequency structure and cannot achieve shape matching between the transmitted signal and the channel scattering characteristics. Summary of the Invention

[0004] The purpose of this invention is to provide a multipath wireless communication method based on AI model intelligent waveform matching, and to solve the following technical problems.

[0005] The objective of this invention can be achieved through the following technical solutions: A multipath wireless communication method based on AI model-based intelligent waveform matching includes the following steps: Step S1: A test signal is preset, and the transmitting point transmits the test signal to the receiving end. The receiving end receives and parses the test signal, and repeats the transmission several times, moving the position of the receiving end each time it is transmitted. Finally, the receiving end obtains the multipath structure parameters of each multipath component, and obtains the channel diffusion characteristics of the channel based on all multipath structure parameters. Step S2: Set several candidate primitive waveforms with different inherent time-frequency parameters to form a waveform primitive library; select several candidate primitive waveforms from the waveform primitive library as basic waveform functions according to the channel diffusion characteristics; obtain the transmission parameters of each basic waveform function according to the multipath structure parameters of each multipath component, wherein the transmission parameters include start time, carrier frequency offset and complex weighting coefficient; Step S3: The transmitting end generates several waveform signals according to the transmission parameters of each basic waveform function, and superimposes the waveform signals to generate a composite transmission signal, and then transmits the composite transmission signal.

[0006] As a further aspect of the present invention: the process of the receiving end parsing the test signal includes: The receiving end performs timing synchronization and carrier synchronization on the test signal, and downconverts the test signal into a baseband signal; based on the known test signal, it performs matched filtering on the baseband signal to obtain the time-domain impulse response of the channel, and performs peak detection on the time-domain impulse response, recording several peak points whose energy exceeds a preset threshold as multipath components, and recording the arrival time of each multipath component; for any multipath component, it records the multipath structure parameters of the multipath component.

[0007] As a further aspect of the present invention: the process of obtaining the channel diffusion characteristics of the channel includes: The multipath structure parameters include the relative delay, complex channel coefficient, and Doppler frequency shift of each multipath component; the relative delay of each multipath component is obtained, the average value of each relative delay is obtained, and the variance of each relative delay is obtained based on the average value, which is denoted as the time-domain spread parameter. For any multipath component, the Doppler frequency shift of the multipath component is treated as a discrete frequency point, and the square of the magnitude of the complex channel coefficient of the multipath component is obtained, denoted as the initial power of the frequency point. The frequency points of all discrete multipath components are smoothed, and all frequency points are transformed into a continuous power spectral density function S(f) based on an interpolation method, where f represents the Doppler frequency. The normalized first moment of the power spectral density function is obtained, denoted as the average Doppler frequency shift. And obtain the normalized second-order central moment of the power spectral density function, denoted as the frequency domain extension parameter. The time-domain spread parameter and the frequency-domain spread parameter are encapsulated together into a two-dimensional channel diffusion feature, denoted as the channel diffusion feature.

[0008] As a further aspect of the present invention: the process of setting several candidate primitive waveforms with different inherent time-frequency parameters includes: The inherent time-frequency parameters include the element duration, element bandwidth, and element center frequency. The element duration, element bandwidth, and element center frequency are all denoted as parameter dimensions. For any parameter dimension, the value range of the parameter dimension is obtained. Within this value range, discretization sampling is performed based on a preset interval threshold to obtain several sampled values. All sampled values ​​for each parameter dimension are obtained, and the sampled values ​​on each dimension are arranged and combined to obtain several parameter combinations. For any parameter combination, a waveform discrete sequence corresponding to the parameter combination is generated according to a predefined waveform mathematical model, and this sequence is denoted as a candidate element waveform.

[0009] As a further aspect of the present invention: the process of selecting several candidate primitive waveforms from the waveform primitive library as basic waveform functions includes: Based on the channel diffusion characteristics, the channel delay spread range [τ] is analyzed. min , τ max Doppler extension range [fd] min ,fd max and channel operating frequency band [F low F high ], where τ min τ represents the minimum time delay. max Indicates the maximum delay, fd min Denotes the minimum Doppler extension, fd max F represents the maximum Doppler extension. low Indicates the maximum and minimum communication frequency band, F high Indicates the highest communication frequency band; obtains the primitive time width T of each candidate primitive waveform in the waveform primitive library. p Element bandwidth B p and the center frequency f of the elementary element cp If there are alternative primitive waveforms that satisfy Then, the candidate primitive waveform is recorded as the adapted primitive waveform; and the overall matching degree of the adapted primitive waveform is obtained. K1, K2, and K3 are preset weights, and W is the anti-interference coefficient of the primitives obtained from the waveform primitive library. The waveforms of each adaptive primitive are sorted according to the magnitude of the comprehensive matching degree, and the top n adaptive primitive waveforms are selected as the basic waveform functions, where n is the preset number.

[0010] As a further aspect of the present invention: the process of resolving the channel's delay spread range, Doppler spread range, and operating frequency band based on the channel diffusion characteristics includes: Based on the multipath structure parameters of each multipath component, the maximum relative time delay of each multipath component is obtained, denoted as τ. max And obtain the minimum relative time delay in each multipath component, denoted as τ. min The time delay extension range [τ] is obtained. min , τ max Based on the multipath structure parameters of each multipath component, obtain the minimum value fd of the Doppler frequency shift in each multipath component. min And obtain the maximum value fd of the Doppler frequency shift in each multipath component. max The Doppler extension range [fd] is obtained. min ,fd max ]; and based on the preset communication frequency band, determine the frequency range, denoted as the channel operating frequency band [F];low F high ].

[0011] As a further aspect of the present invention: the process of obtaining the transmission parameters of the basic waveform function includes: For any fundamental waveform function, the initial time t of the fundamental waveform function start =τ max -τ ref , where τ ref The reference delay is the average relative delay of all multipath components; the carrier frequency offset f of the fundamental waveform function. w =-fd avg +f corr , where fd avg f is the average value of the Doppler frequency shift of all multipath components. corr The preset correction amount; the complex weighting coefficients of the basic waveform function w = A × e （jΦ） Where A is the amplitude, and , where h w h is the average channel gain of all multipath components. all Let Φ be the sum of the channel gains of all multipath components, Φ be the phase, and Φ = -2π(f cp +f w )×τ ref .

[0012] As a further aspect of the present invention: the waveform signal generation process includes: For any basic waveform function, the actual transmission carrier frequency of the basic waveform function is obtained based on the carrier frequency offset and the center frequency of the basic element. The basic waveform function is then modulated to the actual transmission carrier frequency to obtain a carrier modulation signal. The amplitude and phase of the carrier modulation signal are adjusted according to the complex weighting coefficients of the basic waveform function to generate a calibrated waveform signal.

[0013] The beneficial effects of this invention are: Traditional methods often face severe inter-symbol interference and frequency-selective fading in complex multipath environments, leading to increased bit error rate and decreased effective throughput. This invention intelligently analyzes channel delay and Doppler spread characteristics using an AI model, and dynamically selects or synthesizes transmit waveforms with highly matched time-frequency characteristics from a waveform library. This allows signal energy to pass through the advantageous path in the channel more concentratedly and efficiently, mitigating the negative effects of multipath, significantly reducing the bit error rate, improving link reliability, and achieving higher data transmission rates and spectral efficiency under the same bandwidth and power conditions. Furthermore, by assigning appropriate waveform components to different scattering paths and performing time-frequency domain alignment and synthesis, this invention transforms the natural multipath structure into a usable path diversity resource. This proactive signal design enables constructive superposition of the synthesized signal at the receiver, enhancing the strength of the main signal, while also providing natural spatial and temporal filtering effects against interference and noise from mismatched directions. Combined with the small beam and low elevation angle transmission characteristics, this further reduces the probability of signal interception and interference, improving survivability in electronic warfare environments. Attached Figure Description

[0014] The invention will now be further described with reference to the accompanying drawings.

[0015] Figure 1 This is a schematic diagram illustrating the steps of the multipath wireless communication method based on AI model intelligent waveform matching according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Please see Figure 1 As shown, this invention is a multipath wireless communication method based on AI model intelligent waveform matching, comprising the following steps: Step S1: A test signal is preset, and the transmitting point transmits the test signal to the receiving end. The receiving end receives and parses the test signal, and repeats the transmission several times, moving the position of the receiving end each time it is transmitted. Finally, the receiving end obtains the multipath structure parameters of each multipath component, and obtains the channel diffusion characteristics of the channel based on all multipath structure parameters. It is worth noting that the purpose of repeatedly transmitting the test signal and moving the receiver's position each time is: By changing the position of the receiver in space, the test signal is actively induced to propagate through multiple differentiated paths formed by different combinations of scatterers in space, thereby collecting a richer multipath structure sample reflecting the spatial scattering characteristics of the communication environment in a single measurement process than a single-point static measurement. In a preferred embodiment of the present invention, the process of the receiving end parsing the test signal includes: The receiving end performs timing synchronization and carrier synchronization on the test signal, and downconverts the test signal into a baseband signal; based on the known test signal, it performs matched filtering on the baseband signal to obtain the time-domain impulse response of the channel, and performs peak detection on the time-domain impulse response, recording several peak points whose energy exceeds a preset threshold as multipath components, and recording the arrival time of each multipath component; for any multipath component, it records the multipath structure parameters of the multipath component; In a preferred embodiment of the present invention, the multipath structure parameters include the relative time delay, complex channel coefficient, and Doppler frequency shift of each multipath component; In a preferred embodiment of the present invention, the process of obtaining the channel diffusion characteristics of the channel includes: Obtain the relative delay of each multipath component, obtain the average value of each relative delay, and obtain the variance of each relative delay based on the average value, which is denoted as the time domain spread parameter. For any multipath component, the Doppler frequency shift of the multipath component is treated as a discrete frequency point, and the square of the magnitude of the complex channel coefficient of the multipath component is obtained, denoted as the initial power of the frequency point. The frequency points of all discrete multipath components are smoothed, and all frequency points are transformed into a continuous power spectral density function S(f) based on an interpolation method, where f represents the Doppler frequency. The normalized first moment of the power spectral density function is obtained, denoted as the average Doppler frequency shift. And obtain the normalized second-order central moment of the power spectral density function, denoted as the frequency domain extension parameter. The time-domain spread parameter and the frequency-domain spread parameter are encapsulated together into a two-dimensional channel spread feature, denoted as the channel spread feature. Step S2: Set several candidate primitive waveforms with different inherent time-frequency parameters to form a waveform primitive library; select several candidate primitive waveforms from the waveform primitive library as basic waveform functions according to the channel diffusion characteristics; obtain the transmission parameters of each basic waveform function according to the multipath structure parameters of each multipath component, wherein the transmission parameters include start time, carrier frequency offset and complex weighting coefficient; In a preferred embodiment of the present invention, the process of setting several candidate primitive waveforms with different inherent time-frequency parameters includes: The inherent time-frequency parameters include the element duration, element bandwidth, and element center frequency. The element duration, element bandwidth, and element center frequency are all denoted as parameter dimensions. For any parameter dimension, the value range of the parameter dimension is obtained. Within this value range, discretization sampling is performed based on a preset interval threshold to obtain several sampled values. All sampled values ​​for each parameter dimension are obtained, and the sampled values ​​on each dimension are arranged and combined to obtain several parameter combinations. For any parameter combination, a waveform discrete sequence corresponding to the parameter combination is generated according to a predefined waveform mathematical model, and this sequence is denoted as a candidate element waveform. It should be noted that the primitive time width represents the time length of a single primitive waveform, in μs; the primitive bandwidth represents the frequency range of a single primitive waveform, in MHz; and the primitive center frequency represents the frequency center point of the primitive waveform, in GHz. In a preferred embodiment of the present invention, the process of selecting several candidate primitive waveforms from the waveform primitive library as basic waveform functions includes: Based on the channel diffusion characteristics, the channel delay spread range [τ] is analyzed. min , τ max Doppler extension range [fd] min ,fd max and channel operating frequency band [F low F high ], where τ min τ represents the minimum time delay. max Indicates the maximum delay, fd min Denotes the minimum Doppler extension, fd max F represents the maximum Doppler extension. low Indicates the maximum and minimum communication frequency band, F high Indicates the highest communication frequency band; obtains the primitive time width T of each candidate primitive waveform in the waveform primitive library. p Element bandwidth B p and the center frequency f of the elementary element cp If there are alternative primitive waveforms that satisfy Then, the candidate primitive waveform is recorded as the adapted primitive waveform; and the overall matching degree of the adapted primitive waveform is obtained. K1, K2 and K3 are preset weights, and W is the anti-interference coefficient of the primitives obtained from the waveform primitive library. The waveforms of each adaptive primitive are sorted according to the size of the comprehensive matching degree, and the top n adaptive primitive waveforms are selected as the basic waveform functions, where n is the preset number. Specifically, under normal circumstances, K1=0.4, K2=0.3, K3=0.3. The anti-interference coefficient W of the primitive is extracted from the waveform primitive library. It is the anti-interference capability parameter of each primitive. The larger the value, the stronger the primitive's ability to resist multipath interference. W∈[0,1]. Based on the aforementioned channel diffusion characteristics, the process of resolving the channel's delay spread range, Doppler spread range, and operating frequency band includes: Based on the multipath structure parameters of each multipath component, the maximum relative time delay of each multipath component is obtained, denoted as τ. max And obtain the minimum relative time delay in each multipath component, denoted as τ. min The time delay extension range [τ] is obtained. min , τ max Based on the multipath structure parameters of each multipath component, obtain the minimum value fd of the Doppler frequency shift in each multipath component. min And obtain the maximum value fd of the Doppler frequency shift in each multipath component. max The Doppler extension range [fd] is obtained. min ,fd max ]; and based on the preset communication frequency band, determine the frequency range, denoted as the channel operating frequency band [F]; low F high ]; In a preferred embodiment of the present invention, the process of obtaining the emission parameters of the basic waveform function includes: For any fundamental waveform function, the initial time t of the fundamental waveform function start =τ max -τ ref , where τ ref The reference delay is the average relative delay of all multipath components; the carrier frequency offset f of the fundamental waveform function. w =-fd avg +f corr , where fd avg f is the average value of the Doppler frequency shift of all multipath components. corr The preset correction amount; the complex weighting coefficients of the basic waveform function w = A × e （jΦ） Where A is the amplitude, and , where h w h is the average channel gain of all multipath components. all Let Φ be the sum of the channel gains of all multipath components, Φ be the phase, and Φ = -2π(f cp +f w )×τ ref ; Specifically, for each selected basic waveform function, its transmission start time is calculated based on the relative delay in the multipath structure parameters. This start time is set with the goal of ensuring time-domain alignment between the waveform signal and other waveform signals at the receiving end after transmission through the current multipath channel. Based on the Doppler shift information in the multipath structure parameters, its carrier frequency offset is calculated. This offset is used to pre-compensate for the Doppler shift generated during transmission of the waveform signal in the current multipath channel, thereby achieving frequency-domain alignment of each waveform signal at the receiving end. Based on the path gain and delay information in the current multipath structure parameters, a complex weighting coefficient is calculated. This complex weighting coefficient is used to adjust the transmission amplitude and initial phase of the waveform signal, so that all waveform signals can achieve coherent phase superposition and optimized energy synthesis at the receiving end. The process of obtaining the channel gain of the multipath components includes: The impulse response of the channel is estimated using the least squares method or the least mean square error method, resulting in a discrete response sequence characterizing the relationship between the arrival time and amplitude of each multipath component. Effective peak values ​​are extracted from the discrete response sequence, with the time coordinate corresponding to each peak value being the arrival time of a single multipath component, and the peak amplitude being the actual received amplitude of that multipath component at the receiver. The channel gain of the multipath component is then given by g. i =Arx i / Atx, where Arx i The received amplitude of the multipath component is denoted as Δx, and Atx is the known transmitted amplitude of the test signal at the transmitter. It is worth noting that a preset signal-to-noise ratio (SNR) threshold is set, the SNR of each multipath component is calculated, the linear domain channel gain corresponding to the multipath component with an SNR not lower than the threshold is retained, and invalid multipath gain data dominated by noise is removed. Step S3: The transmitting end generates several waveform signals according to the transmission parameters of each basic waveform function, and superimposes the waveform signals to generate a composite transmission signal, and then transmits the composite transmission signal. In a preferred embodiment of the present invention, the waveform signal generation process includes: For any basic waveform function, the actual transmission carrier frequency of the basic waveform function is obtained based on the carrier frequency offset and the center frequency of the basic element. The basic waveform function is then modulated to the actual transmission carrier frequency to obtain a carrier modulation signal. The amplitude and phase of the carrier modulation signal are adjusted according to the complex weighting coefficients of the basic waveform function to generate a calibrated waveform signal. In a preferred embodiment of the present invention, the process of transmitting the composite transmission signal includes: The waveform signals corresponding to each basic waveform function are linearly superimposed in the time domain to obtain a composite transmission signal; the transmitting end transmits the composite transmission signal according to the average start time of each basic waveform function.

[0018] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A multipath wireless communication method based on AI model-based intelligent waveform matching, characterized in that, Includes the following steps: Step S1: Preset a test signal, transmit the test signal from the transmitting point to the receiving end, the receiving end receives and parses the test signal, and repeats the transmission several times, moving the position of the receiving end each time it is transmitted; Finally, the receiving end obtains the multipath structure parameters of each multipath component, and obtains the channel diffusion characteristics of the channel based on all multipath structure parameters; Step S2: Set several candidate primitive waveforms with different inherent time-frequency parameters to form a waveform primitive library; select several candidate primitive waveforms from the waveform primitive library as basic waveform functions according to the channel diffusion characteristics. Based on the multipath structure parameters of each multipath component, the transmission parameters of each basic waveform function are obtained. The transmission parameters include the start time, carrier frequency offset, and complex weighting coefficient. Step S3: The transmitting end generates several waveform signals according to the transmission parameters of each basic waveform function, and superimposes the waveform signals to generate a composite transmission signal, and then transmits the composite transmission signal.

2. The multipath wireless communication method based on AI model intelligent waveform matching according to claim 1, characterized in that, In step S1, the process of the receiving end parsing the test signal includes: The receiving end performs timing synchronization and carrier synchronization on the test signal, and downconverts the test signal into a baseband signal; based on the known test signal, it performs matched filtering on the baseband signal to obtain the time-domain impulse response of the channel, and performs peak detection on the time-domain impulse response, recording several peak points whose energy exceeds a preset threshold as multipath components, and recording the arrival time of each multipath component; for any multipath component, it records the multipath structure parameters of the multipath component.

3. The multipath wireless communication method based on AI model intelligent waveform matching according to claim 1, characterized in that, In step S1, the process of obtaining the channel diffusion characteristics of the channel includes: The multipath structure parameters include the relative delay, complex channel coefficient, and Doppler frequency shift of each multipath component; the relative delay of each multipath component is obtained, the average value of each relative delay is obtained, and the variance of each relative delay is obtained based on the average value, which is denoted as the time-domain spread parameter. For any multipath component, the Doppler frequency shift of the multipath component is treated as a discrete frequency point, and the square of the magnitude of the complex channel coefficient of the multipath component is obtained, denoted as the initial power of the frequency point. The frequency points of all discrete multipath components are smoothed, and all frequency points are transformed into a continuous power spectral density function S(f) based on an interpolation method, where f represents the Doppler frequency. The normalized first moment of the power spectral density function is obtained, denoted as the average Doppler frequency shift. And obtain the normalized second-order central moment of the power spectral density function, denoted as the frequency domain extension parameter. The time-domain spread parameter and the frequency-domain spread parameter are encapsulated together into a two-dimensional channel diffusion feature, denoted as the channel diffusion feature.

4. The multipath wireless communication method based on AI model intelligent waveform matching according to claim 1, characterized in that, In step S2, the process of setting several candidate elementary waveforms with different inherent time-frequency parameters includes: The inherent time-frequency parameters include the element duration, element bandwidth, and element center frequency. The element duration, element bandwidth, and element center frequency are all denoted as parameter dimensions. For any parameter dimension, the value range of the parameter dimension is obtained. Within this value range, discretization sampling is performed based on a preset interval threshold to obtain several sampled values. All sampled values ​​for each parameter dimension are obtained, and the sampled values ​​on each dimension are arranged and combined to obtain several parameter combinations. For any parameter combination, a waveform discrete sequence corresponding to the parameter combination is generated according to a predefined waveform mathematical model, and this sequence is denoted as a candidate element waveform.

5. The multipath wireless communication method based on AI model intelligent waveform matching according to claim 1, characterized in that, In step S2, the process of selecting several candidate primitive waveforms from the waveform primitive library as basic waveform functions includes: Based on the channel diffusion characteristics, the channel delay spread range [τ] is analyzed. min , τ max Doppler extension range [fd] min ,fd max and channel operating frequency band [F low F high ], where τ min τ represents the minimum time delay. max Indicates the maximum delay, fd min Denotes the minimum Doppler extension, fd max F represents the maximum Doppler extension. low Indicates the maximum and minimum communication frequency band, F high Indicates the highest communication frequency band; obtains the primitive time width T of each candidate primitive waveform in the waveform primitive library. p Element bandwidth B p and the center frequency f of the elementary element cp If there are alternative primitive waveforms that satisfy Then, the candidate primitive waveform is recorded as the adapted primitive waveform; and the overall matching degree of the adapted primitive waveform is obtained. K1, K2, and K3 are preset weights, and W is the anti-interference coefficient of the primitives obtained from the waveform primitive library. The waveforms of each adaptive primitive are sorted according to the magnitude of the comprehensive matching degree, and the top n adaptive primitive waveforms are selected as the basic waveform functions, where n is the preset number.

6. The multipath wireless communication method based on AI model intelligent waveform matching according to claim 5, characterized in that, In step S2, the process of resolving the channel's delay spread range, Doppler spread range, and operating frequency band based on the channel diffusion characteristics includes: Based on the multipath structure parameters of each multipath component, the maximum relative time delay of each multipath component is obtained, denoted as τ. max And obtain the minimum relative time delay in each multipath component, denoted as τ. min The time delay range [τ] is obtained. min , τ max Based on the multipath structure parameters of each multipath component, obtain the minimum value fd of the Doppler frequency shift in each multipath component. min And obtain the maximum value fd of the Doppler frequency shift in each multipath component. max The Doppler extension range [fd] is obtained. min ,fd max ]; and based on the preset communication frequency band, determine the frequency range, denoted as the channel operating frequency band [F]; low F high ].

7. The multipath wireless communication method based on AI model intelligent waveform matching according to claim 1, characterized in that, In step S2, the process of obtaining the emission parameters of the basic waveform function includes: For any fundamental waveform function, the initial time t of the fundamental waveform function start =τ max -τ ref , where τ ref The reference delay is the average relative delay of all multipath components; the carrier frequency offset f of the fundamental waveform function. w =-fd avg +f corr , where fd avg f is the average value of the Doppler frequency shift of all multipath components. corr The preset correction amount; the complex weighting coefficients of the basic waveform function w = A × e （jΦ） Where A is the amplitude, and , where h w h is the average channel gain of all multipath components. all Let Φ be the sum of the channel gains of all multipath components, Φ be the phase, and Φ = -2π(f cp +f w )×τ ref .

8. The multipath wireless communication method based on AI model intelligent waveform matching according to claim 1, characterized in that, In step S2, the waveform signal generation process includes: For any basic waveform function, the actual transmission carrier frequency of the basic waveform function is obtained based on the carrier frequency offset and the center frequency of the basic element. The basic waveform function is then modulated to the actual transmission carrier frequency to obtain a carrier modulation signal. The amplitude and phase of the carrier modulation signal are adjusted according to the complex weighting coefficients of the basic waveform function to generate a calibrated waveform signal.

Images