A system and method for dynamic error compensation of a millimeter wave transmission link

Abstract

This invention discloses a dynamic error compensation system and method for millimeter-wave transmission links, relating to the field of millimeter-wave wireless communication transmission technology. It includes a millimeter-wave signal transceiver module operating between 24 GHz and 100 GHz, a multi-dimensional error sensing module employing a synchronous parallel sampling architecture to synchronously detect phase, amplitude, group delay, and frequency offset coupling errors, a digital predistortion compensation module implemented on an FPGA, and an adaptive parameter adjustment module that dynamically adjusts compensation parameters using a dual-driven mechanism of error and environment. Combined with multi-modal environment sensing, it achieves active pre-compensation. Time-division multiplexing within the data frame protection interval enables uninterrupted online calibration, and a fault self-diagnosis and redundancy backup module is integrated to achieve graded fault handling. This invention improves compensation accuracy and response speed through deep multi-module collaboration, enables uninterrupted online calibration, possesses fault self-diagnosis capabilities, and can maintain stable millimeter-wave communication transmission quality over a long period.

Description

Technical Field

[0001] This invention relates to the field of millimeter-wave wireless communication transmission technology, and in particular to a dynamic error compensation system and method for millimeter-wave transmission links. Background Technology

[0002] Existing millimeter-wave transmission link error compensation systems only detect errors in a single dimension, ignoring the coupling characteristics between phase, amplitude, group delay, and frequency offset, and cannot fully capture the true error state of the link. At the same time, they generally lack environmental awareness capabilities, cannot establish a correlation mapping between environmental parameters and link errors, and can only perform passive compensation after the error occurs, making it difficult to cope with microsecond-level rapid error fluctuations caused by sudden environmental changes. Summary of the Invention

[0003] This invention proposes a dynamic error compensation system and method for millimeter-wave transmission links to solve the problems mentioned in the prior art.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: a dynamic error compensation system for millimeter-wave transmission links, comprising the following modules: The millimeter-wave transceiver module operates in the full frequency band from 24GHz to 100GHz and is used to transmit reference millimeter-wave signals and receive the return signals after transmission through the link, and to collect the original transmission data of the link. The multi-dimensional error perception module adopts a synchronous parallel sampling architecture, which is synchronized with the millimeter-wave signal transceiver module in timing. It performs synchronous joint detection of phase error, amplitude error, group delay error and frequency offset error of the transmission link, and outputs multi-dimensional coupled error detection data. The digital predistortion compensation module receives the coupling error detection data output by the multi-dimensional error perception module, generates a digital compensation signal, and outputs it to the millimeter-wave signal transceiver module to achieve signal predistortion compensation. The adaptive parameter adjustment module is used to dynamically adjust the core parameters of the predistortion compensation algorithm in the digital predistortion compensation module based on the coupling error change trend output by the multi-dimensional error perception module, combined with environmental fluctuation information and link transmission quality; it distinguishes between inherent link error and drift error based on accumulated coupling error data, completes closed-loop automatic calibration, evaluates the compensation effect based on link operation indicators, and outputs the control feedback signal in reverse to the digital predistortion compensation module and the multi-dimensional error perception module.

[0005] Furthermore, it also includes a multimodal environmental sensing module, which integrates temperature sensors, humidity sensors, vibration sensors, and electromagnetic interference sensors to synchronously collect environmental parameters around the link. It uses an adaptive weighted fusion algorithm to fuse the multi-sensor data, establishes a nonlinear correlation mapping relationship between environmental parameters and multi-dimensional coupling errors of the millimeter-wave transmission link, and outputs it to the adaptive parameter adjustment module.

[0006] Furthermore, it also includes a fault self-diagnosis and redundancy backup module, which monitors the power supply voltage, operating current, output signal quality and internal register status of the millimeter wave signal transceiver module, multi-dimensional error perception module, digital predistortion compensation module and adaptive parameter adjustment module in real time. It uses fault tree analysis to identify hardware faults and link anomalies, and classifies faults into three levels: low, medium and high, according to the degree of impact of faults on the performance of multi-dimensional coupling error compensation. In the event of a low-level fault, the adaptive parameter adjustment module is linked to adjust the predistortion compensation algorithm parameters for self-healing. In the event of a medium-level failure, the system will automatically switch to a primary and backup hot backup compensation channel. In the event of a high-level fault, an audible and visual alarm signal will be output and a fault log will be recorded.

[0007] Furthermore, the multi-dimensional error perception module includes a phase detection unit, an amplitude detection unit, a group delay detection unit, and a data preprocessing unit. The phase detection unit extracts the in-phase and quadrature components of the received signal using an orthogonal mixing method, calculates the signal phase value through arctangent operation, and compares it with the reference phase of the transmitting end to obtain the phase error. The amplitude detection unit completes the amplitude detection of the received signal using a true RMS detection method. The data preprocessing unit performs multi-domain joint noise reduction processing on the multi-dimensional coupled error data, separates the slow-varying drift error and the fast-varying burst error, and provides error input for the digital predistortion compensation module and the adaptive parameter adjustment module.

[0008] Furthermore, the digital predistortion compensation module includes a digital signal processing unit, a high-speed digital-to-analog converter unit, and an RF front-end compensation unit. The digital signal processing unit adopts a parallel pipeline architecture to simultaneously process the phase, amplitude, and time delay compensation signals. It introduces a memory polynomial predistortion structure to compensate for the nonlinear memory effect of the link, divides the millimeter-wave operating frequency band into multiple continuous sub-bands, and independently trains compensation coefficients for each sub-band to generate a digital compensation signal for compensating for multi-dimensional coupling errors. The high-speed digital-to-analog converter unit converts the digital compensation signal into an analog RF compensation signal. The RF front-end compensation unit injects the analog RF compensation signal into the transmission link through a vector modulator and integrates an adaptive impedance matching network to dynamically adjust the impedance according to changes in link load.

[0009] Furthermore, the adaptive parameter adjustment module includes a standard calibration signal source, a calibration signal receiving unit, and a calibration parameter calculation unit. The standard calibration signal source outputs a single-tone signal and a linear frequency modulated signal with controllable frequency and amplitude, and automatically adjusts the output power according to the link attenuation status. The calibration signal receiving unit collects the calibration signal after transmission through the link. The calibration parameter calculation unit compares the parameter differences between the transmitted calibration signal and the received calibration signal, calculates the inherent insertion loss, phase offset, and group delay characteristics of the link, and transmits the calibration signal within the data frame guard interval using time-division multiplexing.

[0010] Furthermore, the adaptive parameter adjustment module adopts a parameter update mechanism driven by both multi-dimensional coupling error and environment, and the parameter update formula is as follows: ;in For the first The compensation parameter weights for the next iteration; For the first The compensation parameter weights for the next iteration; The coefficient representing the influence of the rate of change of error; For the first The rate of change of normalized multidimensional coupling error in the next iteration; The environmental change rate influence coefficient; For the first The normalized comprehensive environmental change rate of the next iteration is obtained by weighted summation of the changes in temperature, humidity, vibration, and electromagnetic interference.

[0011] Furthermore, this includes the following steps: Complete the initial configuration and perform initial error calibration to establish the baseline error mapping relationship; The millimeter-wave signal transceiver module transmits a reference millimeter-wave signal and collects the received signal transmitted back via the transmission link. The multi-modal environment sensing module simultaneously collects environmental parameters around the link. The multi-dimensional error perception module analyzes the received signal, extracts the phase error, amplitude error, group delay error and frequency offset error of the transmission link, constructs multi-dimensional coupling error, and generates multi-dimensional error feature vector; The adaptive parameter adjustment module calculates the current compensation parameter weights based on the error feature vector and environmental parameters, and updates the compensation parameters of the digital predistortion compensation module based on the compensation parameter weights. The digital predistortion compensation module generates a corresponding digital compensation signal based on the updated compensation parameters, which is then injected into the transmission link after digital-to-analog conversion and radio frequency modulation to complete the dynamic compensation of multi-dimensional coupling errors. The adaptive parameter adjustment module executes a closed-loop online calibration process and updates the reference compensation parameters based on preset period or error threshold trigger conditions.

[0012] Furthermore, generating multi-dimensional error feature vectors includes: sequentially performing down-conversion and quadrature demodulation on the received signal to obtain baseband in-phase and quadrature components; calculating the average phase and average amplitude of multiple consecutive sampling points through a sliding window of preset length, and comparing them with the phase and amplitude of the transmitted reference signal to obtain phase error and amplitude error; obtaining group delay error by calculating the peak position of the cross-correlation function of the transmitted and received signals; extracting the carrier frequency of the received signal based on a phase-locked loop and comparing it with the transmitted carrier frequency to obtain frequency offset error; calculating the correlation coefficient between different error components, identifying multi-dimensional error coupling relationships, and processing each error component using a piecewise normalization method.

[0013] Furthermore, the closed-loop online calibration process includes: triggering online calibration when the system's continuous running time reaches a preset calibration cycle or when the link bit error rate exceeds a set threshold; transmitting calibration signals from the standard calibration signal source of the adaptive parameter adjustment module within the data frame guard interval, and collecting the calibration signals after link transmission by the calibration signal receiving unit of the adaptive parameter adjustment module; calculating the difference between the current error characteristics of the link and the reference error characteristics to generate calibration correction coefficients; using an incremental update method to adjust only compensation parameters whose changes exceed a set threshold; verifying the calibration effectiveness by comparing the link signal-to-noise ratio before and after calibration; restoring normal data transmission after calibration and storing calibration data and correction coefficients.

[0014] Compared with existing technologies, the beneficial effects of this invention are: This invention solves the problems of low compensation accuracy, slow response, and poor continuity by constructing an integrated dynamic error compensation system with deep multi-module collaboration, which organically integrates error perception, environmental prediction, parameter adjustment, real-time compensation, online calibration, and fault protection.

[0015] The multi-dimensional error perception module and the multi-modal environment perception module work together to achieve a fusion of passive detection and proactive prediction. The multi-dimensional error perception module adopts a synchronous parallel sampling architecture to simultaneously detect four types of coupling errors: phase, amplitude, group delay, and frequency offset. It separates slowly changing drift errors and rapidly changing burst errors through multi-domain joint noise reduction. The multi-modal environment perception module integrates multi-source environmental data to establish a nonlinear correlation mapping between environmental parameters and multi-dimensional coupling errors. The collaboration between the two not only comprehensively and accurately captures the current error state of the link but also predicts error change trends in advance. This transforms traditional post-event passive compensation into a combination of pre-event proactive compensation and real-time compensation, significantly improving the system's response speed to sudden environmental changes.

[0016] The adaptive parameter adjustment module and the digital predistortion compensation module work together to achieve dynamic parameter adaptation and accurate compensation across the entire frequency band. The adaptive parameter adjustment module employs a parameter update mechanism driven by both error and environment, dynamically adjusting the weights and update step sizes of the compensation parameters. The digital predistortion compensation module uses a parallel pipeline architecture, independently training compensation coefficients and a memory polynomial structure for each frequency band to compensate for the nonlinear memory effect of the link. The collaboration between the two ensures both rapid convergence when errors change quickly and fine-tuning when the environment is stable. Furthermore, combined with an adaptive impedance matching network, it achieves accurate compensation for multi-dimensional coupling errors across the entire frequency band from 24 GHz to 100 GHz.

[0017] The adaptive parameter adjustment module and the digital predistortion compensation module work in a closed-loop manner to achieve continuous optimization without interrupting service. The calibration parameter calculation unit of the adaptive parameter adjustment module transmits calibration signals within the data frame guard interval, without occupying service bandwidth or interrupting transmission. After calibration, the compensation parameters are adjusted through incremental updates, synchronously updating the lookup table and reference error database of the digital predistortion module, forming a complete closed-loop optimization system to continuously adapt to device aging and link characteristic drift.

[0018] The fault self-diagnosis and redundancy backup module works in synergy with the entire system to achieve tiered fault handling and continuous operation assurance. This module monitors the status of all core modules in real time, uses fault tree analysis to handle faults in a tiered manner, and links parameters for self-healing, hot backup switching, and alarm mechanisms to prevent system paralysis caused by a single module failure. Attached Figure Description

[0019] Figure 1 This is a schematic block diagram of the overall flowchart of the millimeter-wave transmission link dynamic error compensation system proposed in this invention; Figure 2 This is a schematic block diagram of the system physical and functional module composition architecture proposed in this invention; Figure 3 This is the logic diagram of the multi-dimensional error perception and preprocessing proposed in this invention; Figure 4 This is the execution path diagram for the digital predistortion compensation proposed in this invention; Figure 5 This is the closed-loop logic diagram for online calibration and fault self-healing proposed in this invention. Detailed Implementation

[0020] 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.

[0021] Reference Figures 1 to 5 A dynamic error compensation system for millimeter-wave transmission links includes the following modules: The millimeter-wave transceiver module operates in the full frequency band from 24GHz to 100GHz and is used to transmit reference millimeter-wave signals and receive the return signals after transmission through the link, and to collect the original transmission data of the link. The multi-dimensional error perception module adopts a synchronous parallel sampling architecture, which is synchronized with the millimeter-wave signal transceiver module to synchronously and jointly detect the phase error, amplitude error, group delay error and frequency offset error of the transmission link, and outputs multi-dimensional coupled error detection data. The digital predistortion compensation module receives the coupling error detection data output by the multi-dimensional error perception module, generates a digital compensation signal, and outputs it to the millimeter-wave signal transceiver module to achieve signal predistortion compensation. The adaptive parameter adjustment module is used to dynamically adjust the core parameters of the predistortion compensation algorithm in the digital predistortion compensation module based on the coupling error change trend output by the multi-dimensional error perception module, combined with environmental fluctuation information and link transmission quality. It distinguishes between inherent link error and drift error based on accumulated coupling error data, completes closed-loop automatic calibration, evaluates the compensation effect based on link operation indicators, and outputs the control feedback signal in reverse to the digital predistortion compensation module and the multi-dimensional error perception module.

[0022] For example, the millimeter-wave signal transceiver module can support full-duplex transmission mode, with a sampling rate of not less than 2GSa / s and an analog-to-digital conversion accuracy of not less than 12 bits; the phase detection accuracy of the multi-dimensional error sensing module can be not less than 0.05 degrees, the amplitude detection accuracy can be not less than 0.05dB, and the delay detection accuracy can be not less than 0.5ps; the digital predistortion compensation module can be implemented based on a field-programmable gate array architecture, with a processing delay of not more than 50ns.

[0023] In this invention, the millimeter-wave transmission link dynamic error compensation system also includes a multimodal environmental sensing module, which integrates temperature sensors, humidity sensors, vibration sensors, and electromagnetic interference sensors. The sampling frequency can be no less than 10Hz, synchronously collecting environmental parameters around the link. An adaptive weighted fusion algorithm is used to fuse the multi-sensor data, establish a nonlinear correlation mapping relationship between environmental parameters and multi-dimensional coupling errors of the millimeter-wave transmission link, and output it to the adaptive parameter adjustment module to train an environmental error time-series prediction model. This model can predict the link error change trend in the next 10 seconds, providing prior environmental information and prediction support for dynamic compensation.

[0024] In this invention, the millimeter-wave transmission link dynamic error compensation system also includes a fault self-diagnosis and redundancy backup module. This module monitors in real time the power supply voltage, operating current, output signal quality, and internal register status of the millimeter-wave signal transceiver module, multi-dimensional error sensing module, digital predistortion compensation module, and adaptive parameter adjustment module. It uses fault tree analysis to identify hardware faults and link anomalies, and classifies faults into low, medium, and high levels based on their impact on the multi-dimensional coupling error compensation performance. In the event of a low-level fault, the adaptive parameter adjustment module is linked to adjust the predistortion compensation algorithm parameters for self-healing. In the event of a medium-level failure, the system will automatically switch to a primary and backup hot backup compensation channel. In the event of a high-level fault, an audible and visual alarm signal will be output and a fault log will be recorded. Establish a fault knowledge base to enable automatic analysis of fault root causes and ensure continuous and stable system operation.

[0025] The fault level determination threshold can be preset and configured according to the service requirements of different millimeter-wave transmission scenarios. For example, a low-level fault is determined when the following conditions are met simultaneously: the power supply voltage and operating current of each module fluctuate within ±10% of the rated range; the link bit error rate is lower than a preset service threshold (e.g., 1×10⁻⁶). -6 The multi-dimensional error compensation accuracy decreases by no more than 10% from the initial calibration value, and the system can still recover performance by adjusting the predistortion compensation algorithm parameters; no core module register abnormalities or signal link interruptions occur. Low-level faults are self-healed by adjusting the compensation algorithm parameters through the linkage adaptive parameter adjustment module, without interrupting services or switching backup channels. A high-level fault is determined when any of the following occurs: power supply failure of the core modules (millimeter-wave signal transceiver module, digital predistortion compensation module) or operating current exceeding the hardware protection threshold; link bit error rate higher than 1×10. -4 The system is considered a medium-level fault if any of the following conditions occur, and the link signal-to-noise ratio drops by more than 15dB and cannot be recovered through algorithm adjustment; the main compensation channel fails, and the error compensation accuracy is lower than 70% of the design specification, failing to meet service transmission requirements; or anomalies such as critical register errors or signal link interruptions cause the compensation function to completely fail. When a high-level fault is triggered, the system immediately outputs an audible and visual alarm signal and records the fault log, while simultaneously uploading the fault information to the network management platform. A medium-level fault is determined when any of the following conditions occur, but the high-level fault criteria are not met: the power supply voltage / current of a single module exceeds the rated range by ±10% but hardware protection is not triggered; the link bit error rate is within the preset service threshold and 1×102... -4The accuracy of multi-dimensional error compensation decreases by more than 10% compared to the initial calibration value, but is still higher than 70% of the design target. The failure of a single non-core module (such as a specific environmental sensor or auxiliary detection unit) does not affect the operation of the main compensation channel. When a medium-level fault is triggered, the system automatically switches to a primary-standby hot backup compensation channel within 1ms to ensure continuous business operation.

[0026] In this invention, the multi-dimensional error perception module includes a phase detection unit, an amplitude detection unit, a group delay detection unit, and a data preprocessing unit. The phase detection unit uses orthogonal mixing to extract the in-phase and quadrature components of the received signal, calculates the signal phase value through high-precision arctangent operation, and compares it with the reference phase of the transmitting end to obtain the phase error. The amplitude detection unit uses true RMS detection to detect the amplitude of the received signal, for example, to achieve a wide dynamic range of received signal amplitude detection from -60dBm to +10dBm. The group delay detection unit uses linear frequency modulation pulse compression technology to improve the delay measurement resolution. The data preprocessing unit performs multi-domain joint noise reduction processing on the multi-dimensional coupling error data, for example, sequentially performing Kalman filtering, moving average, and wavelet denoising processing on the multi-dimensional coupling error data, extracting the trend term and fluctuation term of the error, separating the slow-varying drift error and the fast-varying burst error, providing error input for the digital predistortion compensation module and the adaptive parameter adjustment module, and eliminating the influence of random noise and burst interference.

[0027] In this invention, the digital predistortion compensation module includes a digital signal processing unit, a high-speed digital-to-analog converter (DAC), and a radio frequency (RF) front-end compensation unit. The digital signal processing unit adopts a parallel pipeline architecture to synchronously process the phase, amplitude, and time delay compensation signals. It introduces a memory polynomial predistortion structure to compensate for the nonlinear memory effect of the link, dividing the millimeter-wave operating frequency band into multiple continuous sub-bands (for example, the millimeter-wave operating frequency band can be divided into 8 continuous sub-bands). Each sub-band independently trains compensation coefficients to generate a digital compensation signal for compensating for multi-dimensional coupling errors. The high-speed DAC can have a conversion rate of no less than 5 GSa / s, converting the digital compensation signal into an analog RF compensation signal. The RF front-end compensation unit injects the analog RF compensation signal into the transmission link through a vector modulator and integrates an adaptive impedance matching network to dynamically adjust the impedance according to changes in link load. The phase compensation range is 0 to 360 degrees, the amplitude compensation range is -10 dB to +10 dB, and the time delay compensation range is 0 to 100 ns. The compensation bandwidth covers the entire operating frequency band of the system.

[0028] In this invention, the adaptive parameter adjustment module includes a standard calibration signal source, a calibration signal receiving unit, and a calibration parameter calculation unit. The standard calibration signal source outputs a single-tone signal and a linear frequency modulated signal with controllable frequency, amplitude, and precise phase control. It supports adaptive adjustment of calibration signal power and automatically adjusts the output power according to the link attenuation status. The calibration signal receiving unit collects the calibration signal after transmission through the link. The calibration parameter calculation unit compares the parameter differences between the transmitted and received calibration signals, calculates the inherent insertion loss, phase offset, and group delay characteristics of the link, and transmits the calibration signal within the data frame guard interval using time-division multiplexing. The calibration process does not interrupt normal data transmission and does not occupy any service transmission bandwidth.

[0029] In this invention, the adaptive parameter adjustment module adopts a parameter update mechanism driven by both multi-dimensional coupling error and environment. The parameter update formula is as follows: ;in For the first The compensation parameter weights for the next iteration are dimensionless. For the first The compensation parameter weights for the next iteration are dimensionless. The error rate of change influence coefficient is dimensionless and ranges from 0.01 to 0.1. For the first The normalized multidimensional coupling error change rate of the next iteration is dimensionless and is obtained by dividing the difference between the current error and the previous error by the system reference error. The environmental change rate influence coefficient is dimensionless and ranges from 0.005 to 0.05. For the first The normalized environmental comprehensive change rate of each iteration is dimensionless and is obtained by weighted summation of changes in temperature, humidity, vibration, and electromagnetic interference. An adaptive adjustment mechanism for parameter update step size is incorporated. When the error exceeds a set threshold, the step size is increased to achieve rapid convergence, and when the error is less than the set threshold, the step size is decreased to achieve fine adjustment. Parameter boundary constraints are set to prevent updates from exceeding a reasonable range. The update process is smoothed by weighted averaging of historical parameters, which improves response speed when the error changes rapidly and improves compensation accuracy when the environment is stable.

[0030] In this invention, the dynamic error compensation method for millimeter-wave transmission links includes the following steps: The initial configuration is completed and the initial error calibration is performed to establish the reference error mapping relationship; the millimeter wave signal transceiver module transmits the reference millimeter wave signal and collects the received signal transmitted back through the transmission link; the multimodal environment perception module synchronously collects the environmental parameters around the link. The multi-dimensional error perception module analyzes the received signal, extracts the phase error, amplitude error, group delay error and frequency offset error of the transmission link, constructs multi-dimensional coupling error, and generates multi-dimensional error feature vector; The adaptive parameter adjustment module calculates the current compensation parameter weights based on the error feature vector and environmental parameters, and updates the compensation parameters of the digital predistortion compensation module based on the compensation parameter weights. The digital predistortion compensation module generates a corresponding digital compensation signal based on the updated compensation parameters, which is then injected into the transmission link after digital-to-analog conversion and radio frequency modulation to achieve dynamic compensation of multi-dimensional coupling errors. The adaptive parameter adjustment module executes a closed-loop online calibration process based on preset period or error threshold trigger conditions to update the system reference compensation parameters.

[0031] In this invention, generating a multi-dimensional error feature vector includes: sequentially performing down-conversion and quadrature demodulation on the received signal to obtain baseband in-phase and quadrature components; calculating the average phase and average amplitude of multiple consecutive sampling points through a sliding window of a preset length (e.g., the preset length can be 128 points), and comparing them with the phase and amplitude of the transmitted reference signal to obtain phase error and amplitude error; obtaining the group delay error by calculating the peak position of the cross-correlation function of the transmitted and received signals; extracting the carrier frequency of the received signal based on phase-locked loop technology, and comparing it with the transmitted carrier frequency to obtain frequency offset error; calculating the correlation coefficient between different error components, identifying multi-dimensional coupling error relationships, and processing each error component using a piecewise normalization method, splicing them together to generate an error feature vector with a dimension of 4.

[0032] In this invention, the closed-loop online calibration process includes: triggering the online calibration process when the system's continuous running time reaches a preset calibration cycle or when the link bit error rate exceeds a set threshold; transmitting a calibration signal from the standard calibration signal source of the adaptive parameter adjustment module within the data frame guard interval, and collecting the calibration signal after link transmission by the calibration signal receiving unit of the adaptive parameter adjustment module; calculating the difference between the current error characteristics of the link and the reference error characteristics to generate calibration correction coefficients; using an incremental update method, only adjusting compensation parameters whose changes exceed a set threshold, and updating the lookup table and filter coefficients of the digital predistortion compensation module; verifying the calibration effectiveness by comparing the link signal-to-noise ratio before and after calibration; automatically restoring normal data transmission after calibration, and storing calibration data and correction coefficients to provide historical data support for subsequent adaptive adjustment.

[0033] The following two examples further illustrate the specific implementation of this system: The millimeter-wave transmission link dynamic error compensation system of this invention adopts a hierarchical architecture of hardware and software collaboration. Through multi-dimensional error synchronous sensing, multi-modal environmental data fusion, digital pre-error real-time compensation, online closed-loop calibration, and adaptive parameter adjustment, it achieves comprehensive and accurate correction of dynamic errors in the millimeter-wave transmission link. The system uses a multi-core embedded processor as its control core, integrating a high-speed RF transceiver unit, a high-precision error detection unit, and a programmable digital signal processing unit. It supports operation across the entire frequency band from 24GHz to 100GHz, with a processing latency of no more than 50ns. The system integrates environmental prediction and error feedforward technologies, achieving a combination of passive compensation and active pre-compensation. It also possesses fault self-diagnosis and redundancy backup capabilities, enabling it to adapt to complex and ever-changing transmission environments in different scenarios, effectively improving the transmission quality and reliability of millimeter-wave communication links. The invention is further described in detail below through two specific embodiments.

[0034] Example 1

[0035] This embodiment applies to a point-to-point millimeter-wave backhaul link between a 5G macro base station and the core network, operating at a frequency of 28GHz, with a transmission distance of 1.2 kilometers and a link transmission rate of no less than 10Gbps. This link is deployed on the rooftops of high-rise buildings in the city, surrounded by numerous mobile communication base stations and broadcast television towers, resulting in a complex electromagnetic environment. Furthermore, it is affected by diurnal temperature variations, seasonal climate changes, and antenna vibrations caused by wind, leading to frequent link error fluctuations. Therefore, it places high demands on the system's real-time compensation and anti-interference capabilities.

[0036] During system deployment, a millimeter-wave transmission link dynamic error compensation system is deployed at both the base station and core network ends, with the two systems communicating synchronously via a dedicated control channel. All hardware modules are integrated into a standard 19-inch rack and powered by redundant power supplies. The millimeter-wave signal transceiver module connects to a high-gain directional antenna via a waveguide interface, and the antenna azimuth and elevation angles can be finely adjusted via an electric adjustment mechanism. The multi-modal environmental sensing module's sensors are installed at the antenna feed port and inside the equipment chassis, synchronously collecting environmental parameters from the link's periphery and the equipment's interior. The fault self-diagnosis and redundancy backup module is configured with a 1+1 hot backup compensation channel; the primary and backup channels use identical hardware architectures and synchronize operating status and compensation parameters in real time.

[0037] After the system powers on, the initialization process is executed first. The central control module performs hardware self-tests on each module sequentially, checking the power supply voltage, operating current, and internal register status. After confirming that all modules are working properly, the factory default parameters are loaded. The adaptive parameter adjustment module executes the initial calibration process. A standard calibration signal source generates a linear frequency modulated signal covering a frequency range of 27.5 GHz to 28.5 GHz, which is then acquired by the calibration signal receiving unit at the other end after passing through the transmission link. The calibration parameter calculation unit compares the parameter differences between the transmitted and received signals, calculates the inherent insertion loss, phase offset, and group delay characteristics of the link, generates initial compensation parameters, and writes them into the lookup table of the digital predistortion compensation module, establishing a system reference error database.

[0038] During normal operation, the millimeter-wave signal transceiver module transmits a reference millimeter-wave signal at a sampling rate of 2GSa / s and simultaneously receives the return signal after transmission via the link, with an analog-to-digital conversion accuracy of 12 bits. The multi-dimensional error sensing module adopts a synchronous parallel sampling architecture, simultaneously detecting the phase, amplitude, group delay, and frequency offset of the received signal. The phase detection unit extracts in-phase and quadrature components through orthogonal mixing, obtains the signal phase value through high-precision arctangent calculation, and compares it with the transmitted reference phase to obtain the phase error. The amplitude detection unit uses true RMS detection technology to achieve amplitude measurement over a wide dynamic range of -60dBm to +10dBm. The group delay detection unit uses linear frequency modulation pulse compression technology to improve the delay detection resolution to 0.5ps. The data preprocessing unit sequentially performs Kalman filtering, moving average, and wavelet denoising on the raw error data to separate slowly varying drift errors and rapidly varying burst errors, eliminate the influence of random noise and sudden interference, and generate a multi-dimensional error feature vector.

[0039] The multimodal environment sensing module synchronously collects temperature, humidity, vibration, and electromagnetic interference data at a sampling frequency of 10Hz, and uses an adaptive weighted fusion algorithm to fuse the multi-sensor data. An environmental error time-series prediction model trained on historical data predicts the link error change trend within the next 10 seconds based on the current environmental parameter change trend, generating prior environmental information and sending it to the adaptive parameter adjustment module. The adaptive parameter adjustment module combines the error feature vector and prior environmental information, employing a dual-driven parameter update mechanism to dynamically adjust the update weights and step sizes of the compensation parameters. When the error exceeds a set threshold, the update step size is increased for rapid convergence; when the error is less than the set threshold, the step size is decreased for fine-tuning. Simultaneously, parameter boundary constraints are set to prevent updates from exceeding a reasonable range.

[0040] The digital predistortion compensation module is implemented based on a field-programmable gate array (FPGA) architecture and employs a parallel pipeline architecture to simultaneously process phase, amplitude, and time delay compensation signals. The module divides the 28GHz operating frequency band into eight consecutive sub-bands, each with independently trained compensation coefficients, introducing a memory polynomial predistortion structure to compensate for the nonlinear memory effect of the power amplifier. A high-speed digital-to-analog converter (DAC) converts the digital compensation signal into an analog RF signal at a conversion rate of 5GSa / s, which is then injected into the transmission link via a vector modulator in the RF front-end compensation unit. An integrated adaptive impedance matching network dynamically adjusts the impedance according to changes in link load, ensuring efficient transmission of the compensation signal. The phase compensation range covers 0 to 360 degrees, the amplitude compensation range is -10dB to +10dB, and the time delay compensation range is 0 to 100ns.

[0041] The adaptive parameter adjustment module performs closed-loop online calibration according to a preset period or error triggering conditions. The calibration process is triggered when the system's continuous operating time reaches the preset calibration period or when the link bit error rate exceeds a set threshold. The standard calibration signal source transmits calibration signals within the data frame guard interval, without occupying any service transmission bandwidth, and the calibration process does not interrupt normal data transmission. The calibration signal receiving unit collects the calibration signal after transmission through the link, calculates the difference between the current error characteristics and the reference error characteristics, and generates calibration correction coefficients. An incremental update method is used to adjust only parameters whose changes exceed a set threshold, updating the lookup table and filter coefficients of the digital predistortion compensation module. The link signal-to-noise ratio before and after calibration is compared to verify the calibration effectiveness. After calibration, normal data transmission is automatically restored, and the calibration data and correction coefficients are saved to the system's non-volatile memory.

[0042] The real-time status monitoring module continuously collects data on the link's transmission rate, bit error rate, signal-to-noise ratio, and the operational status of each module. It generates a link operational status report every 5 minutes and a compensation effect evaluation report every hour. When a module malfunctions or link performance deteriorates, a fault self-diagnosis process is triggered. The fault self-diagnosis and redundancy backup module uses fault tree analysis to identify fault types and severity, classifying faults into low, medium, and high levels. Low-level faults are minor and self-heal through adaptive parameter adjustment. Medium-level faults automatically switch to the hot backup compensation channel within 1ms. High-level faults are severe, generating audible and visual alarms and recording detailed fault logs, while simultaneously uploading the fault information to the network management platform. A fault knowledge base is established to analyze historical fault data, enabling automatic root cause analysis and predictive maintenance.

[0043] Example 2

[0044] This embodiment applies to a millimeter-wave transmission link between CNC machine tools and a central control system in an industrial automation workshop. The operating frequency is 60GHz, the transmission distance is 50 meters, the transmission rate is no less than 5Gbps, and the transmission delay is required to be no more than 1ms. The workshop contains numerous motors, frequency converters, and welding equipment, resulting in high electromagnetic interference. Furthermore, the continuous vibrations generated by the machine tools and reflections from metal components lead to significant multipath effects in the link. The ambient temperature also varies considerably with the production process, placing extremely high demands on the system's anti-interference capabilities, real-time performance, and reliability.

[0045] During system deployment, the system is deployed in both the CNC machine tool control cabinet and the central control room, employing a miniaturized, integrated design. All modules are integrated into an industrial-grade protective enclosure, meeting IP54 protection requirements. The millimeter-wave signal transceiver module uses a flat panel antenna, installed on an unobstructed location on top of the equipment, with a 30-degree antenna beamwidth to ensure signal coverage and anti-interference capabilities. The multimodal environmental sensing module's vibration sensor is mounted on the antenna bracket, the electromagnetic interference sensor is installed at the equipment's power input, and the temperature and humidity sensors are installed inside the enclosure to monitor the equipment's operating environment in real time. The fault self-diagnosis and redundancy backup module is equipped with dual power supplies and dual compensation channels to ensure continuous system operation in the event of a single-path failure.

[0046] During the system power-on initialization phase, the central control module completes industrial-grade hardware self-tests for each module and performs electromagnetic compatibility initialization configuration for the digital signal processing unit, taking into account the electromagnetic interference characteristics of the industrial environment. The adaptive parameter adjustment module performs initial calibration, considering the multipath effect in the industrial environment, and uses a multipath suppression algorithm to process the calibration signal, eliminating the interference of reflected signals on the calibration results, generating initial compensation parameters suitable for the industrial environment, and establishing a reference error database.

[0047] During normal operation, the millimeter-wave signal transceiver module transmits and receives signals at a sampling rate of 2 GSa / s, employing orthogonal frequency division multiplexing modulation to improve spectral utilization and multipath resistance. The multi-dimensional error sensing module uses a synchronous parallel sampling architecture to detect the phase, amplitude, group delay, and frequency offset errors of the link in real time. Addressing the characteristics of frequent sudden interference in industrial environments, the data preprocessing unit strengthens wavelet denoising, effectively filtering out pulse interference generated by motor start-up and shutdown and welding operations, accurately extracting the true error characteristics of the link, and generating a multi-dimensional error feature vector.

[0048] The multimodal environment sensing module collects environmental parameters at a sampling frequency of 10Hz, focusing on monitoring changes in vibration and electromagnetic interference. It establishes a mapping relationship between industrial-specific environmental parameters and link errors, training a specialized time-series prediction model capable of predicting error fluctuations caused by machine tool vibration and equipment start-up / shutdown. The adaptive parameter adjustment module, considering the rapid changes in industrial environment errors, appropriately increases the error change rate influence coefficient to improve the system's response speed to rapid error changes while maintaining compensation accuracy in stable environments.

[0049] The digital predistortion compensation module adopts a parallel pipeline architecture, with processing latency controlled within 50ns, meeting the real-time requirements of industrial transmission. Targeting the transmission characteristics of the 60GHz band, the operating frequency is divided into eight consecutive sub-bands, each with independently optimized compensation coefficients. A memory polynomial predistortion structure is introduced to compensate for the nonlinear memory effect of the power amplifier, and an adaptive impedance matching network is integrated to address dynamic changes in link load in industrial environments, achieving real-time compensation for multi-dimensional errors across the entire frequency band.

[0050] The adaptive parameter adjustment module employs a calibration strategy that prioritizes error triggering and supplements it with timed calibration. When the link bit error rate exceeds a set threshold or environmental parameters undergo drastic changes, the online calibration process is immediately triggered. Calibration signals are transmitted within the data frame guard interval, ensuring uninterrupted transmission of normal industrial data. Incremental updates of compensation parameters minimize the impact of the calibration process on system performance, guaranteeing the continuity of industrial production.

[0051] The real-time status monitoring module monitors the link transmission status and equipment operating status in real time, uploading the data to the workshop production management system. When a fault is detected, the fault self-diagnosis and redundancy backup module completes the primary and backup channel switching within 1ms to ensure uninterrupted industrial control signals. An industrial equipment fault knowledge base is established, and fault prediction is performed in conjunction with production process data to proactively schedule equipment maintenance and prevent production accidents.

[0052] This invention comprehensively addresses the dynamic error problem caused by environmental influences in millimeter-wave transmission links through multi-dimensional synchronous error sensing and real-time compensation across all dimensions. Multi-modal environmental sensing and error prediction technologies enable proactive pre-compensation, significantly improving the system's response speed to environmental changes. Online calibration technology updates parameters without interrupting service transmission, ensuring continuous system operation. Frequency-band compensation and nonlinear memory effect correction technologies enhance the accuracy of full-band compensation. Fault self-diagnosis and redundancy backup mechanisms enhance system reliability, enabling long-term stable operation in complex and changing transmission environments, meeting the high-reliability millimeter-wave communication needs of various fields.

[0053] Reference Figure 2This diagram illustrates the system's layered architecture. The system uses a central control module as its core, responsible for task scheduling and module interaction across the entire system. The overall architecture is divided into three layers: the physical sensing and execution layer, the signal processing and compensation layer, and the logic control and support layer. The physical sensing and execution layer includes a millimeter-wave signal transceiver module and a multimodal environment sensing module. The signal processing and compensation layer includes a multi-dimensional error sensing module and a digital predistortion compensation module. The logic control and support layer includes an adaptive parameter adjustment module, a link online calibration module, and a fault self-diagnosis and redundancy backup module. These modules interact via an internal bus, realizing a complete link from physical link signal extraction to digital algorithm compensation.

[0054] Reference Figure 3 This diagram illustrates the internal logic of the sensing module. The system synchronously extracts four types of coupling errors from the link through orthogonal mixing, true RMS detection, pulse compression, and phase-locked loop techniques. The preprocessing unit introduces multi-domain joint noise reduction and piecewise normalization algorithms to effectively identify the coupling relationship between error components and separate slowly varying drift errors (trends) and rapidly varying burst errors (fluctuations), outputting a standardized error feature vector.

[0055] Reference Figure 4 This diagram illustrates the generation and injection path of the compensation signal. The digital signal processing unit (DSP) is based on a parallel pipeline architecture, dividing the frequency band into multiple sub-bands for independent coefficient training, and introducing a memory polynomial structure to compensate for nonlinear memory effects. The digital compensation signal is converted into an analog signal by a high-speed DAC and injected into the link through a vector modulator integrating an adaptive impedance matching network, achieving precise dynamic compensation of link errors.

[0056] Reference Figure 5 This diagram illustrates the system's self-maintenance logic. Online calibration involves injecting a single-tone / linear frequency modulation signal into the data frame protection interval, calculating the difference between the current and reference characteristics, and implementing incremental updates. The fault self-diagnosis module, based on fault tree analysis, categorizes anomalies into three levels: low (algorithm self-healing), medium (redundant channel switching), and high (alarm / logging), forming a self-evolutionary and defensive safety closed loop.

[0057] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A dynamic error compensation system for millimeter-wave transmission links, characterized in that, include: The millimeter-wave transceiver module operates in the full frequency band from 24GHz to 100GHz and is used to transmit reference millimeter-wave signals and receive the return signals after transmission through the link, and to collect the original transmission data of the link. The multi-dimensional error perception module adopts a synchronous parallel sampling architecture, which is synchronized with the millimeter-wave signal transceiver module in timing. It performs synchronous joint detection of phase error, amplitude error, group delay error and frequency offset error of the transmission link, and outputs multi-dimensional coupled error detection data. The digital predistortion compensation module receives the coupling error detection data output by the multi-dimensional error perception module, generates a digital compensation signal, and outputs it to the millimeter-wave signal transceiver module to achieve signal predistortion compensation. The adaptive parameter adjustment module is used to dynamically adjust the core parameters of the predistortion compensation algorithm in the digital predistortion compensation module based on the coupling error change trend output by the multi-dimensional error perception module, combined with environmental fluctuation information and link transmission quality. It distinguishes between inherent link error and drift error based on accumulated coupling error data, completes closed-loop automatic calibration, evaluates the compensation effect based on link operation indicators, and outputs the control feedback signal in reverse to the digital predistortion compensation module and the multi-dimensional error perception module.

2. The millimeter-wave transmission link dynamic error compensation system according to claim 1, characterized in that, It also includes a multimodal environment sensing module, which integrates temperature sensors, humidity sensors, vibration sensors and electromagnetic interference sensors to synchronously collect environmental parameters around the link. It uses an adaptive weighted fusion algorithm to fuse the multi-sensor data, establishes a nonlinear correlation mapping relationship between environmental parameters and multi-dimensional coupling errors of the millimeter-wave transmission link, and outputs it to the adaptive parameter adjustment module.

3. The millimeter-wave transmission link dynamic error compensation system according to claim 1, characterized in that, It also includes a fault self-diagnosis and redundancy backup module, which monitors the power supply voltage, operating current, output signal quality and internal register status of the millimeter wave signal transceiver module, multi-dimensional error perception module, digital predistortion compensation module and adaptive parameter adjustment module in real time. It uses fault tree analysis to identify hardware faults and link anomalies, and classifies faults into three levels: low, medium and high according to the degree of impact of faults on the performance of multi-dimensional coupling error compensation. In the event of a low-level fault, the adaptive parameter adjustment module is linked to adjust the predistortion compensation algorithm parameters for self-healing. In the event of a medium-level failure, the system will automatically switch to a primary and backup hot backup compensation channel. In the event of a high-level fault, an audible and visual alarm signal will be output and a fault log will be recorded.

4. The millimeter-wave transmission link dynamic error compensation system according to claim 1, characterized in that, The multi-dimensional error perception module includes a phase detection unit, an amplitude detection unit, a group delay detection unit, and a data preprocessing unit. The phase detection unit extracts the in-phase and quadrature components of the received signal using orthogonal mixing, calculates the signal phase value through arctangent operation, and compares it with the reference phase of the transmitting end to obtain the phase error. The amplitude detection unit completes the amplitude detection of the received signal using true RMS detection. The data preprocessing unit performs multi-domain joint noise reduction processing on the multi-dimensional coupled error data, separates the slow-varying drift error and the fast-varying burst error, and provides error input for the digital predistortion compensation module and the adaptive parameter adjustment module.

5. The millimeter-wave transmission link dynamic error compensation system according to claim 1, characterized in that, The digital predistortion compensation module includes a digital signal processing unit, a high-speed digital-to-analog converter unit, and a radio frequency front-end compensation unit. The digital signal processing unit adopts a parallel pipeline architecture to simultaneously process the phase, amplitude, and time delay compensation signals. It introduces a memory polynomial predistortion structure to compensate for the nonlinear memory effect of the link, dividing the millimeter-wave operating frequency band into multiple continuous sub-bands. Each sub-band independently trains compensation coefficients to generate a digital compensation signal for compensating for multi-dimensional coupling errors. The high-speed digital-to-analog converter unit converts the digital compensation signal into an analog radio frequency compensation signal. The RF front-end compensation unit injects analog RF compensation signals into the transmission link through a vector modulator and integrates an adaptive impedance matching network to dynamically adjust the impedance according to changes in link load.

6. The millimeter-wave transmission link dynamic error compensation system according to claim 1, characterized in that, The adaptive parameter adjustment module includes a standard calibration signal source, a calibration signal receiving unit, and a calibration parameter calculation unit. The standard calibration signal source outputs single-tone and linear frequency modulated signals with controllable frequency and amplitude, and automatically adjusts the output power according to the link attenuation status; The calibration signal receiving unit collects the calibration signal after it has been transmitted through the link. The calibration parameter calculation unit compares the parameter differences between the transmitted and received calibration signals, calculates the inherent insertion loss, phase offset, and group delay characteristics of the link, and transmits the calibration signal within the data frame guard interval using time-division multiplexing.

7. The millimeter-wave transmission link dynamic error compensation system according to claim 1, characterized in that, The adaptive parameter adjustment module adopts a parameter update mechanism driven by both multi-dimensional coupling error and environment. The parameter update formula is as follows: ;in For the first The compensation parameter weights for the next iteration; For the first The compensation parameter weights for the next iteration; The coefficient representing the influence of the rate of change of error; For the first The rate of change of normalized multidimensional coupling error in the next iteration; The environmental change rate influence coefficient; For the first The normalized comprehensive environmental change rate of the next iteration is obtained by weighted summation of the changes in temperature, humidity, vibration, and electromagnetic interference.

8. A method for dynamic error compensation of a millimeter-wave transmission link, applied to a dynamic error compensation system for a millimeter-wave transmission link as described in any one of claims 1 to 7, characterized in that, Includes the following steps: Complete the initial configuration and perform initial error calibration to establish the baseline error mapping relationship; The millimeter-wave signal transceiver module transmits a reference millimeter-wave signal and collects the received signal transmitted back via the transmission link. The multi-modal environment sensing module simultaneously collects environmental parameters around the link. The multi-dimensional error perception module analyzes the received signal, extracts the phase error, amplitude error, group delay error and frequency offset error of the transmission link, constructs multi-dimensional coupling error, and generates multi-dimensional error feature vector; The adaptive parameter adjustment module calculates the current compensation parameter weights based on the error feature vector and environmental parameters, and updates the compensation parameters of the digital predistortion compensation module based on the compensation parameter weights. The digital predistortion compensation module generates a corresponding digital compensation signal based on the updated compensation parameters, which is then injected into the transmission link after digital-to-analog conversion and radio frequency modulation to complete the dynamic compensation of multi-dimensional coupling errors. The adaptive parameter adjustment module executes a closed-loop online calibration process and updates the reference compensation parameters based on preset period or error threshold trigger conditions.

9. The method for dynamic error compensation of millimeter-wave transmission links according to claim 8, characterized in that, The generation of multi-dimensional error feature vectors includes: performing down-conversion and quadrature demodulation on the received signal sequentially to obtain baseband in-phase and quadrature components; calculating the average phase and average amplitude of multiple consecutive sampling points through a sliding window of preset length, and comparing them with the phase and amplitude of the transmitted reference signal to obtain phase error and amplitude error; obtaining the group delay error by calculating the peak position of the cross-correlation function of the transmitted and received signals; extracting the carrier frequency of the received signal based on a phase-locked loop and comparing it with the transmitted carrier frequency to obtain frequency offset error; calculating the correlation coefficient between different error components, identifying multi-dimensional error coupling relationships, and processing each error component using a piecewise normalization method.

10. The method for dynamic error compensation of a millimeter-wave transmission link according to claim 8, characterized in that, The closed-loop online calibration process includes: triggering online calibration when the system's continuous running time reaches a preset calibration cycle or when the link bit error rate exceeds a set threshold; transmitting calibration signals from the standard calibration signal source of the adaptive parameter adjustment module within the data frame guard interval, and collecting the calibration signals after link transmission by the calibration signal receiving unit of the adaptive parameter adjustment module; calculating the difference between the current error characteristics of the link and the reference error characteristics to generate calibration correction coefficients; using an incremental update method to adjust only compensation parameters whose changes exceed a set threshold; verifying the calibration effectiveness by comparing the link signal-to-noise ratio before and after calibration; restoring normal data transmission after calibration and storing calibration data and correction coefficients.

Images