High-gain dual-polarization parabolic antenna system and performance maintaining method thereof

Abstract

The invention relates to a high-gain dual-polarization parabolic antenna system and a performance maintaining method thereof. The system comprises a parabolic reflector; the feed source is arranged at a focus or a perifocus area of the parabolic reflector; the feed source comprises a printed circuit board (PCB), and a metal oscillator unit is formed on the PCB through an etching process; the metal oscillator unit comprises at least two groups of orthogonally arranged oscillators to form radiation in two orthogonal polarization directions; a low-frequency radiation part for a first frequency band and a high-frequency radiation part for a second frequency band are integrated on each oscillator of the metal oscillator unit, and the first frequency band and the second frequency band respectively correspond to a 900MHz frequency band and a 1800MHz frequency band; the feed source radiates dual-polarized electromagnetic waves through the metal oscillator unit, and the electromagnetic waves are reflected by the parabolic reflector to form high-gain directional wave beams. According to the invention, while the compact structure and low-cost manufacturing are realized, the electrical performance of the antenna in a wide frequency band is remarkably improved, and an efficient and flexible solution is provided for modern wireless communication.

Description

Technical Field

[0001] The present invention relates to the technical field of antenna system equipment, and in particular to a high-gain dual-polarization parabolic antenna system and a performance maintenance method thereof. Background Art

[0002] With the continued evolution of 4G and future 5G mobile communication technologies, wireless communication systems are placing higher demands on antenna performance. To effectively improve channel capacity and spectrum utilization, MIMO (Multiple Input Multiple Output) technology has become a mainstream deployment solution, placing a clear demand on antennas for dual-polarization capabilities. Furthermore, in specific scenarios such as emergency communications, temporary base station deployments for large-scale events, and wide-area coverage in rural areas, antennas must possess high gain and strong directivity to quickly establish stable and reliable wireless connections. Parabolic antennas, with their inherent advantages of high gain and narrow beamform, are a common choice in these applications.

[0003] In actual deployment, considering the compatibility of different network standards (such as 2G / 3G / 4G) and the efficient use of spectrum resources, antennas usually need to cover a wider frequency band or simultaneously support multiple frequency bands that are far apart, such as 900MHz and 1800MHz. Existing technologies usually increase the diameter of the parabolic reflector to obtain a high gain of 15-25 dBi, use reflector shaping or special feed design techniques to reduce the sidelobe level to below -20 dB, and use a tapered feed structure or a complex impedance matching network to achieve a low voltage standing wave ratio (VSWR ≤ 1.7) within the required frequency band. However, existing technologies still face significant challenges in designing feeds that can simultaneously meet the requirements of high gain, dual polarization, dual frequency bands (especially 900MHz and 1800MHz), and are compact, low-cost, and easy to mass-produce:

[0004] 1. The challenge of dual-band integration and performance balance: For the 900MHz and 1800MHz frequency bands, which operate nearly an octave apart, achieving both good resonant characteristics and impedance matching within a single feed structure is a complex task. Existing PCB feeds, to cover both frequency bands simultaneously, often require multi-layer PCBs or complex coupling structures. This increases the feed size and complexity, and can easily lead to undesirable mutual coupling between radiating elements in different frequency bands, impacting the performance of each band.

[0005] 2. Insufficient integration of dual-polarization and dual-band: Traditional dual-polarization feed designs are implemented through the orthogonal arrangement of two independent single-polarization feeds or complex orthogonal mode couplers. When it is necessary to further integrate dual-band functionality on this basis, existing technologies often struggle to maintain a compact structure while ensuring high polarization isolation between the two orthogonally polarized ports in both operating frequency bands. Integrating two operating frequency bands and two orthogonally polarized radiating elements on a PCB through a planar etched structure, while ensuring low mutual coupling between them and excellent performance within their respective frequency bands, is a technical issue that continues to attract attention in the field of antenna design but has not yet been fully resolved.

[0006] 3. Manufacturing complexity and cost control: Existing dual-frequency, dual-polarization feeds that use non-PCB processes or complex multi-layer PCB designs increase manufacturing complexity, material costs, and production cycles, hindering large-scale, low-cost deployment. Efficiently implementing high-performance dual-frequency, dual-polarization feeds using planar etching on a single or limited number of PCB layers remains a pressing technical bottleneck.

[0007] In view of the above-mentioned deficiencies in the existing technology, there is an urgent need for an innovative solution that can provide a parabolic antenna feed and antenna based on PCB etched metal vibrator, which has a relatively simple structure, is easy to manufacture, and has good impedance matching, high gain and dual-polarization characteristics in both 900MHz and 1800MHz frequency bands. Summary of the Invention

[0008] To address the problems of the prior art to the greatest extent possible, the present invention provides a high-gain dual-polarization parabolic antenna system. By integrating a metal oscillator unit with dual-band (900MHz and 1800MHz) and dual-polarization functions on a single printed circuit board (PCB), the present invention helps overcome the challenges of the prior art in dual-band dual-polarization feed sources in terms of structural complexity, manufacturing difficulty, and achieving good impedance matching and high polarization isolation within a wide bandwidth, thereby providing a high-performance, easy-to-deploy antenna solution for modern wireless communication systems.

[0009] The present invention discloses a high-gain dual-polarized parabolic antenna system. The system's basic structure comprises a parabolic reflector and a feed source. The feed source is strategically positioned at or near the focal point of the parabolic reflector to ensure that electromagnetic waves radiated from the feed source are efficiently collected by the reflector and formed into a highly directional beam. The core innovation of the present invention lies in the specific construction of the feed source. A key component of the feed source is a printed circuit board (PCB). The selective use of a PCB as the feed source substrate provides a foundation for subsequent manufacturing processes and integration. Metal oscillator units are meticulously formed on the PCB using an etching process. Etching is a mature and cost-effective manufacturing method that can precisely form complex metal patterns on the PCB surface. These patterns constitute the antenna's radiating elements. The design of the metal oscillator units is central to the present invention's versatility. They comprise at least two sets of orthogonally arranged oscillators. This orthogonal arrangement is key to achieving dual-polarized radiation, such as vertical and horizontal polarization, or ±45-degree polarization. By independently feeding these two sets of orthogonal dipoles, the antenna can simultaneously receive or transmit two mutually orthogonal polarization signals, thus supporting MIMO (Multiple Input Multiple Output) technology and improving channel capacity and spectrum utilization. A further innovation lies in the design of each metal dipole unit. Each dipole (regardless of which polarization set it belongs to) integrates a low-frequency radiator for the first frequency band (900MHz) and a high-frequency radiator for the second frequency band (1800MHz). This means that traditional designs that require separate antennas or complex multi-layer structures to achieve dual-band operation are cleverly integrated into a single dipole structure in this invention. The low-frequency radiator is typically physically larger to resonate in the lower 900MHz band, while the high-frequency radiator is relatively small to resonate in the higher 1800MHz band. By precisely designing and optimizing the specific geometric shapes, dimensions, relative positions (e.g., nested or integrated) of these radiating parts, and the opening angles of the dipole arms, a single dipole can achieve good impedance matching in two widely separated frequency bands, ensuring efficient and stable signal transmission. Ultimately, the feed radiates dual-polarized electromagnetic waves through the metal dipole unit of this solution. These electromagnetic waves are then reflected by the parabolic reflector, and the beam-forming effect of the reflector converges the scattered electromagnetic waves into a directional beam with high gain and strong directivity, thereby meeting the needs of long-distance communication or high-density user coverage.

[0010] According to a high-gain dual-polarized parabolic antenna system of the present invention, the metal dipole unit includes a first polarized dipole group, which is composed of a first metal dipole arm and a second metal dipole arm symmetrically arranged therewith. The first and second metal dipole arms together generate radiation in a single polarization direction. Specifically, the scheme further defines the specific composition of the metal dipole unit. In one polarization direction, the dipole unit is specifically designed to consist of two symmetrical metal dipole arms (B1 and B3). These two arms are formed on a PCB through an etching process and arranged symmetrically so that they resonate together under feed excitation and generate electromagnetic wave radiation in a specific linear polarization direction. This symmetrical dipole-like structure is the basic unit for achieving radiation in a specific polarization direction. It can be understood that by etching the first polarized dipole group consisting of the first metal dipole arm (B1) and the second metal dipole arm (B3) on the PCB, the present invention provides a basic unit for achieving radiation in a specific polarization direction. This symmetrical arm-shaped structure ensures that the polarization direction of the radiated electromagnetic wave is clear and the radiation pattern has good symmetry. This configuration provides the structural basis for the subsequent realization of dual-polarization function. By precisely controlling the geometric dimensions and relative positions of the arms, the required polarization mode can be effectively excited, thereby ensuring that the energy is concentrated in the desired radiation direction and providing the premise for further gain improvement.

[0011] According to a high-gain dual-polarized parabolic antenna system of the present invention, the metal dipole unit further includes a second polarized dipole group, which consists of a third metal dipole arm and a fourth metal dipole arm arranged symmetrically therewith. This second polarized dipole group is arranged 90 degrees relative to the first polarized dipole group on the PCB plane, radiating in a polarization direction orthogonal to the first polarized dipole group. Specifically, this solution, building on the previous solution, introduces a second polarized dipole group (B2 and B4), which also exhibits a symmetrical structure. Crucially, this second polarized dipole group is arranged 90 degrees relative to the first polarized dipole group (B1 and B3) on the PCB plane. This orthogonal arrangement provides a structural guarantee for achieving dual-polarization functionality, enabling the two groups of dipoles to independently radiate two mutually orthogonal electromagnetic wave polarization directions. It can be understood that by integrating and orthogonally (rotating 90 degrees) the second polarized dipole group (B2 and B4) with the first polarized dipole group (B1 and B3), the present invention successfully achieves dual-polarization functionality. This orthogonal structure enables the two sets of oscillators to independently excite and radiate electromagnetic waves with perpendicular polarizations, allowing multiple independent signals to be transmitted using different polarizations within the same frequency band, or for polarization diversity reception. This dual-polarization capability directly supports the application of advanced wireless communication technologies such as Multiple Input Multiple Output (MIMO), significantly improving channel capacity and spectrum utilization, thereby enhancing the throughput and reliability of communication systems.

[0012] According to a high-gain dual-polarized parabolic antenna system of the present invention, each of the first, second, third, and fourth metal dipole arms includes a low-frequency radiating portion and a high-frequency radiating portion, wherein the high-frequency radiating portion is nested or integrated within or near the structure of the low-frequency radiating portion, and the low-frequency radiating portion is a longer and wider arm-shaped structure, while the high-frequency radiating portion is a shorter and narrower arm-shaped structure. Specifically, the scheme further elaborates on the details of each dipole arm, namely, each arm is not a single structure, but integrates two radiating portions of different sizes: a low-frequency radiating portion for the 900MHz frequency band, and a high-frequency radiating portion for the 1800MHz frequency band. By nesting or integrating the high-frequency radiating portion within or near the structure of the low-frequency radiating portion, and giving them longer / wider and shorter / narrower geometric characteristics, respectively, it is achieved that a single dipole arm can simultaneously support the resonance and radiation of two frequency bands that are far apart. It can be understood that by integrating a low-frequency radiating part and a high-frequency radiating part on each metal vibrator arm, and adopting a compact design in which the high-frequency radiating part is nested or integrated inside or near the low-frequency radiating part, the present invention realizes the function of simultaneously supporting two operating frequency bands of 900MHz and 1800MHz on a single PCB feed. The longer and wider structure of the low-frequency radiating part enables it to effectively resonate and radiate in the 900MHz frequency band, while the shorter and narrower structure of the high-frequency radiating part enables it to effectively resonate and radiate in the 1800MHz frequency band. This design enables the antenna to be compatible with multiple mobile communication network standards such as 2G / 3G / 4G, thereby expanding the scope of application and market compatibility of the antenna. At the same time, this dual-frequency integrated structure significantly reduces the size and complexity of the feed, simplifies the manufacturing process, and is conducive to reducing production costs and improving product consistency.

[0013] According to a high-gain dual-polarized parabolic antenna system according to the present invention, the feed also includes a director. This director is formed on the PCB through an etching process and is located in front of the main radiation direction of the high-frequency radiator. This director is used to expand the operating bandwidth and increase the gain in the 1800 MHz frequency band. Specifically, this solution introduces additional auxiliary structures—directors (D). These director (D) are also formed by etching the PCB and are strategically placed in front of the main radiation direction of the high-frequency radiator. As a parasitic element, the director (D) electromagnetically couples with the high-frequency radiator. By optimizing its size and position, it can adjust the radiation characteristics in the 1800 MHz band. It can be understood that the introduction and optimization of the director (D) etched on the PCB in the feed and its placement in front of the main radiation direction of the high-frequency radiator can significantly improve the antenna's performance in the 1800 MHz band. As an electromagnetic coupling element, the director (D) effectively guides and shapes the radiation field of the high-frequency radiator, concentrating the radiated energy forward. This results in: first, expanding the operating bandwidth in the 1800MHz band, improving the antenna's tolerance to frequency variations within this band; second, increasing the gain in this band, further enhancing the antenna's radiation efficiency and long-distance transmission capabilities. This frequency-specific optimization ensures the antenna's performance advantages in high-frequency applications, and is particularly important for high-speed communication systems such as 4G, which require high gain and wide bandwidth.

[0014] According to the present invention, a high-gain dual-polarized parabolic antenna system achieves a voltage standing wave ratio (VSWR) of less than or equal to 1.7 within the 900MHz to 1800MHz frequency band. Specifically, the solution defines the voltage standing wave ratio (VSWR), a key performance indicator of the antenna system across the entire dual-band (900MHz to 1800MHz), and stipulates that its value must be less than or equal to 1.7. This is a quantitative requirement for the antenna's impedance matching performance, comprehensively reflecting the results of all previous structural and parameter optimizations. It can be understood that the antenna system of the present invention achieves a voltage standing wave ratio (VSWR) of less than or equal to 1.7 across the wide frequency band of 900MHz to 1800MHz. This excellent VSWR directly demonstrates that the antenna has a high impedance match with the feeder throughout the entire operating frequency band. Good impedance matching ensures minimal reflection loss during the transmission of RF signal energy from the transmission line to the antenna, thereby maximizing the radiation of transmitted power or the reception of electromagnetic wave energy. A low VSWR value indicates high power transmission efficiency, low insertion loss, and stable signal transmission. This is crucial for wireless communication systems that require broadband compatibility and high efficiency. It ensures stable connection and efficient transmission of signals in different frequency bands and communication standards, reduces power consumption and damage to equipment at the transmitter, and improves the reliability of the entire system.

[0015] Based on the above, the present invention also discloses a performance maintenance method for a high-gain dual-polarized parabolic antenna system, comprising the following steps: a sensor deployment step: integrating a current monitoring sensor, a port characteristic monitoring sensor, and an environmental parameter sensor on a feed unit; a data acquisition configuration step: configuring a data acquisition unit, establishing a connection relationship with each sensor, and setting a data acquisition period and parameters; a data acquisition and processing step: collecting the output signal of each sensor according to a set period through the data acquisition unit, filtering and digitizing the collected signal; a data output step: transmitting the processed data to an external processing unit through a communication interface or storing it in a local storage medium; wherein the sensor deployment step provides a signal source basis for the data acquisition configuration step, the parameter setting in the data acquisition configuration step controls the execution timing of the data acquisition and processing step, and the output result of the data acquisition and processing step is externally transmitted through the data output step. Specifically, the method first integrates three types of key sensors on the feed unit: a current monitoring sensor for real-time sensing of the RF current state of each metal vibrator arm, a port characteristic monitoring sensor for monitoring the reflection coefficient and mutual coupling characteristics of the polarization port, and an environmental parameter sensor for obtaining external condition information such as temperature and humidity. Subsequently, the data acquisition unit is configured, establishing the sensor network connection topology and the timing control mechanism for data acquisition. The data acquisition and processing phase automatically collects sensor signals according to a preset cycle and performs filtering and digital preprocessing to ensure data quality. Finally, the processed data is transmitted to an external processing unit or local storage via a standardized communication interface, providing a data foundation for subsequent intelligent analysis. The technical benefits are as follows: From a systems engineering perspective, the establishment of this infrastructure first achieves comprehensive sensing capabilities for key feed parameters. By integrating microsensors on the PCB, the installation complexity and reliability issues of traditional external monitoring equipment are avoided, while ensuring accurate and real-time monitoring. The unified configuration of the data acquisition unit and the periodic collection mechanism ensure the time synchronization and format standardization of multi-source heterogeneous data, laying a high-quality data foundation for subsequent feature extraction and pattern recognition. The filtering and digital processing phases effectively suppress noise interference in the RF environment and improve the signal-to-noise ratio of the sensor data. The standardized data output interface ensures system scalability and compatibility with higher-level processing systems. Overall, this approach establishes a highly accurate, reliable, and scalable feed status sensing platform, providing a solid technical foundation for intelligent antenna health management.

[0016] According to the present invention, a performance maintenance method for a high-gain dual-polarized parabolic antenna system further includes: extracting feed source health status characteristic parameters from the output data of the data acquisition and processing, wherein the characteristic parameters include: calculating the current amplitude deviation of each metal dipole arm, obtained by the absolute value of the difference between the current monitored current amplitude and the reference current amplitude; calculating the current phase deviation of each metal dipole arm, obtained by the absolute value of the difference between the current monitored current phase and the reference current phase; extracting the percentage change of the mutual coupling coefficient amplitude and the mutual coupling coefficient phase drift between polarization ports; calculating the mutual coupling fingerprint deviation metric value based on the mutual coupling coefficient amplitude change percentage and phase drift; and combining the current amplitude deviation, current phase deviation, mutual coupling coefficient change parameter, mutual coupling fingerprint deviation metric value and environmental parameters to construct a comprehensive feature vector. In summary, this solution realizes the intelligent conversion from raw sensor data to health status characteristics based on the aforementioned data acquisition. This solution constructs a multi-dimensional health feature extraction algorithm for the electromagnetic characteristics of the dual-polarized feed source. Specifically, by calculating the deviation of the current parameters (amplitude and phase) of each metal dipole arm from a healthy baseline, the performance changes of individual dipoles are quantified. Simultaneously, the amplitude change percentage and phase drift of the mutual coupling coefficient between polarization ports are extracted, establishing the concept of a "mutual coupling fingerprint" and calculating the corresponding deviation metric. This scheme organically integrates individual dipole characteristics, mutual coupling characteristics, and environmental parameters to construct a comprehensive feature vector that comprehensively characterizes the health of the feed source. Its technical benefits lie in: From a signal processing theory perspective, this feature extraction scheme effectively maps high-dimensional sensor data into a low-dimensional feature space. The calculation of current amplitude and phase deviations directly reflects changes in the resonant characteristics of each dipole arm, sensitively detecting performance degradation caused by physical damage, connection degradation, or material aging. The introduction of the mutual coupling fingerprint feature is the innovation of this scheme. It leverages the stability and uniqueness of the coupling relationship between polarization ports in a dual-polarization antenna. By monitoring changes in the coupling characteristics, anomalies in the internal electromagnetic environment can be effectively identified. The comprehensive calculation of the deviation metric improves the sensitivity and accuracy of fault detection by weighted fusion of amplitude and phase information. The construction of a comprehensive feature vector enables the fusion of multi-source information, enhancing the integrity and robustness of health status representation. This solution significantly improves the ability to extract effective health information from raw data, providing high-quality feature input for subsequent intelligent diagnosis, while also reducing data dimensionality and improving processing efficiency.

[0017] According to the present invention, a performance maintenance method for a high-gain dual-polarized parabolic antenna system also includes: constructing a health diagnostic model using a machine learning algorithm; training the diagnostic model using electromagnetic simulation data and laboratory accelerated aging test data to establish a mapping relationship from the comprehensive feature vector to health scores, fault types, and fault locations; deploying the trained diagnostic model to a data acquisition unit or edge computing unit; and inputting the comprehensive feature vectors obtained in real time into the diagnostic model to output health scores for each metal dipole arm, the overall health score of the feed source, fault mode identification results, and fault location information. In summary, this solution builds an artificial intelligence-based health diagnostic system based on the comprehensive feature vectors provided above. This solution utilizes a variety of mature machine learning algorithms, including support vector machines, random forests, gradient boosting trees, or neural networks, to accommodate diagnostic tasks of varying complexity. Training data is derived from electromagnetic simulation and laboratory accelerated aging tests, ensuring the model's comprehensive coverage of various fault modes. By establishing a nonlinear mapping relationship from feature vectors to health scores, fault types, and fault locations, an intelligent conversion from data to diagnostic conclusions is achieved. The model deployment utilizes an edge computing architecture, enabling local real-time inference and outputting multi-level diagnostic information, including individual oscillator health scores, overall health scores, fault pattern identification, and precise location. Its technical benefits are as follows: From a machine learning perspective, this intelligent diagnostic system represents a technological leap from traditional rule-based fault detection to data-driven pattern recognition. A multi-algorithm selection strategy ensures adaptability to diverse fault characteristic patterns. Support vector machines are suitable for small sample sizes and high-dimensional problems, random forests offer excellent generalization, gradient boosting trees excel at handling nonlinear relationships, and neural networks possess robust complex pattern learning capabilities. Training with a combination of simulation and experimental data ensures the model's theoretical integrity and practical applicability. The edge computing deployment strategy enables low-latency real-time diagnosis, avoiding data transmission delays and security risks. The multi-level output structure provides comprehensive diagnostic information, spanning individual to overall, qualitative to quantitative, and detection to location. This solution significantly improves the accuracy, real-time nature, and intelligence of fault diagnosis, enabling a shift from reactive maintenance to proactive prediction, and providing reliable technical support for precision and predictive maintenance.

[0018] According to the present invention, a performance maintenance method for a high-gain dual-polarized parabolic antenna system also includes: initiating polarization intelligent reconstruction when the overall health score of the feed source falls below a preset threshold; identifying harmful coupling paths that cause polarization isolation degradation based on the fault mode identification results and fault location information; estimating the complex coefficients of cross-polarization leakage and calculating compensation parameters for suppressing the harmful coupling paths; performing cross-compensation processing on the original polarization port signals by adjusting the amplitude weights and phase weights of the signals transmitted from the two polarization ports; running an optimization algorithm to search for the optimal complex weight combination to maximize a comprehensive utility function that includes signal-to-interference-plus-noise ratio, cross-polarization leakage power, and health deviation; and applying the calculated optimized weight parameters to the base station digital signal processing link to dynamically adjust the baseband signals fed to the antenna polarization ports. In summary, this solution implements an adaptive performance recovery mechanism for the antenna system based on the aforementioned intelligent diagnosis results. When the diagnostic system detects that the health score falls below a preset threshold, the solution initiates the polarization intelligent reconstruction process. First, based on the fault mode and location information, harmful coupling paths that cause polarization isolation degradation are identified, and the complex characteristic parameters of cross-polarization leakage are estimated. Targeted compensation parameters are then calculated, and digital-domain signal preprocessing techniques are used to precisely adjust the amplitude and phase of the transmitted signals at both polarization ports. This solution employs a multi-objective optimization algorithm, comprehensively considering factors such as signal quality, cross-polarization suppression, and health deviation to search for the optimal complex weight combination. The optimization results are ultimately applied to the base station's digital signal processing link, enabling dynamic adjustment of the baseband signal fed to the antenna. The technical benefits of this approach are as follows: From the perspective of adaptive signal processing theory, this polarization reconstruction scheme enables active performance compensation even when physical defects cannot be repaired. The identification of harmful coupling paths is based on the reciprocity principle in electromagnetic field theory. By analyzing the impact of the fault on the field distribution, the root cause of performance degradation can be precisely located. Estimation of the complex coefficients of cross-polarization leakage establishes a channel model under fault conditions, providing a theoretical basis for the compensation algorithm. Digital-domain signal preprocessing techniques draw on the precoding concept used in MIMO systems, actively introducing reverse compensation at the transmitter to offset harmful effects at the receiver. The introduction of a multi-objective optimization algorithm ensures a balance between different performance metrics during performance recovery, avoiding the potential system performance imbalance that can result from optimizing a single metric. Real-time adjustment of the base station's digital signal processing chain enables software-defined antenna characteristics, greatly enhancing the system's flexibility and adaptability. This solution significantly improves the antenna system's service continuity and performance maintenance capabilities during fault conditions, enabling an innovative solution from hardware failure to software compensation, providing a critical technical guarantee for the high-reliability operation of the communication system.

[0019] The present invention proposes a high-gain dual-polarized parabolic antenna system that utilizes an etching process to form an integrated dual-band (900 MHz and 1800 MHz) dual-polarized metal oscillator unit on a single printed circuit board (PCB). This PCB-integrated design significantly reduces the number of components and assembly steps compared to traditional feed sources that utilize multi-layer boards, multiple components, or complex three-dimensional structures. This simplifies the feed source's manufacturing and assembly processes, effectively reducing production costs while improving product manufacturing consistency and reliability. Furthermore, because each oscillator in the metal oscillator unit integrates a low-frequency radiating portion for the first frequency band and a high-frequency radiating portion for the second frequency band, the antenna achieves excellent impedance matching in both the 900 MHz and 1800 MHz target frequency bands. This design, which integrates dual-frequency resonant characteristics into a single oscillator, avoids the complex coupling issues between radiating elements in different frequency bands. This enables the antenna to maintain an excellent voltage standing wave ratio (VSWR) within a wide frequency band of 900MHz-1800MHz, especially across two frequency bands (900MHz and 1800MHz) that are far apart. For example, a VSWR of ≤1.7 can be achieved. This ensures high efficiency and stability of RF signal transmission and reduces power loss. In addition, the metal vibrator unit includes at least two groups of orthogonally arranged vibrators, forming radiation in two orthogonal polarization directions. This orthogonal layout is a direct and effective way to achieve dual-polarization function, so that there is a high degree of isolation between the two polarization channels. This enables the antenna to better support MIMO (multiple input multiple output) technology in modern wireless communication systems, thereby effectively improving the system's channel capacity, spectrum utilization and communication reliability, and meeting the needs of high data rate transmission. Furthermore, the parabolic reflector used in the present invention has an inherent beamforming effect that converges the electromagnetic waves radiated from the feed source into a high-gain directional beam. At the same time, the dual-band (900MHz and 1800MHz) capability of the feed enables it to support both frequency bands simultaneously, so the antenna can provide excellent coverage and anti-interference capabilities, and is particularly suitable for scenarios that require rapid deployment and high-density user access, such as emergency communications and temporary base stations for large-scale events. In addition, being compatible with both the 900MHz and 1800MHz frequency bands enables the antenna to adapt to and support the coexistence of multiple mobile communication networks such as 2G / 3G / 4G, thereby improving the flexibility and practicality of the system. In summary, the present invention, through the innovative design of integrating dual-frequency dual-polarization metal oscillator units on a PCB, significantly improves the electrical performance of the antenna within a wide frequency band while achieving a compact structure and low-cost manufacturing, providing an efficient and flexible solution for modern wireless communications. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art.

[0021] Figure 1 It is a schematic diagram of the antenna feed structure of the present invention;

[0022] Figure 2 It is a schematic diagram of the layout of the metal vibrator unit of the present invention. DETAILED DESCRIPTION

[0023] The high-gain dual-polarized parabolic antenna system of the present invention will be described in detail below in conjunction with the present application. The embodiments disclosed in this specification are only used to exemplify the present invention and are not used to limit the scope of protection of the present invention.

[0024] The present invention provides a high-gain dual-polarization parabolic antenna system, referring to Figure 1 The high-gain dual-polarized parabolic antenna system of this embodiment primarily includes a parabolic reflector and a feed. The feed is positioned at or near the focal point of the parabolic reflector. This placement ensures that the spherical or quasi-spherical waves radiated by the feed are effectively reflected by the parabolic reflector, forming a highly directional plane beam, thereby achieving high gain characteristics for the antenna system. The parabolic reflector utilizes a parabolic antenna reflector structure well known in the art.

[0025] The antenna system also includes an antenna base, which supports the entire antenna system and provides a stable mounting foundation. The feed is connected to the parabolic reflector via a feed flange mount. Made of precision-machined metal, the feed flange mount ensures precise positioning and a secure connection between the feed and reflector. An aluminum tube connects the feed flange mount to the antenna base, providing mechanical support and accommodating the feed cable for efficient signal transmission.

[0026] To protect the feed's electronic components from environmental influences, an antenna housing is installed outside the feed. Made of low-dielectric-constant engineering plastic, the housing offers excellent weather resistance and electromagnetic transparency. Its dielectric constant is typically controlled between 2.0 and 3.0, and its loss tangent is less than 0.01, minimizing any impact on antenna performance. The housing's design also incorporates waterproof sealing requirements, achieving an IP65 rating and above.

[0027] The core component of the feed is a printed circuit board (PCB), manufactured using standard PCB manufacturing processes, ensuring excellent dimensional accuracy and manufacturing consistency. An etching process is employed on the PCB to form the metal oscillator unit B, which constitutes the feed's primary radiating structure. The technical benefits of this PCB etching process include significantly reduced manufacturing costs, improved production efficiency, and ensured the accuracy and repeatability of the oscillator structure.

[0028] Reference Figure 2The metal vibrator unit B includes at least two groups of orthogonally arranged vibrators, specifically a first polarized vibrator group and a second polarized vibrator group. The first polarized vibrator group consists of a first metal vibrator arm B1 and a second metal vibrator arm B3 symmetrically arranged therewith. The first metal vibrator arm B1 and the second metal vibrator arm B3 together form radiation in one polarization direction. The second polarized vibrator group consists of a third metal vibrator arm B2 and a fourth metal vibrator arm B4 symmetrically arranged therewith. The second polarized vibrator group is arranged on the PCB plane rotated 90 degrees relative to the first polarized vibrator group, forming radiation in a polarization direction orthogonal to the first polarized vibrator group.

[0029] like Figure 2 As shown, the four metal dipole arms B1, B2, B3, and B4 form a symmetrical petal-like or butterfly-wing-like geometry on the PCB. This unique geometric design not only aesthetically pleasing but also, more importantly, optimizes current distribution and radiation characteristics. The edge profile of each dipole arm is carefully designed, with smooth arc transitions to avoid current concentration and unwanted high-frequency harmonics that can be generated by sharp corners. A circular feed area is located in the center of the PCB, integrating the feed network and impedance matching circuitry.

[0030] This orthogonal arrangement design achieves high-quality dual-polarization functionality, supports MIMO technology applications, and effectively enhances channel capacity and communication reliability. The symmetry of the four dipole arms ensures omnidirectional radiation characteristics in the horizontal plane while maintaining excellent polarization purity.

[0031] Each of the first, second, third, and fourth metal dipole arms B1, B3, B2, and B4 includes a low-frequency radiating portion and a high-frequency radiating portion. Specifically, the first metal dipole arm B1 includes a low-frequency radiating portion primarily for radiating in the 900 MHz frequency band and a high-frequency radiating portion primarily for radiating in the 1800 MHz frequency band; the second metal dipole arm B3 includes a low-frequency radiating portion and a high-frequency radiating portion; the third metal dipole arm B2 includes a low-frequency radiating portion and a high-frequency radiating portion; and the fourth metal dipole arm B4 includes a low-frequency radiating portion and a high-frequency radiating portion.

[0032] The low-frequency radiator is a long, wide arm-like structure, approximately one-quarter the operating wavelength of the 900MHz band, suitable for achieving good radiation efficiency in the 900MHz band. The high-frequency radiator is a shorter, narrower arm-like structure, approximately one-quarter the operating wavelength of the 1800MHz band, to meet the radiation requirements of the 1800MHz band. The high-frequency radiator is nested or integrated within or near the low-frequency radiator structure. This integrated design enables a single dipole arm to achieve effective radiation in two different frequency bands, significantly improving spectrum utilization.

[0033] By precisely adjusting the length, width, shape, spacing, and overall opening angle of the low- and high-frequency radiators, the antenna's input impedance in the 900MHz and 1800MHz bands can be optimized to approach the feeder's characteristic impedance of 50 ohms, achieving excellent impedance matching. The arm opening angle has been optimized to ensure optimal impedance matching in both frequency bands.

[0034] Furthermore, the feed source also includes a guide plate D. The guide plate D is formed on the PCB through an etching process and adopts the same manufacturing process as the metal vibrator unit B, ensuring the integration of the structure. Figure 2 As shown, the director D is a circular structure located above or in front of the center area of ​​the metal oscillator unit B, forming a cooperative working relationship with the high-frequency radiator. The circular design of the director D has the advantage of symmetry, which can provide uniform electromagnetic coupling enhancement to the high-frequency radiator in all four directions. The size of the director D and its distance from the center of the metal oscillator unit B are optimized to effectively improve performance in the 1800MHz frequency band. The technical effect of this circular director design is reflected in the fact that the director D can effectively expand the operating bandwidth of the 1800MHz frequency band and increase the forward gain in this frequency band. At the same time, the circular structure ensures uniform improvement of radiation characteristics in all directions, avoiding directional deviation. The operating principle of the director D is based on the secondary radiation effect of electromagnetic waves. When the high-frequency radiator radiates electromagnetic waves in the 1800MHz frequency band, the director D, as a passive radiating element, generates induced current and forms secondary radiation. By precisely controlling the size and position of the director D, the secondary radiation and the primary radiation are superimposed in phase in the forward direction and canceled out in anti-phase in the backward direction, thereby achieving the shaping of the radiation pattern and the improvement of the gain.

[0035] The feeding structure and working principle of this embodiment are as follows:

[0036] The antenna system uses differential feeding. For the first polarization dipole group, the inner conductor of the external feed cable connects to the feed point of the first metal dipole arm B1, and the outer conductor connects to the feed point of the second metal dipole arm B3. The feed point is located at the base of the dipole arm, near the center of the PCB. The first and second metal dipole arms B1 and B3 form a half-wavelength symmetrical dipole excited by the feed cable, radiating electromagnetic waves simultaneously in the 900MHz and 1800MHz bands.

[0037] The second polarization element group is fed similarly to the first polarization element group, using independent feed cables for excitation. The independent feeding design of the two polarization element groups ensures good isolation between the dual-polarization signals, meeting the polarization isolation requirements of MIMO systems.

[0038] The low-frequency and high-frequency radiators integrated into each arm of the metal oscillator unit B work together through electromagnetic coupling. In the 900MHz band, the low-frequency radiator plays a dominant role, while the high-frequency radiator, due to its relatively small size, has a minimal impact on radiation in this band. In the 1800MHz band, the high-frequency radiator plays a dominant role. While the low-frequency radiator is larger, its impact on the 1800MHz band is effectively controlled through precise geometric design, without causing significant performance degradation.

[0039] Through the above-mentioned structural design, especially the specific geometric shape, size, relative layout of the metal oscillator unit B on the PCB and the synergistic effect of the guide plate D, the technical effects achieved by the antenna system of the present invention include: 1. Good impedance matching with a voltage standing wave ratio VSWR less than or equal to 1.7 in a wide frequency band from 900MHz to 1800MHz. This indicator ensures the stability and efficiency of signal transmission. 2. Excellent dual-polarization performance, small gain difference in the two orthogonal polarization directions, high polarization isolation, meeting the requirements of modern communication systems for dual-polarization antennas. 3. High gain characteristics: high gain can be achieved in both 900MHz and 1800MHz frequency bands, meeting the application requirements of high-gain antennas. The diameter and focal diameter ratio of the parabolic reflector are designed according to application requirements to optimize the matching of gain and beam width. 4. Good directivity: It has a narrow beam width in both operating frequency bands, and the sidelobe level is effectively controlled, reducing interference to adjacent base stations.

[0040] Based on the antenna system of the above embodiment, this embodiment also provides a performance maintenance method for a high-gain dual-polarized parabolic antenna system. This method ensures that the antenna can maintain the availability and quality of the communication link to the maximum extent possible when the feed unit performance deteriorates or fails through real-time monitoring, intelligent diagnosis, and adaptive reconstruction. The specific implementation content is as follows:

[0041] 1. Execute the sensor deployment steps, which are as follows:

[0042] On the feed unit's printed circuit board (PCB), sensors are integrated and deployed for each metal dipole arm that constitutes the dual-polarized antenna. Specifically, micro sensors are integrated into the current paths of the first polarized dipole group consisting of the first metal dipole arm B1 and the second metal dipole arm B3, and the second polarized dipole group consisting of the third metal dipole arm B2 and the fourth metal dipole arm B4, using a PCB etching process. Specifically, on the feed unit's PCB, for each of the dual-polarized metal dipole arms B1, B2, B3, and B4, a micro current sensing structure (optionally using Rogowski coils and optional micro strain gauges) is integrated to accurately monitor the RF current amplitude, phase, and micro-structural deformation of each dipole arm. The current monitoring sensor is deployed using a micro Rogowski coil or an equivalent current sensing structure, and the sensor monitors the RF current flowing through each dipole arm in a non-contact manner. The Rogowski coil works on the principle of electromagnetic induction. When radio frequency current flows through the vibrator arm, an induced electromotive force is generated in the coil. The amplitude of the induced electromotive force is proportional to the amplitude of the measured current, and the phase maintains a fixed relationship with the phase of the measured current, thereby achieving accurate monitoring of the radio frequency current.

[0043] The port characteristic monitoring sensors are deployed at the two physical polarization ports of the feed source to achieve monitoring. Among them, the V polarization port is formed by the convergence of the B1 and B3 feed points, and the H polarization port is formed by the convergence of the B2 and B4 feed points. The reflection coefficient of each port is monitored through the RF detection port at each port or through an external micro directional coupler. and , and the isolation or coupling coefficient between ports and The directional coupler uses a microstrip structure with a coupling factor of -20dB to -30dB to minimize impact on the main signal path while providing sufficient monitoring signal strength. These parameters directly reflect the antenna's voltage standing wave ratio (VSWR) and inter-polarization coupling.

[0044] In addition, specific structures are etched near the key stress points of the vibrator arm to serve as micro-strain gauges to sense tiny deformations caused by physical damage such as microcracks and deformation. Environmental parameter sensors include temperature sensors and humidity sensors, which are integrated on the PCB or inside the feed cavity to monitor the feed working environment parameters. The temperature sensor uses a platinum resistance RTD or thermistor NTC with a measurement range of -40°C to +85°C and an accuracy of ±0.5°C. The humidity sensor uses a capacitive humidity sensor with a measurement range of 0% to 100% relative humidity and an accuracy of ±2%RH.

[0045] This step aims to achieve more comprehensive data acquisition, specifically multi-dimensional, refined real-time perception of key electrical parameters such as current, S parameters, and physical parameters such as deformation, temperature and humidity inside the feed unit. Through high-precision sensing technology, early and subtle performance changes of the feed unit due to aging, environmental impact or minor damage are captured, providing the original data basis for subsequent health diagnosis and fault warning. In addition, the physical state and electromagnetic properties of the feed are converted into quantifiable and analyzable digital signals, laying the foundation for the introduction of AI-based intelligent algorithms. Independent monitoring of each vibrator arm makes it possible to locate subsequent faults to specific components.

[0046] 2. Execute the data collection configuration steps, which are implemented as follows:

[0047] The data acquisition unit (DCU) is implemented using an onboard microcontroller (MCU) or a compact data acquisition module. The MCU uses a 32-bit ARM Cortex-M4 architecture processor with floating-point unit (FPU) and digital signal processor (DSP) functionality. It operates at 168MHz and has 512KB of flash memory and 192KB of RAM.

[0048] The detailed process for configuring the data acquisition unit involves integrating an existing onboard microcontroller unit (MCU) or connecting a compact data acquisition module (DCU) to the feed. The DCU is responsible for periodically collecting sensor data, with a configurable acquisition period, for example, every 1 second to several minutes. The connection between the data acquisition unit and the sensors is established as follows: The current monitoring sensor is connected via a high-precision analog-to-digital converter (ADC). The ADC features 16-bit resolution, a sampling rate of 1MSPS, differential inputs, and a programmable gain amplifier (PGA). The port characteristic monitoring sensor is connected to the ADC via an RF switch and downconverter to monitor the amplitude and phase of the reflection coefficient and mutual coupling coefficient. The environmental parameter sensor is connected to the MCU via the I2C bus to support digital output and temperature compensation.

[0049] The specific data collected regularly by the DCU include: the current amplitude of each pendulum arm Bk (k=1,2,3,4) and phase The voltage standing wave ratios VSWR_V and VSWR_H of the two polarized ports can be obtained from the reflection coefficients of the corresponding ports. and The mutual coupling coefficient between polarization ports is obtained by converting the modulus value of and The amplitude and phase of the polarization signal; and the ambient temperature T and humidity H. The voltage standing wave ratio is calculated using the standard formula. The voltage standing wave ratio VSWR_V of the V polarization port is calculated by the relationship VSWR_V = (1 + | |) / (1 - | |) calculation, where | | is the reflection coefficient of the V polarization port The module value of VSWR_H is calculated in the same way, using The modulus value of .

[0050] The data collection cycle parameters are set to configurable mode, with the default cycle ranging from 1 second to several minutes. Specifically, the collection cycle for current monitoring data is set to 1 second to capture rapid changes in RF current; the collection cycle for port characteristic monitoring data is set to 5 seconds to balance monitoring accuracy and system resource consumption; and the collection cycle for environmental parameters is set to 60 seconds because ambient temperature and humidity change relatively slowly.

[0051] The technical benefit of this step is to establish a complete data acquisition architecture, ensuring that all types of sensor data can be collected according to preset timing and accuracy requirements, providing a stable and reliable data source for subsequent feature extraction and AI diagnosis. Through standardized data acquisition configuration, unified management and coordinated collection of multi-source heterogeneous sensor data are achieved.

[0052] 3. Execute the data acquisition and processing steps, which are specifically implemented as follows:

[0053] The detailed process of data acquisition and processing involves the DCU proactively querying all sensors at a preset interval, collecting raw data including arm current, port S-parameters, and ambient temperature and humidity. The DCU performs preliminary signal conditioning on the collected raw data, including filtering, amplification if necessary, and analog-to-digital conversion. The processed data is temporarily stored in the DCU's local memory or transmitted in real time to higher-level processing units such as edge computing units or base station BBUs via standard communication interfaces such as SPI, I2C, UART, and Ethernet.

[0054] The process of filtering and digitizing the collected signal includes: first, anti-aliasing low-pass filtering of the original analog signal, using an 8th-order Butterworth filter with a cutoff frequency set to 40% of the ADC sampling frequency; then, analog-to-digital conversion is performed through a 16-bit ADC with a conversion accuracy of 65,536 discrete levels; then, the digitized signal is digitally filtered, using a finite impulse response (FIR) filter to remove high-frequency noise and power frequency interference; finally, amplitude and phase extraction and calibration are performed, using a digital signal processing algorithm to extract the signal's amplitude and phase information, and amplitude and phase correction is performed based on pre-stored calibration data.

[0055] The technical effect of this step is to achieve high-precision, low-noise data acquisition. Through multi-stage filtering and calibration processing, the accuracy and reliability of the monitoring data are ensured, laying the data quality foundation for subsequent feature extraction and AI diagnosis. Through a standardized data processing process, the physical state and electromagnetic properties of the feed are converted into quantifiable and analyzable digital signals.

[0056] 4. Execute the data output steps, which are specifically implemented as follows:

[0057] Processed data is transmitted externally via standard communication interfaces. These include SPI, I2C, UART, and Ethernet. The SPI interface is used for high-speed data transmission, with a transfer rate of up to 10 Mbps; the I2C interface is used for low-speed control signal transmission; the UART interface provides asynchronous serial communication; and the Ethernet interface supports the TCP / IP protocol stack for connection to upper-layer networks.

[0058] Data storage utilizes a dual-mode design: local storage and network transmission. Local storage uses 32MB of non-volatile flash memory and employs a ring buffer management system, automatically overwriting the oldest data when storage space is full. Network transmission utilizes real-time transmission mode, with processed data transmitted in real time to the edge computing unit or base station BBU via an Ethernet interface.

[0059] The data format uses the standardized JSON format and includes information such as timestamp, sensor ID, data type, value, and unit. The data packet structure is designed as: {"timestamp": "YYYY-MM-DD HH:MM:SS", "sensor_id": "Bk_current", "data_type": "amplitude", "value": xx.xx, "unit": "mA"}.

[0060] The technical benefit of this step is to achieve standardized data transmission and storage, ensuring data integrity and traceability, and providing a foundation for subsequent distributed processing and long-term trend analysis. Support for multiple communication interfaces enables compatibility and interoperability with different types of external systems.

[0061] 5. Execute the steps of extracting health status characteristic parameters, which are specifically implemented as follows:

[0062] The detailed feature extraction process involves a higher-level processing unit or a DCU with sufficient computing power receiving the raw sensor data from step 1. For each individual arm, the deviation between the current amplitude and phase and the healthy baseline is calculated. For the polarization port characteristics, the change in current VSWR and the key "mutual coupling fingerprint" feature are calculated. A comprehensive "mutual coupling fingerprint deviation metric" (MFDD) is calculated based on a specific formula to quantify the degree of deviation from the overall coupling state. The individual arm characteristics, mutual coupling fingerprint features, and environmental parameters are integrated into a multidimensional comprehensive feature vector F.

[0063] Extract characteristic parameters that can effectively characterize the health status of the feeder from the output data of data acquisition and processing. Extraction of health characteristics of individual vibrator arms includes current deviation calculation and characteristic frequency offset analysis.

[0064] The current amplitude deviation of each metal vibrator arm is calculated by the formula:

[0065] ;

[0066] The detailed meaning of each parameter in the formula is as follows: The absolute value of the current amplitude deviation of the kth vibrator arm (Bk), in milliamperes (mA), represents the degree of deviation of the current amplitude from the healthy reference state. is the RF current amplitude currently monitored by the k-th vibrator arm, in milliamperes (mA), obtained in real time by a current monitoring sensor such as a Rogowski coil. is the RF current amplitude of the kth vibrator arm in the healthy reference state, in milliamperes (mA), obtained through factory calibration or statistical averaging during healthy operation. k is the index of the vibrator arm, with values ​​of 1, 2, 3, and 4 corresponding to the first, second, third, and fourth metal vibrator arms, respectively:

[0067] The current phase deviation of each metal vibrator arm is calculated by the formula:

[0068] ;

[0069] The detailed meaning of each parameter in the formula is as follows: The absolute value of the current phase deviation of the kth pendulum arm, expressed in degrees or radians, represents the degree of deviation of the current phase from the healthy reference state. is the RF current phase currently monitored by the k-th dipole arm, in degrees or radians, obtained in real time by the phase detection circuit. is the RF current phase of the kth dipole arm in the healthy reference state, in degrees or radians, which serves as a reference.

[0070] The reference value is derived from the statistical average of factory antenna calibration or healthy operation, using least squares fitting to obtain a stable reference. Phase deviation calculations take into account the periodicity of 2π and use the principle of minimum angle difference.

[0071] The characteristic frequency offset is monitored by analyzing the response of the vibrator arm current or port S parameters under a frequency sweep test. This test can be completed by injecting the test signal during non-service time slots. The resonant frequency offset is calculated using the formula:

[0072] ;

[0073] The detailed meaning of each parameter in the formula is as follows: is the resonant frequency offset of the kth vibrator arm, in Hertz (Hz) or Megahertz (MHz). A positive value indicates an upward shift in frequency, and a negative value indicates a downward shift in frequency. The resonant frequency value of the kth vibrator arm currently monitored by frequency sweep testing or the like, in Hertz (Hz) or Megahertz (MHz). is the resonant frequency value of the kth vibrator arm in a healthy reference state, in Hertz (Hz) or Megahertz (MHz), which serves as a reference for frequency deviation.

[0074] The percentage change in the mutual coupling coefficient amplitude between polarization ports is calculated using the formula:

[0075] ;

[0076] The detailed meaning of each parameter in the formula is as follows: is the percentage change in the mutual coupling coefficient amplitude from the V-polarized port to the H-polarized port, which represents the relative change in the coupling strength from the V to H direction. is the currently monitored mutual coupling coefficient amplitude from the V-polarized port to the H-polarized port, which is a dimensionless linear value or expressed in decibels (dB). is the magnitude of the mutual coupling coefficient from the V-polarized port to the H-polarized port in the healthy baseline state, which serves as the reference for calculating the percentage change. It is the ratio of the signal detected at the H-polarization port to the incident signal at the V-port when the signal is fed from the V-polarization port, and represents the degree of coupling or leakage from V-polarization to H-polarization.

[0077] The phase shift of the mutual coupling coefficient is calculated by the formula:

[0078] ;

[0079] The detailed meaning of each parameter in the formula is as follows: is the phase shift of the mutual coupling coefficient from the V-polarized port to the H-polarized port, in degrees or radians. The phase of the mutual coupling coefficient from the V-polarized port to the H-polarized port currently monitored, in degrees or radians. is the phase of the mutual coupling coefficient from the V-polarized port to the H-polarized port in the healthy reference state, in degrees or radians.

[0080] Similarly, for the mutual coupling coefficient in the H to V direction, there is a corresponding formula:

[0081] ;

[0082] ;

[0083] The mutual coupling fingerprint deviation metric MFDD is calculated using a weighted synthesis method:

[0084] ;

[0085] The detailed meaning of each parameter in the formula is as follows:

[0086] Mutual-coupling Fingerprint Deviation Degree is a measure of mutual coupling fingerprint deviation, which is an indicator that comprehensively quantifies the degree to which the mutual coupling state of the feed source deviates from the baseline. A larger value indicates a more serious deviation. is the weighting factor of the mutual coupling coefficient amplitude change, ranging from 0 to 1, and is used to adjust the importance of the amplitude change in the comprehensive evaluation. is the weighting factor for the phase change of the mutual coupling coefficient, ranging from 0 to 1, usually wamp + wphase = 1, and is used to adjust the importance of phase change in the comprehensive evaluation. is the absolute value of the percentage change in the mutual coupling coefficient amplitude from the V-polarized port to the H-polarized port. is the absolute value of the percentage change in the mutual coupling coefficient amplitude from the H-polarized port to the V-polarized port. is the absolute value of the mutual coupling coefficient phase shift from the V-polarized port to the H-polarized port, in degrees or radians. The absolute value of the mutual coupling coefficient phase shift from the H-polarized port to the V-polarized port, expressed in degrees or radians. is a phase normalization constant, for example, 180 degrees or π radians, used to normalize the phase deviation to a comparable range such as 0-1.

[0087] The introduction of MFDD is necessary because simple single-parameter changes may not fully reflect complex coupling state changes. MFDD provides a comprehensive metric that, by weighting amplitude and phase changes, can more sensitively capture changes in the internal electromagnetic environment caused by damage, deformation, or connection degradation of the transducer arm. These changes often manifest themselves first in the coupling characteristics between ports.

[0088] The construction of the comprehensive feature vector combines the above individual vibrator arm health characteristics, mutual coupling fingerprint characteristics and environmental parameters into a multidimensional vector

[0089] ;

[0090] The detailed meaning of each parameter in the comprehensive feature vector F is as follows: F is the comprehensive feature vector of the feed state, which is a multidimensional array containing 19 elements and comprehensively describes the health status of the feed. , , , are absolute values ​​of the current amplitude deviations of the first to fourth dipole arms respectively. are the absolute values ​​of the current phase deviations of the first to fourth pendulum arms respectively. are the resonant frequency shifts of the first to fourth vibrator arms respectively. is the percentage change in the mutual coupling coefficient amplitude from the V to H port. is the phase shift of the mutual coupling coefficient from V to H port. is the percentage change in the mutual coupling coefficient amplitude from the H to V port. is the phase shift of the mutual coupling coefficient from the H to V port. is the mutual coupling fingerprint deviation metric, calculated using the above formula. T is the currently monitored ambient temperature, in degrees Celsius. H is the currently monitored ambient humidity, in %RH.

[0091] The technical benefits of this step are reflected in data dimensionality reduction and information enhancement, transforming high-dimensional, raw sensor data into a set of core features that are more informative and representative of health status, reducing the processing complexity of subsequent AI models. The innovative application of the "mutual coupling fingerprint" and its proposed deviation metric, MFDD, can sensitively reflect changes in the electromagnetic environment within the feed source caused by factors such as slight structural changes and connection degradation. It is particularly sensitive to polarization performance degradation, improving diagnostic accuracy. The generated comprehensive feature vector F provides a standardized input format for the AI ​​diagnostic model. At the same time, through feature engineering, it enhances the distinguishability between different health states (healthy, slightly degraded, severely faulty), as well as different failure modes.

[0092] 6. Execute the steps for building and deploying the health diagnosis model. The specific implementation is as follows:

[0093] The detailed work process of health diagnosis includes collecting a large number of characteristic vectors of feeds in different health states, including various typical failure modes, locations and degrees, through electromagnetic simulation and laboratory accelerated aging tests in advance. and And manually label the health score and fault label. Using these labeled data sets, train one or more machine learning or deep learning models such as SVM, random forest, XGBoost, small CNN or MLP offline. The training goal is to learn the mapping relationship from the feature vector F to the health score, fault mode and fault location. The trained and optimized AI model such as quantization and pruning is deployed to the feed DCU if the computing power allows, the edge computing unit or the base station BBU. When the system is running, the processing unit extracts the real-time feature vector The data is input into the deployed AI model. The AI ​​model performs inference calculations based on the patterns it has learned internally and outputs a diagnosis of the current feed status.

[0094] The machine learning algorithm used was a multi-model fusion approach, primarily consisting of a support vector machine (SVM), a random forest (RF), and a lightweight neural network (MLP). The SVM model used a radial basis function (RBF) kernel, with the kernel parameter γ optimized via grid search. The random forest model consisted of 100 decision trees with a maximum depth of 10 layers. The multilayer perceptron (MLP) model consisted of two hidden layers, with 64 and 32 neurons, respectively, and used a Reluctant Unit (ReLU) activation function.

[0095] The training data sources include simulation data and laboratory test data. The simulation data uses the electromagnetic simulation software HFSS to simulate different types of damage scenarios, including small deformations, cracks, etching defects, local parameter changes of the dielectric substrate, etc. of one or a combination of the vibrator arms B1-B4. For each simulated failure scenario, the corresponding comprehensive feature vector is recorded. The lab accelerated aging test data is obtained by subjecting actual antenna samples to high temperature, high humidity, salt spray corrosion, vibration, and other tests. Data is collected regularly and combined with manual inspection results to annotate the feature vector Flab and its health status.

[0096] in:

[0097] The comprehensive feature vector is extracted when simulating various fault conditions through electromagnetic simulation software.

[0098] In order to conduct laboratory accelerated aging tests on actual antenna samples, data are collected at different aging or failure stages and the resulting comprehensive feature vectors are extracted.

[0099] The training dataset contains 5,000 simulation samples and 2,000 laboratory test samples, covering four levels: healthy, slightly degraded, moderately faulted, and severely faulted. The health score is scored on a scale of 0-100, with 90-100 indicating healthy, 70-89 indicating slightly degraded, 50-69 indicating moderately faulted, and 0-49 indicating severely faulted.

[0100] The model was trained using a 10-fold cross-validation method, with a training set and a test set split of 8:2. The Adam optimizer was used for training, with a learning rate of 0.001, a batch size of 32, and 200 training epochs. To prevent overfitting, early stopping and dropout regularization were employed, with a dropout ratio of 0.3.

[0101] Model deployment utilizes a quantization and optimization strategy, performing INT8 quantization on trained models, reducing model size to 25% of its original size and increasing inference speed fourfold. The deployment platform can choose to use the source DCU when computing power is sufficient, or a connected edge computing unit. The edge computing unit uses an ARM Cortex-A72 quad-core processor with a main frequency of 1.8GHz and 4GB of memory, supporting machine learning inference acceleration.

[0102] During the real-time diagnosis process, the current comprehensive feature vector Input to the deployed AI model, model output includes:

[0103] : The health score of the kth oscillator arm output by the AI ​​model, usually a value between 0-100, with a higher value indicating better health.

[0104] : The overall health score of the feed output by the AI ​​model.

[0105] : The most likely fault mode or type of the current feed identified by the AI ​​model, such as "microcracks in the B1 vibrator arm", "abnormal increase in VH polarization coupling", etc.

[0106] : The AI ​​model provides fault location information, such as which pendulum arm B1-B4, or which polarization V-Pol or H-Pol has a problem.

[0107] : When the antenna system is actually running, sensor data is collected in real time and the current comprehensive feature vector is extracted.

[0108] The confidence assessment of the output results is achieved through model integration. When the consistency of the prediction results of the three models exceeds 80%, the diagnosis result is considered credible.

[0109] The technical effect of this step is reflected in intelligent diagnosis, which enables automatic and intelligent assessment of the health status of the feed source, replacing the traditional method of relying on manual experience or regular offline testing, and improving diagnostic efficiency and accuracy. By outputting a quantitative health score, operation and maintenance personnel can intuitively understand the health status of the feed source and set multi-level warning thresholds. It can not only determine whether a fault has occurred, but also identify the type of fault with a high probability and locate it to the specific vibrator arm or polarization direction, providing a basis for precise maintenance. At the same time, trend prediction capabilities can realize predictive maintenance. By learning a large amount of data, the AI ​​model has a certain reliability against noise and small disturbances, and can adapt to environmental changes within a certain range.

[0110] 6. Execute the polarization intelligent reconstruction step, which is specifically implemented as follows:

[0111] The detailed working process of polarization intelligent reconstruction includes: when the AI ​​model diagnoses that a certain polarization direction of the feed or the overall performance has significantly degraded and the health score is lower than the preset threshold If the conventional active impedance tuning on the base station side cannot be fully restored, the polarization intelligent reconstruction mechanism is triggered. The reconstruction control logic usually obtains the AI ​​diagnosis results first at the base station BBU. 、 and current "mutual coupling fingerprint" features such as MFDD and 、 Based on this information, the system identifies the "harmful coupling path" caused by the fault and estimates the compensation coefficients required for cross-polarization cancellation, such as 、 The goal is to offset the cross-polarization components introduced by the physical defects of the antenna. The system starts the optimization algorithm according to the preset optimization target. The optimization algorithm searches and determines the best digital domain transmission signal complex weight combination within the allowed range. 、 or cross compensation coefficient 、 The base station BBU applies the calculated optimization parameters to its digital signal processing chain, dynamically adjusting the amplitude and phase of the original baseband signal fed to the two physical polarization ports of the antenna. After the reconstruction is executed, the system continuously monitors key performance indicators (KPIs) and feed sensor data to evaluate the reconstruction effect and makes iterative adjustments or triggers reconstruction again as needed.

[0112] When the feed's overall health score Below the preset threshold When the polarization intelligent reconstruction mechanism is started, the preset threshold It is set according to the system performance requirements, and the value is usually 70 points, indicating that the feed performance has been significantly degraded.

[0113] in: The preset feed health score threshold is used when AI diagnoses When the value is lower than this, the performance is considered to have degraded significantly and a reconstruction needs to be initiated.

[0114] The prerequisite is that the two physical polarization ports V-Pol and H-Pol of the feed source are connected to independent ports of the base transceiver station TRX. The digital baseband processing unit BBU or intermediate frequency radio frequency front end of the TRX has the ability to independently, accurately and dynamically adjust the amplitude and phase of the transmitted signals of the two polarization ports. In other words, the complex weight of the V-Pol signal can be controlled. and the complex weight of the H-Pol signal .

[0115] in, is the complex weight applied to the V polarization port signal. is the amplitude adjustment factor of the V polarization port signal. e is the base of the natural logarithm. j is the imaginary unit ( =). is the phase adjustment factor for the V-polarized port signal (usually in radians). is the complex weight applied to the H-polarized port signal. is the amplitude adjustment factor of the H-polarization port signal. is the phase adjustment factor for the H-polarization port signal (usually in radians).

[0116] The identification of harmful coupling paths is based on the diagnosis results of the AI ​​model and the current "mutual coupling fingerprint" characteristics. 、 Combined with the current "mutual coupling fingerprint" characteristics, particularly MFDD and its components, the primary "harmful coupling paths" that lead to performance degradation, particularly the reduction in polarization isolation (XPI), are identified. For example, if damage to the B1 transducer arm causes abnormal coupling of V-Pol energy to the H-Pol direction, then the abnormal coupling path between B1 and the H-Pol (composed of B2 / B4) is a harmful path.

[0117] The complex coefficient of cross-polarization leakage is estimated by analyzing the change of the current mutual coupling coefficient relative to the reference value. Assuming that the fault mainly affects the V polarization, causing part of its energy EV_leak to leak to the H polarization, the complex coefficient of leakage is Estimated by the formula:

[0118] ;

[0119] The detailed meaning of each parameter in the formula is as follows: is the signal component leaking from the V-polarization path to the H-polarization path, expressed as a complex number, including amplitude and phase. (alpha_VH) is the complex leakage coefficient that characterizes the leakage of the V-polarized signal into the H-polarized path. It describes the amplitude and phase relationship of the leakage signal relative to the original V-polarized intended signal. is the original signal component intended to be transmitted from the V polarization path, expressed in complex numbers.

[0120] Similarly, if H-Pol is damaged, This coefficient can be analyzed by 、 Relative to 、 The changes in are estimated by combining the fault model. (alpha_HV) is the complex leakage or coupling coefficient that represents the H-polarized intended transmitted signal that leaks or contributes to the V-polarization direction or is received by the V-polarization port due to factors such as internal non-ideal coupling within the antenna.

[0121] The calculation of compensation parameters uses a cross-coupling compensation algorithm. To offset this leakage, it is necessary to adjust the excitation weights of the healthy port or two ports. For example, to suppress the leakage from V-Pol to H-Pol, adjust the excitation weight of H-Pol. , so that it generates a compensation field that is in phase with the leakage field. More generally, the adjusted port excitation signal can be expressed as:

[0122] ;

[0123] ;

[0124] The detailed meaning of each parameter in the formula is as follows: The baseband signal is transmitted from the V polarization port after compensation adjustment. The baseband signal is transmitted for the H-polarization port after compensation adjustment. is the main path complex weight applied to the original V polarization signal, ,in is the amplitude adjustment factor, is the phase adjustment factor. is the main path complex weight added to the original H-polarized signal, ,in is the amplitude adjustment factor, is the phase adjustment factor. is the original, uncompensated V-polarized transmit baseband signal. is the original, uncompensated H-polarized transmit baseband signal. is the cross-compensation complex coefficient applied to the H-polarized path signal. The product of the H-polarized signal and the main path weight WH constitutes a compensation term superimposed on the V-polarized signal path. This compensation term is intended to actively introduce an H-polarized source signal component of a specific amplitude and phase into the V channel to offset or weaken the signal caused by Described as the unwanted leakage or coupling effect from the original H-polarized signal to the V-polarized path. is the cross-compensation complex coefficient applied to the V-polarized path signal. The product of the V-polarized signal and the main path weight WV constitutes a compensation term superimposed on the H-polarized signal path. This compensation term is intended to actively introduce a V-polarized source signal component of a specific amplitude and phase into the H channel to offset or weaken the V-polarized signal caused by Described as the unwanted leakage or coupling effect from the original V-polarized signal to the H-polarized path.

[0125] This cross-compensation is necessary because when physical impairments degrade the inherent isolation between polarizations, simple single-port amplitude and phase adjustments are insufficient to restore it. This formula draws on the principles of signal preprocessing or interference cancellation in MIMO. By actively compensating for cross-coupling in the digital domain, it can preemptively "cancel" the cross-polarization components introduced by physical imperfections in the antenna feed before the signal leaves the antenna.

[0126] The amplitude and phase weights of the transmitted signals at the two polarization ports are adjusted through digital baseband processing. The complex weight of the V polarization port is expressed as:

[0127] : Where AV is the amplitude adjustment factor of the complex weight of the main path of the V polarization port transmission signal, is the phase adjustment factor.

[0128] : Where AH is the amplitude adjustment factor of the complex weight of the main path of the H-polarized port transmission signal, is the phase adjustment factor.

[0129] KPI-driven optimization search defines a comprehensive utility function U:

[0130] ;

[0131] The detailed meanings of the parameters in this formula are as follows: is a comprehensive utility function whose value is a scalar, which represents the comprehensive performance of the system under the current reconstruction parameter combination. The optimization goal is to find the parameter that maximizes U k1 is the weight coefficient of the first performance indicator, which is a positive value and is used to adjust the importance of SINR in the optimization target. (or more precisely ) is the signal-to-interference-plus-noise ratio (SIN) achievable by the target user or link after reconstruction, expressed in decibels (dB). Maximizing this factor helps improve communication quality. k2 is the weighting factor for the second performance metric, cross-polarization power. A positive value multiplied by a negative sign indicates a desire to minimize this factor. is the estimated total cross-polarization leakage power, expressed in watts (W) or decibel milliwatts (dBm). Reducing this term helps improve polarization purity. k3 is the weighting coefficient for the third performance indicator, healthiness deviation. A positive value multiplied by a negative sign indicates the desire to minimize this term. The overall health score of the feed diagnosed by the current AI model. is a desired baseline or target value for feed health, such as a pre-failure health score or an acceptable minimum health level, | |Indicates the gap between current health and target.

[0132] The reason and necessity for implementing this utility function lies in the fact that actual reconstruction requires a trade-off between multiple, potentially conflicting objectives. This formula provides a framework that unifies the desired communication performance (SINR), the antenna's own polarization purity (Pcross-pol), and the degree of deviation from the original healthy state, representing possible beam deformation, into a single optimization objective. By adjusting the weights ki, the reconstruction strategy can be flexibly adjusted to suit different scenarios, such as coverage priority, capacity priority, and interference suppression priority.

[0133] The optimization algorithm uses particle swarm optimization (PSO) to search for the best combination of complex weights. The control system is usually on the base station BBU side, receiving diagnostic information from the feed MCU or edge computing unit and the current S parameters to run optimization algorithms such as hill climbing, simulated annealing, particle swarm optimization, or gradient-based optimization methods to search for the best combination of complex weights. or compensation factor To maximize the utility function U. During the optimization process, it is necessary to consider the adjustable parameters 、 、 、 The actual dynamic range and accuracy of

[0134] The PSO algorithm parameters were set as follows: a swarm size of 30, a maximum number of iterations of 100, an inertia weight of 0.7, and a learning factor of c1 = c2 = 2.0. The algorithm updates the particle positions and velocities in each iteration, gradually converging to the optimal weight combination through guidance from global and individual optima.

[0135] The optimized weight parameters are applied through the base station's digital signal processing (DSP) link. The base station BBU applies the calculated optimized complex weights or compensation coefficients to the DSP link, adjusting the transmitted signals of the two physically polarized ports. After reconfiguration, the system continuously monitors KPIs and feed sensor data to evaluate the reconfiguration results. If the results are poor or the status continues to deteriorate, reconfiguration can be attempted or a higher-level alarm can be triggered. The base station BBU receives the weight parameters calculated by the optimization algorithm and applies them to the digital predistortion (DPD) unit and beamforming unit. Digital signal processing includes complex weighting of the baseband signal, cross-coupling compensation, digital upconversion, and power amplifier predistortion.

[0136] The technical effect of this step is reflected in adaptive performance recovery. When the feed source is physically damaged or its performance degrades, the polarization mode of the transmitted signal is intelligently adjusted to actively compensate for antenna defects, thereby restoring or maintaining the key performance indicators of the communication link to a certain extent. Through precise cross-polarization cancellation, the inter-polarization interference caused by feed source failure is effectively suppressed, improving the performance of the MIMO system. This allows the antenna system to continue to provide services after certain types of failures, delaying the need for physical repair or replacement and reducing operation and maintenance costs. It improves the robustness and service continuity of the communication system in the face of antenna component failures and reduces service interruptions caused by antenna problems. The introduction of the utility function U allows the reconstruction strategy to be flexibly adjusted according to current network requirements, such as coverage priority and capacity priority, to achieve more refined performance management.

[0137] 7. Execute the alarm and maintenance instructions as follows:

[0138] The detailed working process of alarm and maintenance guidance includes the system continuously monitoring the overall health score of the feed diagnosed by AI and restructured KPIs. Continuously below the preset critical alarm threshold , or when polarization intelligent reconstruction fails to maintain key KPIs above an acceptable level, or when the AI ​​model predicts that the feed source is about to suffer a serious failure in a short period of time, the system automatically triggers the alarm mechanism. The system generates a detailed alarm report, including: the unique identifier of the faulty antenna, geographic location information (if integrated with GPS), alarm occurrence time, alarm level, detailed results of AI diagnosis, and a summary of the sensor raw data and feature vector when the alarm was triggered. , attempted reconstruction strategies and their effectiveness evaluations, and a preliminary fault cause analysis based on the system's failure mode. This alarm report is submitted to the upper-level network operations and maintenance management system (OMS) via a network interface. The OMS receives and analyzes the alarm information, presents it to maintenance personnel, and automatically generates maintenance work orders based on pre-set rules.

[0139] The alarm trigger mechanism is based on multiple conditions: when the overall health score of the AI ​​diagnosis feed Continuously below the critical alarm threshold More than 5 minutes; when after polarization intelligent reconstruction is executed, key KPIs still cannot be maintained above acceptable levels, such as the SINR drops by more than 3dB; when the AI ​​model predicts that the feed source will suffer a serious failure within the next 24 hours, the health score prediction value is less than 30 points.

[0140] in: The critical alarm threshold for the preset overall health score of the feed. Below this value, the feed is in very poor condition and requires immediate attention or maintenance.

[0141] Alarm information is generated in a structured report format, which includes the following: antenna unique identifier, GPS location information latitude and longitude coordinates, alarm timestamp in UTC format, alarm levels 1-5, with level 5 being the most severe, and detailed results of AI diagnosis. , , , , the original data snapshot of the sensor when the alarm is triggered, the current feature vector The complete numerical value of the system, the reconstruction strategies that have been tried and their effect evaluation, and the preliminary failure cause analysis given by the system based on the failure mode.

[0142] Fault cause analysis utilizes a combined rule engine and knowledge graph. The rule engine contains 150 expert experience rules, covering the characteristic patterns and possible causes of common fault modes. The knowledge graph establishes relationships between antenna components, fault types, and environmental factors, supporting reasoning analysis. For example, if the current amplitude deviation of the B1 dipole arm is detected to be greater than 20%, the VH polarization mutual coupling coefficient increases by more than 15%, and the ambient humidity exceeds 85%, the system infers that "the insulation performance of the B1 dipole arm may have degraded due to moisture intrusion, causing current abnormalities and increased polarization coupling."

[0143] Alarm information is reported via the SNMP protocol and the RESTful API. SNMP version 3 supports encryption and authentication, ensuring secure transmission of alarm information. The RESTful API uses the HTTPS protocol and supports data exchange in JSON format. Reporting targets the network operations management system (OMS), supporting multi-level reporting and cascading alarms.

[0144] Maintenance guidance is generated based on expert systems and decision tree algorithms. Based on the alarm information, OMS provides operation and maintenance personnel with accurate fault antenna location and preliminary fault cause analysis, and guides on-site maintenance work such as checking specific vibrator arms, feeder connections, cleaning foreign objects, etc., thus shortening the troubleshooting and repair time. , automatically generating detailed maintenance instructions, including: a precise description of the fault location, recommended inspection steps, a list of required tools and spare parts, safety precautions, and an estimated repair time. For example, for a B1 arm failure, the maintenance instructions include: "1. Turn off the antenna power; 2. Remove the antenna cover; 3. Inspect the B1 arm for physical damage; 4. Use a multimeter to measure the DC resistance of the B1 arm; 5. Check the connection between the B1 arm and the feed network; 6. If damage is found, replace the B1 arm assembly."

[0145] The technical benefit of this step is reflected in timely and proactive alarms, enabling early detection and timely reporting of faults, thus preventing the accumulation of problems that could lead to more serious system failures or prolonged service quality degradation. The detailed diagnostic information and preliminary cause analysis provided greatly assist operations and maintenance personnel in quickly locating problematic antennas and fault points, shortening troubleshooting time. It provides clear guidance for on-site maintenance work, such as checking specific antenna arms and feeder connections, improving maintenance efficiency and reducing the heavy reliance on maintenance personnel's experience. Accumulated alarm and diagnostic data can be used to analyze common patterns and trends in antenna failures, providing data support for spare parts management, the development of preventive maintenance plans, and optimizing overall operations and maintenance strategies. Maintenance is triggered by alarms, and maintenance results can be fed back through system monitoring, forming a closed loop of fault management and performance maintenance.

[0146] By implementing the above-mentioned steps, the performance maintenance method for the high-gain dual-polarized parabolic antenna system of the present invention can achieve refined and intelligent health status monitoring of the feed unit, early fault warning, and maintain the quality and availability of the communication link as much as possible through intelligent reconstruction when performance degradation occurs, thereby improving the resilience and adaptability of the antenna system. Through real-time monitoring, intelligent diagnosis and adaptive reconstruction, this method ensures that the antenna can still maintain the availability and quality of the communication link to the greatest extent possible when the feed unit performance deteriorates or fails, significantly improving the resilience and adaptability of the antenna system, and providing strong guarantees for the reliability and service quality of the wireless communication system.

Claims

1. A high-gain dual-polarization parabolic antenna system, comprising: parabolic reflector; A feed source, the feed source being arranged at a focal point or a near-focal area of ​​the parabolic reflector; It is characterized in that the feed source includes a printed circuit board (PCB), and a metal vibrator unit is formed on the PCB through an etching process; The metal vibrator unit includes at least two groups of orthogonally arranged vibrators to form radiation in two orthogonal polarization directions; Each vibrator of the metal vibrator unit is integrated with a low-frequency radiation portion for a first frequency band and a high-frequency radiation portion for a second frequency band, wherein the first frequency band and the second frequency band correspond to the 900 MHz frequency band and the 1800 MHz frequency band respectively; The feed source radiates dual-polarized electromagnetic waves through the metal vibrator unit, and the electromagnetic waves are reflected by the parabolic reflector to form a high-gain directional beam.

2. The high-gain dual-polarization parabolic antenna system according to claim 1, characterized in that: The metal dipole unit includes a first polarization dipole group, which is composed of a first metal dipole arm and a second metal dipole arm symmetrically arranged therewith, and the first metal dipole arm and the second metal dipole arm together form radiation in one polarization direction.

3. The high-gain dual-polarization parabolic antenna system according to claim 2, characterized in that: The metal vibrator unit also includes a second polarized vibrator group, which is composed of a third metal vibrator arm and a fourth metal vibrator arm symmetrically arranged therewith. The second polarized vibrator group is arranged on the PCB plane rotated 90 degrees relative to the first polarized vibrator group to form radiation in a polarization direction orthogonal to the first polarized vibrator group.

4. The high-gain dual-polarization parabolic antenna system according to claim 3, characterized in that: Each of the first metal vibrator arm, the second metal vibrator arm, the third metal vibrator arm and the fourth metal vibrator arm includes a low-frequency radiation part and a high-frequency radiation part, and the high-frequency radiation part is nested or integrated in or near the structure of the low-frequency radiation part. The low-frequency radiation part has an arm-shaped structure with a longer length and a wider width, and the high-frequency radiation part has an arm-shaped structure with a shorter length and a narrower width.

5. The high-gain dual-polarization parabolic antenna system according to claim 4, characterized in that: The feed source also includes a guide plate, which is formed on the PCB through an etching process. The guide plate is located in front of the main radiation direction of the high-frequency radiation part and is used to expand the operating bandwidth of the 1800MHz frequency band and increase the gain value of this frequency band.

6. The high-gain dual-polarization parabolic antenna system according to claim 5, characterized in that: The voltage standing wave ratio of the antenna system in the frequency band of 900 MHz to 1800 MHz is less than or equal to 1.

7.

7. A method for maintaining performance of a high-gain dual-polarization parabolic antenna system, characterized in that: The following steps are involved: Sensor deployment steps: Integrate current monitoring sensors, port characteristic monitoring sensors, and environmental parameter sensors on the feed unit; Data acquisition configuration steps: configure the data acquisition unit, establish connections with each sensor, and set the data acquisition cycle and parameters; Data acquisition and processing steps: collecting the output signals of each sensor according to a set period through the data acquisition unit, filtering and digitalizing the collected signals; Data output step: transmitting the processed data to an external processing unit through a communication interface or storing it in a local storage medium; Among them, the sensor deployment step provides a signal source basis for the data acquisition configuration step, the parameter setting in the data acquisition configuration step controls the execution timing of the data acquisition and processing step, and the output results of the data acquisition and processing step are externally transmitted through the data output step.

8. The method for maintaining performance of a high-gain dual-polarization parabolic antenna system according to claim 7, wherein: Also includes: Extract the feed source health status characteristic parameters from the output data of the data acquisition and processing, wherein the characteristic parameters include: Calculate the current amplitude deviation of each metal vibrator arm, which is obtained by the absolute value of the difference between the current monitoring current amplitude and the reference current amplitude; Calculate the current phase deviation of each metal vibrator arm, which is obtained by the absolute value of the difference between the current monitoring current phase and the reference current phase; Extract the percentage change of mutual coupling coefficient amplitude and phase shift between polarization ports; Calculating a mutual coupling fingerprint deviation metric value based on the mutual coupling coefficient amplitude change percentage and the phase drift; The current amplitude deviation, current phase deviation, mutual coupling coefficient variation parameter, mutual coupling fingerprint deviation measurement value and environmental parameters are combined to construct a comprehensive feature vector.

9. The method for maintaining performance of a high-gain dual-polarization parabolic antenna system according to claim 8, wherein: Also includes: Use machine learning algorithms to build health diagnosis models; Using electromagnetic simulation data and laboratory accelerated aging test data to train the diagnostic model, and establish a mapping relationship from the comprehensive feature vector to health score, fault type and fault location; Deploy the trained diagnostic model to the data acquisition unit or edge computing unit; The comprehensive feature vector obtained in real time is input into the diagnostic model, and the health score of each metal vibrator arm, the overall health score of the feed source, the fault mode recognition result and the fault location information are output.

10. The method for maintaining performance of a high-gain dual-polarization parabolic antenna system according to claim 9, wherein: Also includes: When the overall health score of the feed is lower than a preset threshold, polarization intelligent reconstruction is initiated; Based on the fault mode identification result and the fault location information, identifying a harmful coupling path that causes a decrease in polarization isolation; estimating complex coefficients of cross-polarization leakage and calculating compensation parameters for suppressing the harmful coupling path; By adjusting the amplitude weights and phase weights of the transmitted signals of the two polarization ports, cross compensation processing is performed on the original polarization port signals; An optimization algorithm is run to search for the best combination of complex weights to maximize a comprehensive utility function including signal-to-interference-plus-noise ratio, cross-polarization leakage power, and health deviation. The calculated optimized weight parameters are applied to the base station digital signal processing link to dynamically adjust the baseband signal fed to the antenna polarization port.

Images