Intelligent auxiliary debugging method and system for instrument landing system parameters

Abstract

The invention relates to the technical field of flight verification, and discloses an intelligent auxiliary debugging method and system for instrument landing system parameters, and the method comprises the steps: constructing a multi-dimensional coupling virtual state space and a standard working geometric polyhedron, and positioning a current equipment state vector based on flight verification original data; calculating a multi-dimensional displacement deviation with a target center, generating an initial suggestion adjustment vector in combination with the hybrid model, and performing simulation deduction to obtain a candidate adjustment strategy; after steady-state judgment and filtering in a hysteresis silence zone, packaging into an atomization execution instruction set, performing closed-loop verification after execution, and adaptively updating the hybrid model; according to the method, debugging deadlock is broken through virtual rehearsal and coupling adaptation, and steady-state operation of the system is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of flight calibration, and more specifically, to a smart assisted debugging method and system for instrument landing system parameters. Background Art

[0002] The Instrument Landing System (ILS) is the standard approach and landing navigation equipment commonly used in international civil aviation at present, and its performance directly relates to the flight safety of aircraft. During the whole life cycle of the instrument landing system, flight calibration is a key link to ensure that the signal space quality meets the requirements of civil aviation regulations.

[0003] In the prior art, a Chinese patent with the authorization announcement number CN113282098B discloses a method for improving the flight calibration accuracy of an instrument landing system. This technology collects signals through an airborne antenna array and processes them to obtain guidance data, guiding the unmanned aerial vehicle to fly along the optimal route, mainly solving the problems of the flight accuracy of the calibration platform itself and the accuracy of data collection. A Chinese patent with the authorization announcement number CN111824453B discloses a method and system for comparative analysis of flight calibration results of instrument landing system equipment. This solution focuses on the backend processing of data. By associating the calibration results with the transmitter setting values, calculating parameter fluctuations and issuing optimization suggestions, it realizes data visualization and trend analysis of equipment maintenance.

[0004] However, although the above prior art has improved the acquisition accuracy and analysis dimension of calibration data, when facing the core control link of ground transmitter parameter debugging, it still faces the deep technical problem of "debugging deadlock and transient out-of-bounds induced by strong coupling of multi-dimensional parameters". Specifically, there is a non-linear dynamic coupling relationship among the core parameters of the instrument landing system (such as course alignment, course width, warning threshold, etc.) at the physical level. That is, the adjustment of a single parameter often causes unexpected offsets of related parameters. In the traditional debugging mode, operators lack the ability to globally quantify and predict this coupling effect, and often fall into the dilemma of "pressing the gourd and floating the ladle": adjusting the transmitter attenuation value to correct the width deviation, but unexpectedly causing the relative value of the originally normal warning threshold to exceed the safety boundary. More critically, this "serial and step-by-step" execution method based on manual experience cannot guarantee the atomicity of changes in multiple related parameters. During the time interval of instruction execution, the device is extremely likely to be in an intermediate state with parameter mismatch, thus triggering an instantaneous alarm of the monitor and causing the device to shut down. This debugging blindness caused by the lack of a virtual rehearsal mechanism and atomic execution means not only greatly prolongs the expensive flight test time, but also increases the safety hazards of the device operating in a critical state. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of existing technologies, this invention provides an intelligent assisted debugging method and system for instrument landing system parameters. By constructing a multi-dimensional coupled virtual state space and defining a standard working geometric polyhedron, discrete electrical parameters are transformed into state vectors capable of geometric operations. Combined with a hybrid model, an initial suggested adjustment vector is generated, and simulation deduction and collision detection are performed in the virtual space. Through anti-deadlock logic, hysteresis silent zone filtering, and atomic instruction encapsulation mechanisms, forward-looking prediction of parameter adjustments and multi-parameter collaborative execution are achieved. This effectively avoids debugging deadlock and equipment alarm triggering caused by parameter coupling, improves flight calibration efficiency, and reduces impact losses on transmitter hardware.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A smart assisted debugging method for instrument landing system parameters includes: Construct a multidimensional coupled virtual state space, and define a standard working geometric polyhedron within the multidimensional coupled virtual state space; obtain the original data sequence of the instrument landing system flight verification, and locate the current equipment state vector of the current ground transmitter in the multidimensional coupled virtual state space based on the original data sequence of the instrument landing system flight verification. Determine the target center coordinates of the standard working geometric polyhedron, calculate the vector difference between the current equipment state vector and the target center coordinates to obtain the multidimensional displacement deviation, generate an initial suggested adjustment vector based on the multidimensional displacement deviation and the pre-built hybrid model, substitute the initial suggested adjustment vector into the multidimensional coupled virtual state space to perform simulation deduction, and generate candidate adjustment strategies based on the deduction results; Candidate adjustment strategies are determined and filtered for steady-state conditions through a preset hysteresis quiescent zone, and then encapsulated into an atomic execution instruction set. The atomic execution instruction set is executed to obtain a new round of current device state vector, and closed-loop verification and hybrid model adaptive update are performed based on the new round of current device state vector.

[0007] The multidimensional coupled virtual state space includes four dimensions: heading straightness dimension, heading width dimension, glide slope dimension, and glide slope width dimension. The method for generating the standard working geometric polyhedron includes: obtaining the upper limit and lower limit of the normal operating parameters of the current ground transmitter, and forming the standard working geometric polyhedron by enclosing the upper limit and lower limit of the normal operating parameters within a multi-dimensional coupled virtual state space.

[0008] The method for locating the current device state vector of the current ground transmitter includes: Receive the original flight verification data sequence of the instrument landing system transmitted by the verification aircraft, remove the high-frequency noise components in the original flight verification data sequence of the instrument landing system and smooth the abrupt change points to obtain a high-confidence real-time flight verification net value. By combining the high-confidence real-time flight verification net value with the current ground transmitter's launch parameter settings, and through multi-dimensional parameter normalization mapping, the current device state vector is located in the multi-dimensional coupled virtual state space.

[0009] The method for generating the initial proposed adjustment vector includes: The displacement deviation components of the heading straightness dimension, heading width dimension, glide slope angle dimension, and glide slope width dimension in the multidimensional displacement deviation are extracted. The displacement deviation components of each dimension are calculated separately, and the calculation results are vectorized to generate the initial suggested adjustment vector.

[0010] The hybrid model includes a microampere-DDM conversion model combined with a linear negative feedback model and a logarithmic nonlinear correction model; The method for calculating the displacement deviation components in each dimension includes: For the displacement deviation components in the heading alignment dimension and glide slope dimension, the micro-ampere-DDM conversion model combined with the linear negative feedback model is used to calculate the heading alignment dimension adjustment component and glide slope dimension adjustment component; For the displacement deviation components in the heading width dimension and the glide slope dimension, a logarithmic nonlinear correction model with a smoothing correction coefficient is used to calculate the adjustment components in the heading width dimension and the glide slope dimension.

[0011] The simulation results include at least the predicted values ​​for each dimension; The method for generating candidate adjustment strategies based on the inference results includes: performing collision detection between the predicted values ​​of each dimension and the standard working geometric polyhedron; if a collision occurs, activating the deadlock prevention logic and generating a deadlock prevention suggested adjustment strategy as a candidate adjustment strategy; if no collision occurs, using the initial suggested adjustment vector as a candidate adjustment strategy.

[0012] The deadlock prevention logic includes: performing a safety clamping process on the initial suggested adjustment vector to generate a safety compromise adjustment vector, calculating a threshold linkage adjustment vector based on the safety compromise adjustment vector, and combining the safety compromise adjustment vector and the threshold linkage adjustment vector to generate a deadlock prevention suggested adjustment strategy.

[0013] The method for steady-state determination and filtering includes: Extract the adjustment amounts of parameters for each dimension from the candidate adjustment strategies, calculate the expected state value after adjustment for each dimension, and if the expected state value of a certain dimension falls within the hysteresis quiescent zone, then the adjustment amount of the parameter for that dimension is forcibly set to zero; if it does not fall within the hysteresis quiescent zone, then the adjustment amount of the parameter for that dimension is retained, and valid adjustment data is generated.

[0014] When the candidate adjustment strategy is the initial suggested adjustment vector, the adjustment amounts of each dimension parameter in the candidate adjustment strategy include the heading straight dimension adjustment component, the heading width dimension adjustment component, the glide slope dimension adjustment component, and the glide slope width dimension adjustment component; when the candidate adjustment strategy is the deadlock prevention suggested adjustment strategy, the adjustment amounts of each dimension parameter include the deadlock prevention heading straight dimension adjustment component, the deadlock prevention heading width dimension adjustment component, the deadlock prevention glide slope dimension adjustment component, the deadlock prevention glide slope width dimension adjustment component, and the threshold linkage adjustment vector.

[0015] A smart auxiliary adjustment system for instrument landing system parameters, used to implement the aforementioned smart auxiliary adjustment method for instrument landing system parameters, the system comprising: State space positioning module: used to construct a multidimensional coupled virtual state space, define a standard working geometric polyhedron within the multidimensional coupled virtual state space; acquire the original data sequence of the instrument landing system flight verification, and locate the current equipment state vector of the current ground transmitter in the multidimensional coupled virtual state space based on the original data sequence of the instrument landing system flight verification. Strategy generation module: used to determine the target center coordinates of the standard working geometric polyhedron, calculate the vector difference between the current device state vector and the target center coordinates to obtain the multidimensional displacement deviation, generate an initial proposed adjustment vector based on the multidimensional displacement deviation and the pre-built hybrid model, substitute the initial proposed adjustment vector into the multidimensional coupled virtual state space to perform simulation deduction, and generate candidate adjustment strategies based on the deduction results; Closed-loop update module: This module performs steady-state determination and filtering of candidate adjustment strategies through a preset hysteresis quiescent zone, and encapsulates them into an atomic execution instruction set; it executes the atomic execution instruction set and obtains the current device state vector for the next round, and performs closed-loop verification and hybrid model adaptive update based on the new round of current device state vector.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention constructs a multi-dimensional coupled virtual state space and a standard working geometric polyhedron with embedded parameter physical coupling characteristics. Combined with the fusion data acquisition mechanism of the airborne ILS signal receiving sensor and the satellite positioning system, and the real-time status feedback from multiple types of monitoring sensor groups on the ground transmitter, it achieves precise quantitative characterization of equipment operating status and parameter boundaries, providing a systematic spatial reference framework for parameter adjustment. By using simulation to predict the adjusted system state in advance, it overcomes the limitations of the traditional "adjust first, observe later" delayed feedback, accurately identifying boundary exceedance risks caused by parameter coupling. After filtering through the hysteresis quiescent zone steady-state determination, the atomic execution instruction set ensures the synchronization and integrity of the associated parameter adjustments, avoiding transient exceedances caused by timing misalignments. Combined with closed-loop verification and hybrid model adaptive updates, the debugging process can dynamically adapt to changes in equipment characteristics, continuously optimizing adjustment accuracy, effectively resolving logic deadlocks caused by dynamic parameter coupling, ensuring the equipment remains within safe operating boundaries during debugging, avoiding equipment shutdown and parameter rollback caused by alarm triggers, and significantly improving the stability, reliability, and efficiency of instrument landing system parameter debugging. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating a method for intelligent assisted debugging of instrument landing system parameters provided in an embodiment of the present invention; Figure 2 A schematic diagram of an instrument landing system heading guidance scenario provided in an embodiment of the present invention. Figure 3 A flowchart illustrating the principle of the candidate adjustment strategy provided in this embodiment of the invention; Figure 4 This is a schematic diagram of collision detection and determination provided in an embodiment of the present invention; Figure 5 A flowchart illustrating the principle of steady-state determination and filtering provided in this embodiment of the invention; Figure 6 This is a functional block diagram of an intelligent auxiliary debugging system for instrument landing system parameters provided in an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1 Please see Figure 1 As shown, this embodiment provides a smart assisted debugging method for instrument landing system parameters, including: Step S10: Construct a multidimensional coupled virtual state space, define a standard working geometric polyhedron within the multidimensional coupled virtual state space; obtain the original data sequence of the instrument landing system flight verification, and locate the current equipment state vector of the ground transmitter in the multidimensional coupled virtual state space based on the original data sequence of the instrument landing system flight verification. Further, step S10 includes: Step S11: Receive the original data sequence of the instrument landing system flight verification transmitted by the verification aircraft, remove the high-frequency noise components in the original data sequence of the instrument landing system flight verification and smooth the abrupt change points to obtain the high-confidence real-time flight verification net value. The system receives raw data sequences of the Instrument Landing System (ILS) flight calibration continuously transmitted by the calibration aircraft during approach via a standardized data interface. These data sequences are acquired in real-time by ILS signal receiving sensors on the calibration aircraft and spatially calibrated using precise three-dimensional aircraft position information obtained from the onboard satellite positioning module. The data sequences comprise two main categories: heading equipment parameters and glide slope equipment parameters. (See also...) Figure 2 This is a schematic diagram (top view) of the heading guidance scenario of the instrument landing system provided in the embodiments of this application. Figure 2 The diagram shows the localizer beacon at the runway end, the centerline extending from the beacon along the runway, and the runway sector boundaries indicated by dashed lines on both sides; the calibration aircraft is approaching the runway along the approach direction. (Combined with...) Figure 2 The geometric relationships shown include heading equipment parameters such as heading alignment parameters, alignment alarm thresholds, heading width parameters, heading width alarm thresholds, and heading narrowing alarm thresholds. The heading alignment parameters, defined as CLDDM, characterize the horizontal deviation of the aircraft being checked relative to the heading centerline in the diagram. Figure 2 The physical meaning of the horizontal deviation line segments marked in the diagram; when verifying the heading alignment deviation reported by the aircraft crew, microamperes (μA) are used as the unit of measurement, and the report content is in the form of "leftward / rightward deviation by how many microamperes", where there is a conversion relationship between the microampere value and the DDM value, approximately 10μA corresponds to 1DDM; the alignment alarm threshold is the alarm threshold of CLDDM; the heading width parameter is defined as DSDDM, corresponding to Figure 2 The width range of the glide path sector defined by the dashed lines on both sides is measured in degrees (°) when the aircraft crew reports the heading width, directly reporting the angle width value of the glide path sector. The glide path equipment parameters include glide angle parameters, glide angle lower limit, glide angle upper limit, glide path width parameters, glide path width alarm threshold, and glide path narrowing alarm threshold. The glide angle parameter is measured by the airborne glide path signal receiving sensor, which characterizes the degree of deviation of the aircraft from the glide path centerline in the vertical direction. At the same time, the aircraft altitude and distance information obtained by the airborne satellite positioning module are used to assist in the verification of the glide angle measurement value. When the aircraft crew reports the glide angle deviation, microamperes (μA) are used as the unit of measurement. The glide angle lower limit and glide angle upper limit are the glide angle thresholds. The glide path width parameter is defined as glide path DSDDM. When the aircraft crew reports the glide path width, angles (°) are used as the unit of measurement, directly reporting the angle width value of the glide path sector. The heading alignment parameter (CLDDM) is a core indicator measuring the aircraft's deviation from the runway horizontal centerline, and its value changes directly reflect the guidance accuracy of the ground-based heading beacon signal. The glide slope parameter is a core indicator measuring the aircraft's deviation from the glide slope vertical centerline. The heading width parameter (DSDDM) and glide slope parameter determine the range of the heading and glide slope sectors, respectively. Considering that aircraft are susceptible to turbulence and multipath effects caused by reflections from ground structures at high altitudes, the raw data sequence for instrument landing system flight calibration exhibits fluctuations with random spikes in the time domain. Directly using this data for calculations would lead to significant deviations in the adjustment values. To address this, a hybrid processing strategy combining Fast Fourier Transform (FFT) and median filtering is employed. First, the original data sequence is transformed from the time domain to the frequency domain using FFT, identifying and removing high-frequency noise components with frequencies more than three times the fundamental frequency of the signal. Then, the filtered frequency-domain data is inversely transformed back to the time domain, and median filtering is applied to smooth abrupt changes in the time-domain curve. The window size for median filtering is dynamically set according to the data sampling frequency; for example, when the sampling frequency is 10Hz, the window size is set to 5 data points, ensuring that isolated abrupt changes are suppressed while preserving the signal trend. Through this hybrid filtering process, environmental interference in the original data is effectively removed, ultimately yielding a high-confidence real-time flight verification net value that accurately reflects the launch performance of the ground transmitter. Step S11 eliminates the interference of noise on the authenticity of the data, ensuring that all subsequent calculations are based on reliable input; it avoids misjudgment of adjustment direction or deviation of adjustment amplitude caused by noise, and provides a clean data foundation for accurate positioning of equipment status. At the same time, compared with the single filtering method, the synergy of hybrid filtering not only preserves the core features of the signal, but also removes different types of interference to the maximum extent, solving the problem of difficulty in balancing "noise reduction and feature preservation" in traditional data processing.

[0021] Step S12: Construct a multi-dimensional coupled virtual state space including heading alignment dimension, heading width dimension, glide slope dimension, and glide slope width dimension; obtain the upper limit and lower limit of the normal operation index of the current ground transmitter; and form a standard working geometric polyhedron by enclosing the upper limit and lower limit of the normal operation index within the multi-dimensional coupled virtual state space. A four-dimensional orthogonal, multi-dimensional coupled virtual state space is constructed within the software. The four dimensions correspond to the heading alignment dimension (CLDDM), heading width dimension (DSDDM), glide slope angle dimension, and glide slope width dimension (glide slope DSDDM), respectively. The selection of these dimensions is based on the core working mechanism of the ILS device: the heading alignment dimension represents the guidance accuracy of the horizontal centerline, the heading width dimension represents the coverage range of the horizontal channel sector, the glide slope angle dimension represents the guidance accuracy of the vertical centerline, and the glide slope width dimension represents the coverage range of the vertical glide slope. These four dimensions together constitute the core elements for the normal operation of the ILS device. Specifically, the heading device provides horizontal centerline guidance, and the glide slope device provides vertical centerline guidance. Together, they form a complete three-dimensional glide slope guideline and are physically coupled; for example, changes in the heading width directly affect the effective range of the alarm threshold. Unlike traditional abstract parameter coordinate systems, this multi-dimensional coupled virtual state space embeds the electrical and physical characteristics of the ILS device, with the scale of each dimension strictly corresponding to the actual range of variation of the physical parameters.

[0022] Subsequently, the upper and lower limits of the normal operating parameters of the equipment as specified in civil aviation regulations and equipment technical manuals, i.e., alarm thresholds, are obtained. Within a multi-dimensional coupled virtual state space, these upper and lower limits enclose a closed standard working geometric polyhedron. The geometric shape of this standard working geometric polyhedron is determined by the boundaries of each dimension. For example, the upper and lower limits of the heading alignment dimension are the DDM range specified by the alignment alarm threshold; the upper and lower limits of the heading width dimension are the angle range specified by the heading width alarm threshold and the heading narrow alarm threshold; the upper and lower limits of the glide slope angle dimension are the DDM range specified by the glide slope angle lower limit threshold and the upper limit threshold; and the upper and lower limits of the glide slope width dimension are the angle range specified by the glide slope width and narrow alarm thresholds. These four dimensions enclose a four-dimensional hypergeometric polyhedron, which figuratively defines the "safe working corridor" of the equipment. Only when the coordinate point corresponding to the parameter combination of the equipment falls within this polyhedron is the equipment in a normal operating state without alarms. Step S12 transforms the abstract electrical parameter relationships into concrete geometric spatial relationships, clarifying the boundary range of normal equipment operation. It establishes a systematic view of the physical coupling between parameters, overcoming the shortcomings of parameter discretization and lack of correlation representation in traditional technologies. This allows the mutual influence between parameters to be intuitively reflected through geometric positional relationships. Simultaneously, the construction of the standard working geometric polyhedron provides a clear basis for subsequent collision detection, laying the spatial foundation for the "pre-simulation followed by adjustment" anti-deadlock logic. For example, the boundaries of each dimension of the standard working geometric polyhedron of a runway ILS device are shown in Table 1: Table 1. Boundaries of the standard working geometric polyhedron of a runway ILS device in various dimensions. Step S13: Combine the high-confidence real-time flight verification net value with the current ground transmitter's launch parameter settings, and locate the current device state vector in the multi-dimensional coupled virtual state space through multi-dimensional parameter normalization mapping.

[0023] The high-confidence real-time flight verification net value obtained in step S11 is fused with the current launch parameter settings of the ground transmitter. These settings include adjustable parameters such as heading alignment settings, heading width settings, glide slope settings, glide slope settings, and various alarm threshold settings. These parameters are read in real-time through the transmitter's control interface and verified in real-time by the integrated status monitoring sensor group within the transmitter to ensure consistency between the read values ​​and the actual launch status. Since the high-confidence real-time flight verification net value reflects the equipment's "actual output performance," and the launch parameter settings reflect the equipment's "current input configuration," their fusion comprehensively characterizes the equipment's operating status. The current equipment state vector is generated using a multi-dimensional parameter normalization mapping method. For the heading alignment dimension, the heading alignment microampere measurement value (in μA) reported by the flight crew is extracted from the high-confidence real-time flight verification net value. First, the heading alignment microampere measurement value is converted into a DDM value according to the physical conversion relationship between microampere and DDM. The conversion formula is: Heading alignment DDM value = Heading alignment microampere measurement value / φ, where φ is the microampere-DDM conversion factor, with an initial value of 10μA / DDM. This value represents the number of microamperes corresponding to 1DDM. This conversion relationship is determined based on the modulation signal detection principle of the ILS equipment. Then, the converted heading alignment DDM value is divided by the scale factor of the heading alignment dimension (i.e., 0.001DDM / unit scale) to obtain the heading alignment state component θ1. For the heading width dimension, the heading width angle measurement value (in °) reported by the calibration aircraft crew is extracted from the high-confidence real-time flight calibration net value. The angle measurement value is divided by the scale coefficient of the heading width dimension (i.e., 0.01° / unit scale) to obtain the heading width state component θ2. For the glide slope dimension, the glide slope microampere measurement value (in μA) from the flight crew report is extracted from the high-confidence real-time flight calibration net value. First, the glide slope microampere measurement value is converted into a DDM value according to the physical conversion relationship between microampere and DDM. The conversion formula is: glide slope DDM value = glide slope microampere measurement value / φ. Then, the converted glide slope DDM value is divided by the scale factor of the glide slope dimension (i.e., 0.001 DDM / unit scale) to obtain the glide slope state component θ3. For the glide width dimension, the glide width angle measurement value (in °) reported by the calibration aircraft crew is extracted from the high-confidence real-time flight calibration net value. This angle measurement value is divided by the scale coefficient of the glide width dimension (i.e., 0.01° / unit scale) to obtain the glide width state component θ4.

[0024] The four-dimensional state components constitute the current device state vector θ=[θ1,θ2,θ3,θ4], which corresponds to a unique coordinate point in the multi-dimensional coupled virtual state space. To improve the state vector's resistance to transient fluctuations, the system uses a sliding window averaging method to smooth the state components across multiple consecutive sampling times. The window size is determined based on the sampling frequency, ensuring that the state vector reflects the steady-state characteristics of the device within a certain time interval rather than instantaneous disturbances. Step S13 quantifies the device's "input configuration-output performance" relationship into spatial coordinates, achieving precise positioning of the device state. This solves the problem that traditional technologies cannot quantify the comprehensive state of the device, transforming the originally abstract operating state into calculable and comparable geometric coordinates, providing a core basis for subsequent calculation of state deviations and planning adjustment paths. It should be noted that, for the sake of clarity in subsequent calculations, in the calculations of steps S21 to S33, the components of the current equipment state vector θ in the heading alignment dimension and glide slope dimension, the components of the adjustment component ∆, and the components of the theoretical state vector θpred are all calculated in DDM (which can be converted to the microampere value reported by the aircraft through the conversion factor φ). The components of the heading width dimension and glide slope width dimension are calculated in angle (°). The scale coefficients in Table 1 (0.001 DDM / unit scale) are only used for the visualization of the multidimensional coupled virtual state space and the intuitive understanding of the spatial geometric relationship, and do not affect the accuracy and correctness of the actual numerical calculations.

[0025] Step S20: Determine the target center coordinates of the standard working geometric polyhedron, calculate the vector difference between the current equipment state vector and the target center coordinates to obtain the multidimensional displacement deviation, generate an initial suggested adjustment vector based on the multidimensional displacement deviation and the pre-built hybrid model, substitute the initial suggested adjustment vector into the multidimensional coupled virtual state space to perform simulation deduction, and generate candidate adjustment strategies based on the deduction results; Further, step S20 includes: Step S21: Extract the displacement deviation components of the heading straightness dimension, heading width dimension, glide slope angle dimension, and glide slope width dimension from the multidimensional displacement deviation. Perform calculations on the displacement deviation components of each dimension separately, and synthesize the calculation results into a vector to generate the initial suggested adjustment vector. The target center coordinates are the median points of the boundaries of each dimension of the standard working geometric polyhedron. They are determined based on the optimal operating characteristics of the ILS device. When the parameter combination is at the center of each dimension, the device's signal stability is strongest, its anti-interference capability is optimal, and it is far from the alarm boundary, providing the maximum safety margin. The target center coordinates are calculated as follows: For each dimension (heading alignment dimension, heading width dimension, glide slope dimension, glide slope width dimension), the arithmetic mean of the upper and lower limits of that dimension is taken. The median points of the four dimensions together constitute the target center coordinates, denoted as C=[C1,C2,C3,C4], where C1 is the center value of the heading alignment dimension (in DDM), C2 is the center value of the heading width dimension (in °), C3 is the center value of the glide slope angle dimension (in DDM), and C4 is the center value of the glide slope width dimension (in °). The vector difference between the current device state vector θ and the target center coordinate C is calculated to obtain the multidimensional displacement deviation ∆θ = θ - C. This difference directly quantifies the degree and direction of deviation between the current device state and the optimal state. For example, ∆θ1 = θ1 - C1 represents the deviation in the heading alignment dimension (unit: DDM). A positive ∆θ1 indicates that the current heading alignment value is greater than the optimal value, requiring adjustment in the direction of reduction; a negative ∆θ1 indicates the opposite. To generate precise adjustment commands, the following components are extracted from the multidimensional displacement deviation: heading alignment dimension displacement deviation ∆θ1, heading width dimension displacement deviation ∆θ2, glide slope dimension displacement deviation ∆θ3, and glide slope width dimension displacement deviation ∆θ4. These four are core parameters that directly affect navigation accuracy. A pre-constructed hybrid model is used for calculation. The hybrid model includes a micro-ampere-DDM conversion model combined with a linear negative feedback model and a logarithmic nonlinear correction model. The model selection is based on the physical response characteristics of the parameters. For the heading alignment dimension, a microampere-DDM conversion model combined with a linear negative feedback model is adopted. The physical basis of this model is that the heading alignment deviation reported by the aircraft crew is in microamperes (μA), while the CLDDM setting value of the ground transmitter is in DDM. There is a conversion relationship of approximately 10 μA to 1 DDM. The operational logic of the microampere-DDM conversion model combined with the linear negative feedback model is as follows: First, the heading alignment microampere deviation value reported by the crew is converted into a DDM deviation value. Then, the deviation between the current heading alignment DDM value and the target center value C1 is calculated. This deviation is multiplied by the damping coefficient λ and taken as negative, so that the adjustment direction is opposite to the deviation direction, ensuring that the parameters converge towards the target center. The specific calculation process is as follows: the heading alignment dimension adjustment component ∆1 is equal to the negative λ times the displacement deviation component ∆θ1, that is, ∆1 = -λ × ∆θ1. The initial value of the damping coefficient λ is set to 0.1. This value is the best empirical value summarized from long-term flight calibration practice. At the same time, the value of λ will be adaptively updated in step S33 according to the measured deviation.

[0026] For the glide slope dimension, a microampere-DDM conversion model combined with a linear negative feedback model is employed. The physical basis of this model is that the glide slope deviation reported by the aircraft crew is in microamperes (μA), while the ground transmitter's adjustment parameters are in DDM, with an approximate conversion of 10 μA to 1 DDM. The operational logic of the microampere-DDM conversion model combined with the linear negative feedback model is as follows: First, the microampere deviation value of the glide slope reported by the crew is converted to a DDM deviation value. Then, the deviation between the current glide slope DDM value and the target center value C3 is calculated, and this deviation is multiplied by the damping coefficient λ and negative. The specific calculation formula is: glide slope dimension adjustment component ∆3 = -λ × ∆θ3. The adjustment logic for the lower limit of the glide slope is the same as that for the glide slope.

[0027] For the heading width and glide slope dimensions, a logarithmic nonlinear correction model is adopted. The physical basis of this model is that the width parameter is determined by the sideband power. The ground transmitter changes the measured width angle value by adjusting the SBO (sideband signal) attenuation value (in dB). There is a logarithmic physical relationship between the attenuation value (dB) and the actual width angle value. According to the power propagation theory of radio signals, the power expression in decibels is itself logarithmic, i.e., dB = 20 × lg (voltage ratio) or dB = 10 × lg (power ratio), where lg() is a logarithmic function with a fixed base of 10. When adjusting the sideband signal attenuation value, the change in the width angle value and the change in the attenuation value follow a logarithmic rather than a nonlinear relationship. The operational logic of the logarithmic nonlinear correction model is as follows: First, calculate the ratio k of the measured width angle value to the standard width angle value in the high-confidence real-time flight verification net value. Perform a logarithmic operation on the ratio k to adapt to the nonlinear relationship. Then, introduce a smoothing correction coefficient μ for scaling, finally obtaining the attenuation adjustment amount. The specific calculation formula is as follows: First, calculate the width ratio k = measured width angle value / standard width angle value, where the standard width angle value is the center value of the target center coordinates in the width dimension, C2 (for heading width) or C4 (for glide slope width); then calculate the attenuation adjustment ΔdB = μ × 20 × lg(k), where 20 × lg(k) is an industry-known expression reflecting the relationship between the attenuation value and the width angle value, and the initial value of the smoothing correction coefficient μ is set to 2 / 3. The function of the smoothing correction coefficient μ is to suppress overshoot in nonlinear adjustment, and its value of 2 / 3 is to ensure convergence speed while avoiding overshoot, thus achieving smooth adjustment. The heading width dimension adjustment component ∆2 is equal to the negative attenuation adjustment multiplied by the angle conversion factor ζ, that is, ∆2 = -μ × 20 × lg(k) h )×ζ, where k h=Measured heading width angle value / Standard heading width angle value. The initial value of ζ is set to 0.05° / dB. This value is determined based on the physical characteristic that a 1dB change in attenuation value in a typical ILS device corresponds to a change in the heading width angle value of approximately 0.05°. ζ is determined through device calibration, converting the logarithmic calculation result into an adjustment amount in angle units. The introduction of a negative sign ensures that the adjustment amount is negative when the measured width is greater than the standard width and positive when the measured width is less than the standard width, causing the width parameter to converge towards the target center. The value of ζ will be adaptively updated in step S33 based on the measured deviation. The glide width dimension adjustment component ∆4 is equal to the negative attenuation value adjustment amount multiplied by the angle conversion factor ζ, i.e., ∆4 = -μ × 20 × lg(k x )×ζ, where k x =Measured angle value of sliding width / Standard angle value of sliding width

[0028] Finally, the calculated heading straightness dimension adjustment component ∆1 (in DDM), heading width dimension adjustment component ∆2 (in °), glide slope dimension adjustment component ∆3 (in DDM), and glide slope width dimension adjustment component ∆4 (in °) are vectorized to obtain the initial suggested adjustment vector ∆=[∆1,∆2,∆3,∆4]. This vector contains the core adjustment information that moves the equipment state toward the target center.

[0029] Step S21 selects a suitable computational model based on the physical characteristics of the parameters, ensuring the accuracy of the initial adjustment. It avoids the adaptation defects of traditional linear models to nonlinear parameters, enabling the adjustment to respond quickly to deviations without inaccurate adjustments due to model mismatch. Furthermore, the calibration mechanism for the damping coefficient and smoothing correction coefficient gives the model cross-device adaptability, solving the problem of poor universality of traditional empirical parameters. Without step S21, subsequent virtual simulations would lack a clear adjustment target, leading to aimless debugging. The absence of a hybrid model would also cause the adjustment to fail to adapt to the physical characteristics of the parameters, further increasing the risk of debugging deadlock.

[0030] Step S22: Input the initial suggestion adjustment vector into the multidimensional coupled virtual state space for simulation and deduction, simulate the physical response after the change of the current ground transmitter's launch parameter settings, and obtain the theoretical state vector and the predicted values ​​of each dimension; The built-in physical coupling model is the core of the simulation. Based on the electrical principles of the ILS equipment, this model includes dynamic coupling relationships between parameters, such as the impact of heading width adjustment on the heading width warning angle value, the correlation between heading straightening changes and straightening warnings, and the impact of glide slope adjustment on the glide slope warning angle value. These coupling relationships are obtained by collecting a large amount of equipment operating data and performing regression analysis, ensuring that the model can accurately reproduce the actual operating characteristics of the equipment. The basic structure of the physical coupling model is a multi-input, multi-output coupling response matrix: the inputs are the adjustment components of each dimension in the initial suggested adjustment vector ∆=[∆1,∆2,∆3,∆4], and the outputs are the predicted state changes of each dimension after adjustment. The elements of the coupling response matrix represent the cross-influence coefficients between the adjustment amounts of different dimension parameters. For example, the cross-influence coefficient κ' of the heading width adjustment component ∆2 on the heading width warning angle value represents the change in the width warning angle caused by each unit attenuation value adjustment. This coefficient is determined through the regression relationship between attenuation value changes and width warning angle changes in historical calibration data. The specific process of simulation is as follows: The parameter adjustment amount corresponding to the initial suggested adjustment vector ∆ is applied to the current equipment state vector θ. The chain reaction after each parameter adjustment is calculated through the coupling response matrix to generate the theoretical state vector θpred=θ+∆+∆cross, where ∆cross is the additional state change caused by cross coupling. The components of each dimension of ∆cross are obtained by multiplying the coupling response matrix and the initial suggested adjustment vector. The focus is on extracting the predicted values ​​of each dimension from the theoretical state vector θpred, including the heading straight DDM predicted value (which can be converted into the heading straight microampere predicted value), the heading width angle predicted value, the glide slope DDM predicted value (which can be converted into the glide slope angle microampere predicted value), and the glide slope width angle predicted value. These predicted values ​​are the core indicators for the equipment to determine whether to trigger an alarm. If their values ​​exceed the boundary of the standard working geometric polyhedron, it means that the equipment will trigger an alarm. This method can predict the adjusted equipment status in advance, breaking the traditional "adjust first, observe later" delayed feedback mode; it provides a quantitative basis for subsequent collision detection, enabling potential alarm shutdown risks to be identified before adjustment is executed, avoiding equipment downtime during debugging. At the same time, the simulation is based on a multi-dimensional coupled virtual state space, ensuring the accuracy of the prediction results and solving the defect that the influence of parameter coupling cannot be predicted in traditional debugging.

[0031] See Figure 3 Step S23: Perform collision detection between the predicted values ​​of each dimension and the standard working geometric polyhedron. See Figure 4 This is a schematic diagram of collision detection and determination provided in an embodiment of this application. Figure 4The diagram visually illustrates the safety boundary region enclosed by the upper threshold, lower threshold, narrow alarm threshold, and wide alarm threshold, as well as the geometrical relationship between the adjusted state trajectory and this region. The collision detection is used to determine whether the predicted values ​​of each dimension derived from the initial suggested adjustment vector overflow the standard working geometric polyhedron, i.e., whether the adjusted state point overflows the standard range. The specific collision detection methods include: comparing the predicted heading width angle with the heading narrow warning threshold and heading width warning threshold of the standard working geometry polyhedron; if the predicted heading width angle is less than or greater than the heading narrow warning threshold, a collision is determined to have occurred; comparing the predicted heading straightening distance (DDM) value with the straightening warning threshold; if the predicted heading straightening distance (DDM) value exceeds the straightening warning threshold range, a collision is determined to have occurred; comparing the predicted glide slope width angle with the glide slope narrow warning threshold and glide slope width warning threshold; if the predicted glide slope width angle is less than or greater than the glide slope width warning threshold, a collision is determined to have occurred; and comparing the predicted glide slope angle (DDM) value with the glide slope lower limit and glide slope upper limit; if the predicted glide slope angle (DDM) value is less than or greater than the glide slope upper limit, a collision is determined to have occurred. Figure 4 As shown, the adjusted state trajectory simulates the parameter change path. When the predicted value, as indicated by the point where it overflows the boundary in the figure, falls outside the wide alarm threshold, a collision is determined to have occurred. A collision indicates that adjusting according to the initial suggested adjustment vector will cause the equipment parameters to exceed the limits, triggering an alarm and requiring the activation of anti-deadlock. Conversely, when the predicted value, as indicated by the point where it lies within the boundary in the figure, does not touch any threshold boundary, a collision is determined to have not occurred, indicating that the adjustment amount corresponding to the initial suggested adjustment vector is within a safe range. The alarm thresholds for heading straightness and glide slope angle dimensions are in DDM units (convertible to μA), while the alarm thresholds for heading width and glide slope width dimensions are in degrees (°). These are based on the boundary values ​​of the standard working geometric polyhedron determined by civil aviation regulations in each dimension, ensuring that the collision detection judgment criteria are consistent with the actual alarm logic of the equipment. This method can accurately identify the adjusted alarm risks, clarify the triggering conditions of the anti-deadlock logic, and transform the abstract parameter over-limit risks into intuitive geometric collision judgments. This gives risk identification clear quantitative standards, avoids the subjectivity and uncertainty of relying on human experience to judge risks in traditional debugging, and ensures the forward-looking nature of risk identification based on virtual pre-simulation results, thus gaining adjustment space for subsequent anti-deadlock corrections.

[0032] Step S24: If a collision occurs, activate the deadlock prevention logic and generate a proposed adjustment strategy to prevent deadlock as a candidate adjustment strategy; if no collision occurs, use the initial proposed adjustment vector as a candidate adjustment strategy. The deadlock prevention logic includes: performing safety clamping on the initial suggested adjustment vector to generate a safety compromise adjustment vector, calculating a threshold linkage adjustment vector based on the safety compromise adjustment vector, and combining the safety compromise adjustment vector and the threshold linkage adjustment vector to generate a deadlock prevention suggested adjustment strategy.

[0033] Methods for applying safety clamping to the initial proposed adjustment vector include: reducing the smoothing correction coefficient in the logarithmic nonlinear correction model, simultaneously reducing the damping coefficient of the microampere-DDM conversion model combined with the linear negative feedback model, and reducing the adjustment step size.

[0034] Step S24 generates an adjustment strategy based on the collision detection results: If a collision occurs, the anti-deadlock logic is activated, and a proposed anti-deadlock adjustment strategy is generated through a combination of safety clamping and threshold linkage, serving as a candidate adjustment strategy; if no collision occurs, the initial proposed adjustment vector is used as a candidate adjustment strategy and enters the hysteresis silent zone filtering in subsequent step S30. The core of the anti-deadlock logic is to decouple deadlock caused by parameter coupling through the synergy of "limiting the adjustment step size" and "dynamic adjustment boundary". The specific implementation of the safety clamping process is as follows: reducing the smoothing correction coefficient μ in the logarithmic nonlinear correction model, reducing the adjustment step size in the heading width dimension and glide width dimension, and generating a safety compromise adjustment vector ∆'. The adjustment range of the smoothing correction coefficient μ is determined based on the over-limit amount during collision detection: the over-limit amount ∆over between the predicted value and the triggered alarm threshold boundary value is calculated (the over-limit amount in the heading straight and glide slope dimensions is in DDM, and the over-limit amount in the heading width and glide slope dimensions is in angle). The ratio η = ∆over / ∆step between the over-limit amount ∆over and the change amount ∆step corresponding to the original adjustment step size is calculated. This ratio η represents the proportion of the over-limit degree to the adjustment step size; ∆step is the change amount corresponding to the adjustment component of the initial suggested adjustment vector in the dimension where the collision occurred. The adjusted smoothing correction coefficient μ' = μ × (1 - η - q), where q is the safety margin coefficient, and the value of q ranges from 0.05 to 0.15, ensuring that the adjusted state point not only does not exceed the limit, but also maintains a certain safe distance from the boundary; to avoid the adjustment failure due to μ' being too small, when the calculated μ' is lower than the preset lower limit value of 0.1, μ' is forcibly set to 0.1. For example, if the initial μ value is 2 / 3, the collision detection judgment will overflow the heading narrow warning threshold after adjustment. The over-limit amount ∆over is 30% of the change amount ∆step corresponding to the adjustment step size (i.e., η=0.3), and the safety margin coefficient q is taken as 0.1. Then, the adjusted smoothing correction coefficient μ'=2 / 3×(1-0.3-0.1)=0.4. The heading width dimension adjustment component is recalculated and defined as the heading width dimension adjustment component for deadlock prevention ∆2'=-0.4×20×lg(k)×ζ, so that the adjustment step size is reduced and the adjusted state point is ensured to fall within the boundary of the standard working geometry polyhedron. The glide width dimension adjustment component is recalculated using the same method and defined as the glide width dimension adjustment component for deadlock prevention ∆'4.

[0035] Simultaneously, safety clamping is performed on both the heading alignment dimension and the glide slope dimension. For the heading alignment dimension, the adjusted damping coefficient λ' = λ × (1 - η - q). When the calculated λ' is lower than the preset lower limit of 0.02, λ' is forcibly set to 0.02, and the anti-lockdown heading alignment dimension adjustment component ∆'1 = -λ' × ∆θ1 is recalculated. For the glide slope dimension, the same adjusted damping coefficient λ' is used, and the anti-lockdown glide slope dimension adjustment component ∆'3 = ∆3 × λ' is recalculated to ensure that the adjustment step size of each dimension is reduced in a coordinated manner, avoiding new coupling overlimits caused by excessively large step sizes in a single dimension. After the safety compromise adjustment vector ∆' is generated, the physical coupling model in step S22 is called again for simulation. The safety compromise adjustment vector ∆'=[∆'1,∆'2,∆'3,∆'4] is used as input to replace the initial proposed adjustment vector ∆. The additional state change ∆'cross caused by cross coupling is calculated through the coupling response matrix, and the theoretical state vector θ'pred=θ+∆'+∆'cross based on the safety compromise adjustment vector is generated. This theoretical state vector θ'pred will replace the θpred originally generated in step S22 and be used for closed-loop verification and model parameter correction in the subsequent step S33.

[0036] The specific implementation of threshold linkage adjustment is as follows: The threshold linkage adjustment vector ∆th is calculated synchronously. While adjusting the centerline parameters (including heading alignment parameters and glide slope parameters) and width parameters (including heading width parameters and glide slope width parameters), the boundary position of the standard working geometric polyhedron is dynamically shifted, i.e., the alarm threshold value is adjusted. For example, when it is necessary to increase the heading width and it is about to touch the upper boundary of the heading narrow alarm, the threshold linkage adjustment vector ∆th indicates that the heading narrow alarm threshold angle value is shifted upward by a certain amount, so that the adjusted heading width angle prediction value is still within the shifted boundary. After the centerline parameters and width parameters are adjusted to be close to the target center, the alarm threshold is gradually pulled back to the original standard value. The design basis of threshold linkage is: during the debugging process, temporarily relaxing the alarm threshold aims to provide a larger safety margin for parameter adjustment and avoid adjustment deadlock caused by fixed boundaries. The subsequent pullback mechanism ensures that the equipment ultimately operates within the range required by civil aviation standards and will not affect navigation safety. The strategy combination of the safety compromise adjustment vector and the threshold linkage adjustment vector is as follows: the corresponding adjustment instructions of the two are bound together to form an anti-deadlock suggested adjustment strategy. This ensures that the centerline and width parameter adjustments are executed synchronously with the threshold adjustments, avoiding new risks caused by adjusting the centerline and width parameters or the threshold separately. If no collision occurs, the initial suggested adjustment vector does not need to be modified and is directly used as a candidate adjustment strategy because the adjustment amount corresponding to the initial suggested adjustment vector has adapted to the physical characteristics of the parameters and has not exceeded the safety boundary, and can directly enter the subsequent steady-state determination stage.

[0037] Step S24, through the synergy of step size limitation and dynamic boundary adjustment, resolves the deadlock caused by parameter coupling while ensuring the safety of the adjustment. This breaks the traditional debugging process's vicious cycle where "adjusting a single parameter easily triggers other parameters exceeding alarm thresholds" due to strong parameter coupling, achieving smooth adjustment of coupled parameters. Simultaneously, the combined strategy of safety clamping and threshold linkage balances debugging efficiency and safety, avoiding slow convergence or boundary control issues caused by a single strategy. Without step S24, a safe adjustment strategy cannot be generated when a collision occurs; either initial adjustment will cause equipment shutdown, or adjustment will be abandoned, causing debugging to stall, making effective debugging of coupled parameters impossible.

[0038] Step S30: The candidate adjustment strategies are steadily determined and filtered through a preset hysteresis quiescent zone, and encapsulated into an atomic execution instruction set; the atomic execution instruction set is executed and a new round of current device state vector is obtained, and closed-loop verification and hybrid model adaptive update are performed based on the new round of current device state vector.

[0039] Further, step S30 includes: Step S31, see Figure 5 Extract the adjustment amount of each dimension parameter in the candidate adjustment strategy, calculate the expected state value after each dimension adjustment, and if the expected state value of a certain dimension falls within the hysteresis quiescent zone, then the adjustment amount of the parameter of that dimension is forcibly set to zero; if it does not fall within the hysteresis quiescent zone, then the adjustment amount of the parameter of that dimension is retained, and valid adjustment data is generated. The hysteresis quiet zone is a tolerance area defined around the target center coordinates of a standard working geometric polyhedron. Defined by the hysteresis quiet zone threshold, it appears as a closed neighborhood with the target center coordinates as its geometric center in the multidimensional coupled virtual state space. Physically, it defines the "steady-state convergence boundary" of equipment parameters: when equipment parameters are already sufficiently close to the target optimal state, further minor adjustments not only fail to produce substantial improvement but also cause repeated oscillations of parameters near the target value, similar to underdamped oscillations in a mechanical system, increasing hardware wear and extending calibration time. The hysteresis quiet zone threshold is determined based on the steady-state characteristics of the equipment operation and the allowable minor deviations according to civil aviation standards. By collecting parameter fluctuation data from multiple instrument landing systems under stable operating conditions, the average peak value of parameter fluctuations in each dimension is statistically analyzed. 1.2 times this average value is set as the corresponding dimension's hysteresis quiet zone threshold, ensuring that the hysteresis quiet zone threshold filters out meaningless minor fluctuations without affecting adjustments for normal deviations.

[0040] During the determination process, the adjustment amounts of each dimension parameter in the candidate adjustment strategy are first extracted: when the candidate adjustment strategy is the initial suggested adjustment vector, the adjustment amounts of each dimension parameter in the candidate adjustment strategy include the heading straightness dimension adjustment component ∆1, the heading width dimension adjustment component ∆2, the glide slope dimension adjustment component ∆3, and the glide slope width dimension adjustment component ∆4; when the candidate adjustment strategy is the deadlock prevention suggested adjustment strategy, the adjustment amounts of each dimension parameter include the deadlock prevention heading straightness dimension adjustment component ∆'1, the deadlock prevention heading width dimension adjustment component ∆'2, the deadlock prevention glide slope dimension adjustment component ∆'3, the deadlock prevention glide slope width dimension adjustment component ∆'4, and the threshold linkage adjustment vector ∆th.

[0041] The expected state values ​​after adjustments in each dimension are calculated by adding the corresponding adjustment amounts to the components of the current device state vector θ: θ'1 = θ1 + ∆1 (or θ1 + ∆'1) for the heading straightening dimension, θ'2 = θ2 + ∆2 (or θ2 + ∆'2) for the heading width dimension, θ'3 = θ3 + ∆3 (or θ3 + ∆'3) for the glide slope angle dimension, and θ'4 = θ4 + ∆4 (or θ4 + ∆'4) for the glide slope width dimension. The expected state values ​​represent the state the device parameters will reach after executing the current adjustment command. The expected state values ​​in each dimension are compared with the corresponding components of the target center coordinates to determine whether the expected state value falls within the hysteresis quiet zone. The criterion is |θ' i -C i |<ε i , where θ' i Let C be the expected state value in the i-th dimension. i Let ε be the component of the target center coordinates in the i-th dimension. i Let be the threshold for the hysteresis quiescent zone in the i-th dimension. If a certain dimension satisfies |θ' i -C i |<ε i If the expected state value of that dimension has fallen into the hysteresis quiescent region, it indicates that the adjusted parameter of that dimension is close enough to the target center, and the device is in a steady state in that dimension. Continuing to perform adjustments will cause meaningless perturbations, so the adjustment amount of that dimension is forcibly set to zero. If a dimension does not meet the above conditions, i.e., |θ' i -C i |≥ε i If the value is zero, it indicates that the expected state value for that dimension has not yet entered the steady-state region, and the adjustment amount for that dimension is retained. After completing the above determination for all dimensions, the retained non-zero adjustment amounts are summarized to form the final effective adjustment data.

[0042] Step S31 filters out invalid fine-tuning commands, avoiding the problem of frequent perturbations to the equipment in pursuit of absolute mathematical zero error in traditional debugging. It adds a "mechanical damping" characteristic to the system, automatically suppressing adjustment actions and reducing parameter oscillations when parameters approach the target. Through steady-state determination and filtering, the stability of equipment operation is significantly improved, avoiding wear and tear on transmitter hardware caused by frequent adjustments, and reducing the wasted calibration time due to invalid adjustments. This lays a precise foundation for subsequent atomic command execution. Without this step, frequent adjustments caused by minor fluctuations would lead to repeated parameter oscillations around the target value, not only failing to achieve steady-state debugging but also potentially causing parameters to deviate from the safety boundary due to continuous adjustments, rendering the virtual pre-simulation and deadlock prevention logic in step S20 meaningless.

[0043] Step S32: Encapsulate the effectively adjusted data into an atomic execution instruction set; The centerline adjustment commands, width adjustment commands, and threshold-linked adjustment commands for alarm thresholds from the effective adjustment data are packaged to construct an indivisible multi-parameter collaborative atomic execution command set. The effective adjustment data filtered in step S31 is then encapsulated into a multi-parameter collaborative atomic execution command set, which includes centerline adjustment commands for heading straightening and glide slope angle, width adjustment commands for heading width and glide slope width, and threshold-linked adjustment commands for width and narrowness alarm thresholds. An atomic execution command set is a command encapsulation form that binds multiple interrelated parameter adjustment commands into an indivisible execution unit. Its core feature is that all adjustment commands within the command set must be synchronously sent to the control interface of the ground transmitter as a whole, and operators are not allowed to selectively execute some commands while ignoring others. The design basis of atomized execution stems from the strong coupling characteristics between parameters of the Instrument Landing System (ILS): adjusting the heading width inevitably causes a change in the heading width alarm DDM value, and the change in the width alarm DDM value directly affects the triggering state of the alarm threshold; adjusting the glide width inevitably causes a change in the glide width alarm DDM value; adjusting the heading straightening affects the triggering state of the straightening alarm; adjusting the glide angle affects the triggering state of the glide angle lower limit. If the operator is allowed to execute only some commands, such as only adjusting the heading width while ignoring the linkage correction of the alarm threshold, the adjusted width DDM value may immediately trigger an alarm because the threshold is not updated synchronously, causing the commissioning to be interrupted and reverting to the state before the adjustment. This is precisely the human oversight problem that frequently occurs in traditional commissioning.

[0044] The generation of centerline adjustment and width adjustment commands is based on the valid adjustment data output in step S31. For the width adjustment command: For the heading width dimension and glide slope width dimension, the width adjustment command includes a target adjustment amount of attenuation (dB) derived from the valid adjustment data. This adjustment amount is calculated from the width dimension adjustment components ∆2 / ∆'2 and ∆4 / ∆'4 using the angle conversion coefficient ζ (° / dB), and then superimposed with the current attenuation setting of the ground transmitter to obtain the target attenuation value. When a new attenuation value is written, the width value on the software interface changes accordingly, and the width angle value measured by the aircraft in the air also changes accordingly. For the centerline adjustment command: For the heading alignment dimension and glide slope angle dimension, the centerline adjustment command includes a target adjustment value for the CLDDM (Clearing, Distance, and Difference) adjustment, which is obtained by adding the current CLDDM setting value of the ground transmitter to the corresponding dimension adjustment component. The target values ​​of the centerline adjustment command and the width adjustment command are encapsulated in absolute value form rather than incremental form. The choice of absolute value form is to avoid incremental cumulative error: in the process of multi-round iterative debugging, if incremental form is used, the execution accuracy error of the control interface will accumulate with the number of adjustments, resulting in unpredictable deviations between the actual parameters and the target parameters; the absolute value form ensures that each adjustment is based on the target center coordinates, eliminating the possibility of error accumulation.

[0045] The threshold linkage adjustment command is generated based on the threshold linkage adjustment vector ∆th calculated in step S24. When a collision is determined to have occurred in step S23, the threshold linkage adjustment vector ∆th indicates that the heading narrow warning threshold, heading wide warning threshold, glide slope narrow warning threshold, or glide slope wide warning threshold will be temporarily shifted. The shift direction is consistent with the adjustment direction of the centerline parameter and the width parameter, and the shift magnitude is determined by the excess amount of the predicted value and the original threshold boundary during collision detection. For example, if the heading width predicted value exceeds the heading narrow warning threshold angle value by ∆over, the threshold linkage adjustment command will temporarily shift the heading narrow warning threshold upward by ∆over×γ, where γ is the threshold shift safety factor, and the value of γ ranges from 1.05 to 1.20. The threshold linkage adjustment command also includes the threshold callback trigger condition. When the centerline parameter and width parameter are adjusted and all dimensions of the current device state vector fall into the hysteresis quiet zone defined in step S31, it indicates that the device has entered a steady state. At this time, the system automatically generates a threshold callback command to gradually restore the temporarily shifted alarm threshold to the standard value specified by civil aviation regulations. The threshold callback adopts a gradual adjustment rather than a one-step adjustment. Each callback amplitude is one-quarter to one-third of the original shift amplitude to avoid new alarm triggering caused by sudden changes in the threshold.

[0046] The core idea of ​​the linkage mechanism is revised as follows: the width alarm threshold and the width parameter (DSDDM / sliding DSDDM) are coupled in the control logic. The threshold can be implemented using either an "absolute threshold value" or a "relative difference". In this embodiment, the standard working geometric polyhedron boundary constructed in step S12 is used as the benchmark. The threshold linkage adjustment is uniformly described as the synchronous translation and callback of the alarm threshold boundary value to ensure that collision detection and threshold linkage are consistent in the closed loop under the same boundary semantics. Specifically, when the width adjustment command causes a change in the width dimension parameter, the system synchronously calculates and issues the corresponding threshold linkage adjustment command, so that the adjusted alarm threshold boundary value is consistent with the width dimension change in the multi-dimensional coupled virtual state space, thereby avoiding transient out-of-bounds alarms caused by "width has been adjusted but the boundary is not synchronized".

[0047] The encapsulation order of various instructions in the atomic execution instruction set follows the causal logic of parameter adjustment. Threshold linkage adjustment instructions, as the pre-processing part of the instruction set, ensure that the alarm boundary has been temporarily extended before the centerline and width parameter adjustments are executed, providing a safety margin for these adjustments. Centerline and width adjustment instructions, as the core parts of the instruction set, execute the target value writing for heading straightening CLDDM, glide slope CLDDM, and heading and glide slope parameters. All instructions are bound together using a unified timestamp and executed synchronously. Upon receiving the instruction set, the ground transmitter's control interface completes the execution of all instructions within a single control cycle, ensuring that parameter changes take effect synchronously within a very short time.

[0048] Step S32 encapsulates the effective adjustment data into an atomic execution instruction set, merging multiple adjustment steps that previously required separate operations by the operator into a single operation. This fundamentally eliminates human error caused by operators forgetting to adjust a related parameter, leading to equipment alarms in traditional debugging. The indivisible nature of atomic encapsulation ensures the integrity of parameter adjustments. When the operator triggers the execution instruction, the centerline parameter, width parameter, and alarm threshold are updated synchronously. Any change to any parameter is accompanied by the coordinated correction of related parameters, avoiding the risk of transient out-of-bounds errors caused by timing misalignments in parameter adjustments. The pre-execution and progressive callback mechanism of the threshold-linked adjustment instruction ensures that the centerline and width parameter adjustments are always performed within the extended safety boundaries. Even if the adjustment step size approaches the original threshold boundary, the temporarily shifted threshold can accommodate the adjusted DDM value, achieving zero alarms during the adjustment process. The deadlock prevention suggestion adjustment strategy generated in step S32 and step S24 works synergistically. Step S24 ensures that the adjustment amount is within a safe range through a safety clamping and threshold linkage strategy, while step S32 transforms this strategy into a directly executable instruction form. Together, they constitute a complete link from strategy generation to strategy execution, enabling the deadlock prevention logic to be effectively implemented. Without step S32, the candidate adjustment strategy generated in step S24 would only remain at the numerical level. The operator would need to manually input the adjustment amounts for each dimension into the transmitter control interface, which not only increases operational complexity but also fails to guarantee the atomicity of parameter adjustments due to the lack of mandatory synchronous execution constraints, causing the deadlock prevention mechanism built in the preceding steps to fail during execution.

[0049] Step S33: After the atomic execution instruction set is executed, the updated instrument landing system flight verification original data sequence is received and a new round of current equipment state vector is regenerated. The deviation between the new round of current equipment state vector and the theoretical state vector is calculated. Based on the deviation, the damping coefficient of the microampere-DDM conversion model combined with the linear negative feedback model and the smoothing correction coefficient of the logarithmic nonlinear correction model are corrected.

[0050] Specifically, the execution of the atomized execution instruction set is completed through the control interface of the ground transmitter. The control interface writes the target values ​​of each item in the instruction set into the transmitter's parameter register, and the transmitter adjusts the transmission characteristics of the radio frequency signal according to the new parameter configuration. After the atomized execution instruction set is completed, the system enters a waiting state, waiting for the calibration aircraft to complete a new round of airborne signal acquisition in the next approach segment and transmit the updated Instrument Landing System (ILS) flight calibration raw data sequence. The updated ILS flight calibration raw data sequence reflects the actual signal transmission performance of the ground transmitter after parameter adjustment. This data sequence is re-acquired by the airborne ILS signal sensor of the calibration aircraft and transmitted after spatial registration with the real-time position information of the aircraft provided by the satellite positioning system. The acquisition and processing of the data sequence follows the process of step S11: receiving the raw data transmitted by the calibration aircraft through the standardized data interface, using Fast Fourier Transform to remove high-frequency noise components, applying medium-range filtering to smooth abrupt changes, and obtaining a new round of high-confidence real-time flight calibration net values.

[0051] The generation of the new round of current equipment state vector follows the process of step S13: the net value of the new round of high-confidence real-time flight verification is fused with the adjusted ground transmitter launch parameter settings. Using a multi-dimensional parameter normalization mapping method, the new round of current equipment state vector θnew is located in the multi-dimensional coupled virtual state space. The components of each dimension of θnew correspond to the adjusted heading alignment state value θnew1, heading width state value θnew2, glide slope state value θnew3, and glide slope width state value θnew4, respectively. The new round of current equipment state vector θnew represents the actual operating state of the equipment after parameter adjustment and forms a comparison relationship with the theoretical state vector derived from the physical coupling model in the previous steps: if the collision detection judgment in step S23 does not result in a collision, the comparison object is the theoretical state vector θpred generated in step S22; if the collision detection judgment in step S23 results in a collision, the comparison object is the theoretical state vector θ'pred re-derived in step S24. The theoretical state vector is the theoretical expectation of the adjusted state before adjustment, while θnew is the actual measurement result after adjustment. The difference between the two reflects the prediction accuracy of the physical coupling model.

[0052] The specific method for calculating the deviation is as follows: The current device state vector θnew of the new round is compared dimension-by-dimensionally with the theoretical state vector (θpred or θ'pred) generated in the previous step, and the prediction deviation for each dimension is calculated. The theoretical state vector has already been calculated based on the physical coupling model in step S22 or S24. It includes the direct effect of the adjustment vector and the cross-coupling effect obtained from the coupling response matrix calculation, representing the adjusted ideal state predicted based on the physical coupling model. The formula for calculating the prediction deviation for each dimension is e. i =θnewi -θpred i (or e) i =θnew i -θ'pred i ), where e i Let θnew be the prediction bias in the i-th dimension. i Let θpred be the component of the current device state vector in the i-th dimension for the new round. i (or θ'pred) i ) represents the component of the theoretical state vector in the i-th dimension.

[0053] The model parameter correction based on prediction bias employs an adaptive update mechanism. The correction targets are the damping coefficient λ (initial value 0.1) of the micro-ampere-DDM conversion model combined with the linear negative feedback model used in step S21, and the smoothing correction coefficient μ (initial value 2 / 3) of the logarithmic nonlinear correction model. The correction logic for the damping coefficient λ is as follows: calculate the ratio r1 = e1 / ∆1 of the prediction bias e1 in the heading correction dimension and the adjustment component ∆1 in that dimension. r1 reflects the proportional relationship between the prediction bias and the adjustment magnitude. If r1 is positive and has a large absolute value, it indicates that the model systematically underestimates the heading correction effect, and the actual state after adjustment exceeds expectations. In this case, the damping coefficient λ needs to be reduced to decrease the step size of subsequent adjustments. If r1 is negative and has a large absolute value, it indicates that the model systematically overestimates the effect, and the actual state after adjustment does not meet expectations. In this case, the damping coefficient λ needs to be increased to accelerate the convergence speed. The update formula for the damping coefficient is λ. new =λ×(1-α×r1), where α is the learning rate coefficient. The value of α determines the sensitivity of the model parameters to single bias. Too large an α will cause the parameters to fluctuate drastically, while too small an α will cause the adaptive convergence to be slow. The range of α was determined to be 0.05 to 0.15 through verification of multiple sets of calibration flight data, so as to ensure that the parameter update achieves a balance between stability and response speed.

[0054] The correction logic for the smoothing correction coefficient μ is as follows: Calculate the ratio r2 = e2 / ∆2 of the prediction deviation e2 in the heading width dimension to the adjustment component ∆2 in that dimension, and the ratio r4 = e4 / ∆4 of the prediction deviation e4 in the glide slope width dimension to the adjustment component ∆4 in that dimension, and take the average of the two, r avg =(r2+r4) / 2 is used as the comprehensive prediction bias ratio for the width dimension, r avg This reflects the overall prediction accuracy of the logarithmic nonlinear correction model. Since both the heading width and glide slope width are changed by adjusting the attenuation value (dB) to alter the DSDDM value, and there is a logarithmic physical relationship between the attenuation value and the actual width DDM, prediction bias may originate from two aspects: firstly, the linearization error of the logarithmic operation itself; and secondly, the degree of nonlinearity between the attenuation value and the actual width DDM changes with equipment aging. If r avgA consistently positive value indicates that the smoothing correction coefficient μ is too large, and the adjustment step size exceeds the actual response capability of the device. In this case, the step size should be reduced by decreasing μ; if r avg A persistently negative value for μ indicates that μ is too small, and adjusting the step size is insufficient to drive the parameters towards the target convergence. In this case, increasing μ can accelerate convergence. The update formula for the smoothing correction coefficient is μ. new =μ×(1-β×r avg ), where β is the learning rate coefficient for nonlinear adjustment. The value of β is slightly larger than α due to the special nature of logarithmic operations, ranging from 0.08 to 0.20, ensuring the model can adapt to the dynamic changes in the decay value-width DDM relationship in a timely manner. The correction logic for the angle conversion coefficient ζ is similar to that of the smoothing correction coefficient μ: since ζ is a conversion coefficient that converts the logarithmic result (dB) into a DDM unit adjustment, the prediction accuracy of ζ is also reflected in the prediction bias in the heading width dimension and glide slope width dimension. The update formula for the angle conversion coefficient is ζ. new =ζ×(1-β×r avg The updated value of ζ is limited to a preset upper and lower limit range, such as 0.001 to 0.01, to prevent the subsequent width adjustment from getting out of control due to a single abnormal deviation causing the value of ζ to be too large or too small.

[0055] The adaptive update mechanism also includes parameter boundary constraints: the updated value of the damping coefficient λ is limited to a preset upper and lower limit range (0.05 to 0.20) to prevent subsequent adjustments from becoming uncontrollable due to excessively large or small values ​​caused by a single abnormal deviation. The upper limit of λ is determined based on system stability conditions to ensure that the closed-loop control will not oscillate due to excessive gain. The lower limit of λ is determined based on convergence speed requirements to ensure that debugging can be completed within the specified calibration flight time. The smoothing correction coefficient μ is also subject to boundary constraints. The upper limit of μ is 1 to avoid the adjustment step size exceeding the theoretical value of logarithmic calculation, and the lower limit of μ is 0.3 to ensure that the adjustment amount has sufficient driving force to make the parameter leave the current state. The specific values ​​of the parameter boundaries are determined by a combination of offline simulation and online calibration: In the offline simulation stage, a simulation environment is built based on historical calibration flight data of multiple instrument landing systems, and the closed-loop control performance under different boundary values ​​is traversed to select the boundary range that meets the requirements for both convergence speed and stability; In the online calibration stage, the boundary values ​​are dynamically fine-tuned according to the update trajectory of the model parameters during the actual calibration flight process to ensure that the boundary constraints are adapted to the characteristics of specific equipment.

[0056] Step S33 constructs a complete closed loop from command execution to effect verification and model optimization, enabling the debugging system to self-correct based on measured data. The calculation of prediction deviation quantitatively compares the theoretical expectations of the physical coupling model with actual measurement results, ensuring that the model's prediction accuracy is no longer statically fixed but continuously approximates the actual physical characteristics of the equipment through accumulated calibration data. The adaptive update mechanism of the damping coefficient and smoothing correction coefficient allows the hybrid model to adapt to different models and aging levels of instrument landing systems, eliminating the shortcomings of traditional empirical coefficients that only fit specific equipment and become inaccurate when applied across different equipment. The introduction of learning rate coefficients α and β ensures that parameter updates can respond promptly to prediction deviations without over-correcting due to single abnormal data, guaranteeing the smoothness and robustness of the adaptive process. The setting of parameter boundary constraints, while endowing the model with adaptive capabilities, limits parameter changes within a safe range, avoiding the risk of parameter runaway caused by adaptive updates. Step S33 and step S21 form a collaborative closed loop. Step S21 uses the damping coefficient λ and the smoothing correction coefficient μ to generate an initial suggested adjustment vector. Step S33 corrects the values ​​of λ and μ based on the actual effect after adjustment. The corrected coefficients are used in step S21 in the next round of debugging. Together, they constitute an iterative loop of parameter calculation-execution verification-model optimization, which gradually improves the debugging accuracy with the increase of calibration rounds. Step S33 also forms a verification relationship with the virtual pre-run of step S22. Step S22 predicts the adjusted predicted values ​​of each dimension based on the physical coupling model. Step S33 evaluates the accuracy of the physical coupling model by comparing the predicted values ​​with the measured values. If the deviation persists and shows regularity, it indicates that the physical coupling model needs structural optimization rather than just coefficient adjustment, providing data support for the long-term maintenance of the model. If step S33 is missing, the damping coefficient and smoothing correction coefficient in step S21 will always remain at their initial values, which will not be able to adapt to the drift of parameter response characteristics caused by factors such as equipment aging and environmental changes. After long-term use, the model prediction accuracy will gradually decrease, eventually causing the adjustment amount to deviate from the optimal value, reducing debugging efficiency or even causing adjustment failure.

[0057] Example 2 This embodiment, based on Embodiment 1, provides an intelligent auxiliary debugging system for instrument landing system parameters, such as... Figure 6 As shown, it includes: State space positioning module: used to construct a multidimensional coupled virtual state space, define a standard working geometric polyhedron within the multidimensional coupled virtual state space; acquire the original data sequence of the instrument landing system flight verification, and locate the current equipment state vector of the current ground transmitter in the multidimensional coupled virtual state space based on the original data sequence of the instrument landing system flight verification. Strategy generation module: used to determine the target center coordinates of the standard working geometric polyhedron, calculate the vector difference between the current device state vector and the target center coordinates to obtain the multidimensional displacement deviation, generate an initial proposed adjustment vector based on the multidimensional displacement deviation and the pre-built hybrid model, substitute the initial proposed adjustment vector into the multidimensional coupled virtual state space to perform simulation deduction, and generate candidate adjustment strategies based on the deduction results; Closed-loop update module: This module performs steady-state determination and filtering of candidate adjustment strategies through a preset hysteresis quiescent zone, and encapsulates them into an atomic execution instruction set; it executes the atomic execution instruction set and obtains the current device state vector for the next round, and performs closed-loop verification and hybrid model adaptive update based on the new round of current device state vector.

[0058] Furthermore, in the state space positioning module, the method for locating the current device state vector of the current ground transmitter includes: receiving the original data sequence of the instrument landing system flight verification transmitted by the verification aircraft, removing high-frequency noise components and smoothing abrupt changes in the original data sequence of the instrument landing system flight verification, and obtaining a high-confidence real-time flight verification net value; combining the high-confidence real-time flight verification net value with the current ground transmitter's transmission parameter settings, and locating the current device state vector in the multi-dimensional coupled virtual state space through multi-dimensional parameter normalization mapping.

[0059] Furthermore, in the strategy generation module, the method for generating candidate adjustment strategies based on the deduction results includes: performing collision detection between the predicted values ​​of each dimension and the standard working geometric polyhedron; if a collision occurs, activating the deadlock prevention logic and generating a deadlock prevention suggested adjustment strategy as a candidate adjustment strategy; if no collision occurs, using the initial suggested adjustment vector as a candidate adjustment strategy.

[0060] Furthermore, in the closed-loop update module, the method for closed-loop verification and hybrid model adaptive update based on the new round of current device state vector includes: calculating the deviation between the new round of current device state vector and the theoretical state vector, and correcting the damping coefficient of the microampere-DDM conversion model combined with the linear negative feedback model and the smoothing correction coefficient of the logarithmic nonlinear correction model based on the deviation.

[0061] The methods and systems of this application may be implemented in many ways. For example, they may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above-described order of steps for the method is for illustrative purposes only, and the steps of the method of this application are not limited to the order specifically described above, unless otherwise specifically stated.

[0062] In addition, the parts of the technical solutions provided in the embodiments of this application that are consistent with the implementation principles of the corresponding technical solutions in the prior art have not been described in detail, so as to avoid excessive elaboration.

[0063] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the invention. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for intelligently assisting in the debugging of parameters of an instrument landing system, characterized in that, The method includes: Construct a multidimensional coupled virtual state space, and define a standard working geometric polyhedron within the multidimensional coupled virtual state space; obtain the original data sequence of the instrument landing system flight verification, and locate the current equipment state vector of the current ground transmitter in the multidimensional coupled virtual state space based on the original data sequence of the instrument landing system flight verification. Determine the target center coordinates of the standard working geometric polyhedron, calculate the vector difference between the current equipment state vector and the target center coordinates to obtain the multidimensional displacement deviation, generate an initial suggested adjustment vector based on the multidimensional displacement deviation and the pre-built hybrid model, substitute the initial suggested adjustment vector into the multidimensional coupled virtual state space to perform simulation deduction, and generate candidate adjustment strategies based on the deduction results; Candidate adjustment strategies are determined and filtered for steady-state conditions through a preset hysteresis quiescent zone, and then encapsulated into an atomic execution instruction set. The atomic execution instruction set is executed to obtain a new round of current device state vector, and closed-loop verification and hybrid model adaptive update are performed based on the new round of current device state vector.

2. The intelligent assisted debugging method for instrument landing system parameters according to claim 1, characterized in that, The multidimensional coupled virtual state space includes four dimensions: heading straightness dimension, heading width dimension, glide slope dimension, and glide slope width dimension. The method for generating the standard working geometric polyhedron includes: obtaining the upper limit and lower limit of the normal operating parameters of the current ground transmitter, and forming the standard working geometric polyhedron by enclosing the upper limit and lower limit of the normal operating parameters within a multi-dimensional coupled virtual state space.

3. The intelligent assisted debugging method for instrument landing system parameters according to claim 2, characterized in that, The method for locating the current device state vector of the current ground transmitter includes: Receive the original flight verification data sequence of the instrument landing system transmitted by the verification aircraft, remove the high-frequency noise components in the original flight verification data sequence of the instrument landing system and smooth the abrupt change points to obtain a high-confidence real-time flight verification net value. By combining the high-confidence real-time flight verification net value with the current ground transmitter's launch parameter settings, and through multi-dimensional parameter normalization mapping, the current device state vector is located in the multi-dimensional coupled virtual state space.

4. The intelligent assisted debugging method for instrument landing system parameters according to claim 3, characterized in that, The method for generating the initial proposed adjustment vector includes: The displacement deviation components of the heading straightness dimension, heading width dimension, glide slope angle dimension, and glide slope width dimension in the multidimensional displacement deviation are extracted. The displacement deviation components of each dimension are calculated separately, and the calculation results are vectorized to generate the initial suggested adjustment vector.

5. The intelligent assisted debugging method for instrument landing system parameters according to claim 4, characterized in that, The hybrid model includes a microampere-DDM conversion model combined with a linear negative feedback model and a logarithmic nonlinear correction model; The method for calculating the displacement deviation components in each dimension includes: For the displacement deviation components in the heading alignment dimension and glide slope dimension, the micro-ampere-DDM conversion model combined with the linear negative feedback model is used to calculate the heading alignment dimension adjustment component and glide slope dimension adjustment component; For the displacement deviation components in the heading width dimension and the glide slope dimension, a logarithmic nonlinear correction model with a smoothing correction coefficient is used to calculate the adjustment components in the heading width dimension and the glide slope dimension.

6. The intelligent assisted debugging method for instrument landing system parameters according to claim 5, characterized in that, The simulation results include at least the predicted values ​​for each dimension; The method for generating candidate adjustment strategies based on the inference results includes: performing collision detection between the predicted values ​​of each dimension and the standard working geometric polyhedron; if a collision occurs, activating the deadlock prevention logic and generating a deadlock prevention suggested adjustment strategy as a candidate adjustment strategy; if no collision occurs, using the initial suggested adjustment vector as a candidate adjustment strategy.

7. The intelligent assisted debugging method for instrument landing system parameters according to claim 6, characterized in that, The deadlock prevention logic includes: performing a safety clamping process on the initial suggested adjustment vector to generate a safety compromise adjustment vector, calculating a threshold linkage adjustment vector based on the safety compromise adjustment vector, and combining the safety compromise adjustment vector and the threshold linkage adjustment vector to generate a deadlock prevention suggested adjustment strategy.

8. The intelligent assisted debugging method for instrument landing system parameters according to claim 7, characterized in that, The method for steady-state determination and filtering includes: Extract the adjustment amounts of parameters for each dimension from the candidate adjustment strategies, calculate the expected state value after adjustment for each dimension, and if the expected state value of a certain dimension falls within the hysteresis quiescent zone, then the adjustment amount of the parameter for that dimension is forcibly set to zero; if it does not fall within the hysteresis quiescent zone, then the adjustment amount of the parameter for that dimension is retained, and valid adjustment data is generated.

9. The intelligent assisted debugging method for instrument landing system parameters according to claim 8, characterized in that, When the candidate adjustment strategy is the initial suggested adjustment vector, the adjustment amounts of each dimension parameter in the candidate adjustment strategy include the heading straight dimension adjustment component, the heading width dimension adjustment component, the glide slope dimension adjustment component, and the glide slope width dimension adjustment component; when the candidate adjustment strategy is the deadlock prevention suggested adjustment strategy, the adjustment amounts of each dimension parameter include the deadlock prevention heading straight dimension adjustment component, the deadlock prevention heading width dimension adjustment component, the deadlock prevention glide slope dimension adjustment component, the deadlock prevention glide slope width dimension adjustment component, and the threshold linkage adjustment vector.

10. A smart auxiliary debugging system for instrument landing system parameters, used to implement the smart auxiliary debugging method for instrument landing system parameters according to any one of claims 1-9, characterized in that, The system includes: State space positioning module: used to construct a multidimensional coupled virtual state space, define a standard working geometric polyhedron within the multidimensional coupled virtual state space; acquire the original data sequence of the instrument landing system flight verification, and locate the current equipment state vector of the current ground transmitter in the multidimensional coupled virtual state space based on the original data sequence of the instrument landing system flight verification. Strategy generation module: used to determine the target center coordinates of the standard working geometric polyhedron, calculate the vector difference between the current device state vector and the target center coordinates to obtain the multidimensional displacement deviation, generate an initial proposed adjustment vector based on the multidimensional displacement deviation and the pre-built hybrid model, substitute the initial proposed adjustment vector into the multidimensional coupled virtual state space to perform simulation deduction, and generate candidate adjustment strategies based on the deduction results; Closed-loop update module: This module performs steady-state determination and filtering of candidate adjustment strategies through a preset hysteresis quiescent zone, and encapsulates them into an atomic execution instruction set; it executes the atomic execution instruction set and obtains the current device state vector for the next round, and performs closed-loop verification and hybrid model adaptive update based on the new round of current device state vector.

Images