A method, apparatus and storage medium for evaluating landing trajectories based on avionics data

By constructing a grounding point identification model and an ideal landing reference curve based on avionics data, and combining area deviation and angle deviation, the problem of strong subjectivity in the evaluation standards of existing technologies has been solved, and an objective, comprehensive, and accurate quantitative evaluation of aviation flight training and flight quality monitoring has been achieved.

CN123115580APending Publication Date: 2026-11-17CIVIL AVIATION FLIGHT UNIV OF CHINA +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611018028.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-09
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In the existing technology, the landing trajectory evaluation method in the field of aviation flight training and flight quality monitoring relies on a fixed theoretical standard trajectory, which leads to the evaluation standard being highly subjective and lacking universality. It cannot fully reflect the actual landing trajectory deviation under different scenarios, and the evaluation results are not accurate enough and lack persuasiveness.

Method used

By extracting avionics data, a landing point identification model is constructed, a trajectory coordinate function is generated, a subset of normal trajectory data is selected based on a confidence interval model, an optimal approximation algorithm is used to fit the ideal landing reference curve, deviation index is calculated by combining area deviation and angle deviation, and an evaluation model is established based on historical data to output the landing operation quality evaluation results.

Benefits of technology

It enables objective, comprehensive, and precise quantitative evaluation of landing operation quality, abandons fixed theoretical standards, improves the objectivity and universality of evaluation, and can more accurately reflect reasonable trajectory deviations under different scenarios, providing more convincing evaluation results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123115580A_ABST
    Figure CN123115580A_ABST
Patent Text Reader

Abstract

This application discloses a landing trajectory evaluation method, device, and storage medium based on avionics data, used to achieve an objective, comprehensive, and accurate quantitative evaluation of landing operation quality. The method includes: extracting and preprocessing the data sequence during the landing phase to obtain a trajectory coordinate function; constructing a confidence interval model based on the trajectory coordinate function and selecting a subset of trajectory data representing a conventional landing from the confidence interval model; averaging the trajectory data subset to obtain average trajectory data points; using an optimal approximation algorithm to perform polynomial curve fitting on the average trajectory data points to generate an ideal landing reference curve; calculating a deviation index based on the coordinate function of a single actual landing trajectory and the ideal landing reference curve; establishing an evaluation model based on the statistical distribution of the deviation index corresponding to historical landing trajectory data; inputting the deviation index of a single actual landing trajectory into the evaluation model and outputting the landing operation quality evaluation result.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of evaluation technology, and in particular to a landing trajectory evaluation method, device and storage medium based on avionics data. Background Technology

[0002] In the field of aviation flight training and flight quality monitoring, landing trajectory is an important basis for evaluating a pilot's landing operation skills. To ensure flight safety, it is necessary to quantify the deviation between the actual landing trajectory and the standard trajectory, thereby achieving an objective assessment of the landing operation quality.

[0003] In existing technologies, a fixed theoretical standard trajectory is typically pre-defined, and the actual landing trajectory is compared with this standard trajectory. The landing trajectory is evaluated by calculating the positional deviation between the two. During the evaluation process, the distance deviation between the trajectories is generally selected directly as the evaluation index, and the landing operation is judged as qualified based on a manually set fixed threshold.

[0004] However, existing technical solutions do not rely on the statistical distribution patterns of real historical landing data to establish evaluation models. Evaluation thresholds are set based on human experience, resulting in highly subjective evaluation standards with poor universality, making it difficult to objectively reflect the real and reasonable level of landing trajectory deviation under different scenarios. At the same time, single position deviation indicators cannot comprehensively characterize the overall deviation of the trajectory, making the final landing trajectory evaluation results inaccurate and unconvincing. Summary of the Invention

[0005] This application discloses a landing trajectory evaluation method, device, and storage medium based on avionics data, which is used to achieve an objective, comprehensive, and accurate quantitative evaluation of landing operation quality.

[0006] The first aspect of this application discloses a landing trajectory evaluation method based on avionics data, including:

[0007] Extract the data sequence of the landing phase and preprocess the data sequence to construct a grounding point identification model and obtain the trajectory coordinate function, wherein the x-axis of the trajectory coordinate function is the horizontal distance and the y-axis is the vertical height; A confidence interval model is constructed based on the statistical distribution of altitude data at preset feature locations of the trajectory coordinate function, and a subset of trajectory data representing normal landing is selected from the confidence interval model. The average trajectory data points are obtained by averaging the subset of trajectory data. The average trajectory data points are fitted with a polynomial curve using the optimal approximation algorithm to generate an ideal landing reference curve. The deviation index of the landing is calculated based on the coordinate function of the actual landing trajectory and the ideal landing reference curve. The deviation index includes area deviation and angle deviation. An evaluation model is established based on the statistical distribution of deviation indices corresponding to a sufficient number of similar landing trajectory data. The deviation index of the single actual landing trajectory is input into the evaluation model, and the landing operation quality evaluation result is output.

[0008] Optionally, the step of constructing a confidence interval model based on the statistical distribution of height data at preset feature locations using the trajectory coordinate function includes: The height values ​​of all trajectories at the starting feature position before the grounding point are selected to form a sample set; The sample set conforms to or approximately conforms to a normal distribution using a QQ graph. Calculate the mean and standard deviation of the sample set; A confidence interval model is constructed based on the mean and the standard deviation.

[0009] Optionally, the step of using an optimal approximation algorithm to perform polynomial curve fitting on the average trajectory data points to generate an ideal landing reference curve includes: The least squares-based optimal approximation algorithm is used to perform polynomial curve fitting on the average trajectory data points to generate an ideal landing reference curve.

[0010] Optionally, the least squares-based optimal approximation algorithm performs polynomial curve fitting on the average trajectory data points to generate an ideal landing reference curve, including: Preset the fitting error convergence threshold; The average trajectory data points are fitted with polynomials sequentially from low to high order, and the sum of squared residuals for each fit is calculated. When the ratio of the sum of squared residuals of the subsequent fit to the previous fit is less than the convergence threshold, the order of the fit is stopped and the optimal fit order is determined. Based on the optimal fitting order, a polynomial curve is fitted to generate the ideal landing reference curve.

[0011] Optionally, the deviation index is calculated based on the coordinate function of a single actual landing trajectory and the ideal landing reference curve. The deviation index includes area deviation and angle deviation, including: Determine the corresponding data points of the single actual landing trajectory coordinate function and the ideal landing reference curve within the analysis interval; Based on the corresponding data points, the area deviation and angle deviation are calculated respectively. The area deviation is used to characterize the spatial position difference of the trajectory, and the angle deviation is used to characterize the flight attitude difference of the trajectory.

[0012] Optionally, after calculating the deviation index based on the coordinate function of a single actual landing trajectory and the ideal landing reference curve, wherein the deviation index includes area deviation and angle deviation, and before establishing an evaluation model based on the statistical distribution of the deviation index corresponding to historical landing trajectory data, the following further steps are included: The overall deviation is calculated based on the area deviation and angle deviation.

[0013] Optionally, the evaluation model established based on the statistical distribution of the deviation index corresponding to historical landing trajectory data includes: The maximum likelihood estimation method is used to fit the comprehensive deviation to a Gamma distribution to obtain a Gamma probability distribution model; The evaluation threshold range for the comprehensive deviation is set based on the quantiles of the Gamma probability distribution model. The evaluation model is constructed based on the Gamma probability distribution model and the evaluation threshold interval.

[0014] A second aspect of this application provides a landing trajectory evaluation device based on avionics data, comprising: An extraction unit is used to extract the data sequence during the landing phase and preprocess the data sequence to construct a grounding point identification model and obtain a trajectory coordinate function, wherein the horizontal coordinate of the trajectory coordinate function is the horizontal distance and the vertical coordinate is the vertical height. The construction unit is used to construct a confidence interval model based on the statistical distribution of altitude data at preset feature positions of the trajectory coordinate function, and to select a subset of trajectory data representing conventional landing from the confidence interval model; The processing unit is used to perform averaging processing on the subset of trajectory data to obtain average trajectory data points; The fitting unit is used to perform polynomial curve fitting on the average trajectory data points using the optimal approximation algorithm to generate an ideal landing reference curve. The calculation unit is used to calculate the deviation index based on the coordinate function of a single actual landing trajectory and the ideal landing reference curve. The deviation index includes area deviation and angle deviation. Establish a unit to build an evaluation model based on the statistical distribution of deviation indicators corresponding to historical landing trajectory data; The output unit is used to input the deviation index of the single actual landing trajectory into the evaluation model and output the landing operation quality evaluation result.

[0015] A third aspect of this application provides a landing trajectory evaluation device based on avionics data, comprising: Processor, memory, input / output units, and bus; The processor is connected to memory, input / output units, and a bus; The memory holds a program, which the processor calls to execute, as in the first aspect and any optional method of the first aspect.

[0016] The fourth aspect of this application provides a computer-readable storage medium on which a program is stored, which, when executed on a computer, performs the methods of the first aspect and any optional methods of the first aspect.

[0017] As can be seen from the above technical solutions, the embodiments of this application have the following advantages: 1. This application obtains the trajectory coordinate function by preprocessing the raw avionics data, and constructs a confidence interval model based on the statistical distribution of altitude data at preset feature positions. It then selects a subset of trajectory data that represents conventional landing, and performs averaging and polynomial fitting on this subset to generate an ideal landing reference curve based on real conventional landing data. This approach abandons the fixed theoretical standard trajectory setting method in the prior art, making the reference curve more in line with the actual landing scenario and more representative and applicable.

[0018] 2. Using both area deviation and angle deviation as deviation evaluation indicators can not only characterize the positional difference between the actual trajectory and the ideal reference curve, but also reflect the trajectory attitude difference. Compared with the single positional deviation indicator of the existing technology, the evaluation dimensions are more comprehensive and can more completely and accurately characterize the overall deviation of the landing trajectory.

[0019] 3. An evaluation model is established based on the statistical distribution of deviation indicators from historical landing trajectory data. The evaluation threshold is objectively determined by the statistical distribution law of the data, without relying on human experience to set it. This reduces the subjectivity of the evaluation standard, improves the universality and objectivity of the evaluation model, and can truly reflect the reasonable level of landing trajectory deviation under different scenarios. Ultimately, it achieves a more accurate and convincing quantitative evaluation of the landing operation quality. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application, 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A schematic flowchart of an embodiment of a landing trajectory evaluation method based on avionics data provided in this application; Figure 2 A schematic diagram of an embodiment of the process for constructing a confidence interval model provided in this application; Figure 3A schematic diagram of an embodiment for generating an ideal landing reference curve provided in this application; Figure 4 A schematic diagram of an embodiment of the calculation of the deviation index provided in this application; Figure 5 A schematic diagram of an embodiment of the evaluation model provided in this application; Figure 6 A structural diagram of an embodiment of a landing trajectory evaluation device based on avionics data provided in this application; Figure 7 Another embodiment of the landing trajectory evaluation device based on avionics data provided in this application is shown in the structural diagram. Figure 8 Comparative analysis charts of multiple batches of actual landing trajectories and ideal baseline trends provided for this application; Figure 9 The QQ diagram showing the height data at 193m provided for this application; Figure 10 The probability density distribution map of the height data at 193m provided in this application; Figure 11 A comparative analysis diagram of the aircraft landing trajectory and the ideal landing reference curve provided for this application. Detailed Implementation

[0022] It should be noted that the landing trajectory evaluation method based on avionics data provided in this application can be applied to servers, devices, terminals, or other devices with logic processing capabilities; therefore, this application does not impose any limitations on it. For ease of description, the following description takes the system that records the landing trajectory as the executing entity.

[0023] Please see Figure 1 This application provides an embodiment of a landing trajectory evaluation method based on avionics data, comprising: 101. Extract the data sequence of the landing phase and preprocess the data sequence to build a grounding point identification model and obtain the trajectory coordinate function. The x-axis of the trajectory coordinate function is the horizontal distance and the y-axis is the vertical height. This step collects the raw flight data sequence during the landing phase from the system and preprocesses the raw data sequence, including data filtering, outlier removal, data interpolation, and coordinate normalization.

[0024] A touchdown point identification model is constructed based on preprocessed avionics data. The model employs a multi-feature fusion identification algorithm based on vertical velocity and altitude change rate. In practice, vertical velocity thresholds and altitude change rate thresholds are set, and the landing phase data sequence is traversed. When consecutive data points simultaneously meet the above two threshold features, the moment is determined to be the main wheel touchdown moment, thus completing the accurate identification of the touchdown point.

[0025] Using the identified grounding point as the origin, and the horizontal forward direction as the abscissa, and the vertical upward direction as the ordinate, the preprocessed landing phase data points are mapped onto this two-dimensional coordinate system. Spline interpolation is performed on the discrete data points to generate a continuous and smooth functional relationship, ultimately obtaining the trajectory coordinate function. The trajectory coordinate function, with horizontal distance as the abscissa and vertical height as the ordinate, is used to characterize the flight trajectory changes during the aircraft's landing phase.

[0026] The algorithm, based on decision trees and grounding logic, automatically segments the complete landing phase from the 50-foot clearance to complete stop from continuous flight data. The altitude (feet) and ground speed (knots) of each recorded point are converted into planar coordinates in meters using the following formula, forming the landing trajectory function f(x) represented by a polyline segment: ; ; in, Let be the horizontal coordinate of the i-th record point. Let be the ground velocity at the i-th recording point. Let be the vertical coordinate of the i-th record point. Let be the indicated height of the i-th recording point. This indicates the height of the grounding point.

[0027] 102. Construct a confidence interval model based on the statistical distribution of altitude data at preset feature locations using trajectory coordinate functions, and select a subset of trajectory data representing normal landing from the confidence interval model; Based on the altitude data of each trajectory coordinate function at preset feature locations, statistical analysis is performed, and a confidence interval model is constructed. The original landing trajectories are then filtered using the confidence interval model, retaining those with altitude data within the confidence interval and removing outlier trajectories, thereby obtaining a subset of trajectory data that can represent a normal landing state.

[0028] The core of the "noise removal screening" of the original landing trajectory lies in distinguishing between "normal landings" and "abnormal landings." In practice, altitude data of the trajectory coordinate function at preset characteristic positions are selected as analysis samples. These preset characteristic positions represent key flight phases during the landing process and can be considered the starting position before the touchdown point. This position is the critical node where the landing trajectory transitions from gliding to drifting, and its altitude value is crucial for characterizing the stability of the landing operation.

[0029] Collect the altitude values ​​of all landing trajectories to be analyzed at this characteristic location to form an altitude sample set. Perform statistical distribution analysis on the sample set, specifically calculating the mean, standard deviation, and other statistics of the sample set, and plot a QQ plot to verify the distribution characteristics of the sample set.

[0030] A confidence interval model is constructed based on statistical measures (mean and standard deviation), such as setting a 95% or 99% confidence level to determine a reasonable altitude range. The altitude values ​​of all original landing trajectories at this characteristic location are then substituted into this confidence interval model, and trajectories falling within the confidence interval are selected. These trajectories represent normal, conventional landing operations and constitute a subset of the trajectory data for normal landings. Trajectories falling outside the confidence interval are considered abnormal landings (such as hard landings, excessive flattening, or insufficient flattening).

[0031] 103. Average the subset of trajectory data to obtain the average trajectory data points; In order to eliminate random fluctuations in individual trajectories and reflect the overall statistical regularity of the landing trajectories, the subset of trajectory data obtained by screening in step 102 is averaged in this step.

[0032] In practice, using horizontal distance as a common reference axis, the trajectory coordinate functions of all conventional landing trajectories are aligned to the same horizontal distance coordinate system. For each discrete point with the same horizontal distance, the arithmetic mean of the vertical height values ​​of all conventional trajectories at that point is calculated to obtain a unique average vertical height value.

[0033] By repeating the above averaging calculation operation on all discrete points within the entire horizontal distance interval, a set of average trajectory data points is finally obtained, with horizontal distance as the x-axis and average vertical height as the y-axis. This average trajectory data point eliminates the interference of individual abnormal trajectories and can objectively and stably represent the overall morphological trend of a large number of conventional landing trajectories.

[0034] 104. Employ the optimal approximation algorithm to perform polynomial curve fitting on the average trajectory data points to generate an ideal landing reference curve; This step employs a least-squares-based optimal approximation algorithm to fit a polynomial curve to the average trajectory data points. The core principle of the least-squares method is to minimize the sum of squared errors between the fitted curve and the average data points to obtain the optimal polynomial coefficients, ensuring the highest possible fit between the fitted curve and the average trajectory.

[0035] In order to achieve a balance between fitting accuracy and curve smoothness during the fitting process, an adaptive order determination strategy is adopted to perform polynomial fitting and determine the optimal fitting order.

[0036] Based on the determined optimal fitting order, the polynomial coefficients are solved using the least squares method, ultimately generating an ideal landing reference curve that is continuous, smooth, and accurately represents the overall trend of a conventional landing trajectory. This ideal landing reference curve is constructed based on real flight data and has extremely high practical representativeness.

[0037] 105. Calculate the deviation index of the landing based on the coordinate function of the actual landing trajectory and the ideal landing reference curve. The deviation index includes area deviation and angle deviation. After obtaining the trajectory coordinate function of a single actual landing trajectory and the ideal landing reference curve, the corresponding data points of the two within the same analysis interval are first determined. The analysis interval is the complete landing trajectory interval from the flattening start position to the touchdown point.

[0038] Based on the corresponding data points, calculate the deviation indices for both dimensions: Area deviation: The area of ​​the closed figure enclosed by the actual trajectory coordinate function and the ideal reference curve within the entire analysis interval is calculated. This area is then compared with the reference area under the ideal curve to obtain the area deviation index. This area deviation, from the perspective of spatial positional deviation, comprehensively characterizes the degree of spatial offset of the actual trajectory relative to the ideal trajectory.

[0039] Angle deviation: The slopes of the tangents to the actual trajectory and the ideal reference curve at preset key feature points (such as the leveling start point and the landing point) are extracted separately. The slopes are converted into the corresponding glide angles, and the absolute difference between the two angles is calculated. This indicator characterizes the standardization of the glide angle during actual landing from the perspective of flight attitude deviation. Furthermore, due to GPS data drift, it is difficult to effectively determine the landing position based solely on the latitude and longitude coordinates of the landing point. The application of angle deviation can indirectly characterize the standardization of the landing distance. In actual flight, factors such as GPS signal drift, airport reference coordinate errors, and sensor positioning accuracy make it difficult to accurately determine the landing position based solely on the latitude and longitude coordinates of the landing point. Angle deviation, by quantifying the deviation of the glide attitude, can indirectly reflect the rationality of the pilot's leveling and glide control: if the angle deviation is large, it indicates a significant deviation between the actual glide angle and the ideal reference, which can easily lead to the landing point being ahead or behind, thus indirectly characterizing the standardization of the landing distance and compensating for the limitations of pure position coordinate determination due to positioning errors.

[0040] By jointly calculating the area deviation and the angle deviation, the deviation between the actual landing trajectory and the ideal trajectory of a single landing is comprehensively and three-dimensionally quantified from two different dimensions: spatial position and flight attitude, resulting in a more comprehensive and accurate evaluation.

[0041] 106. Establish an evaluation model based on the statistical distribution of deviation indices corresponding to a sufficient number of similar landing trajectory data; This step involves collecting a large amount of historical normal landing trajectory data. For each historical trajectory, steps 101 to 105 are repeated to calculate the corresponding historical comprehensive deviation index (which can be further integrated from area deviation and angle deviation). The deviation indices of all historical normal landing trajectories are used as a statistical sample set.

[0042] By performing statistical distribution analysis on the sample set, the Gamma distribution fitting method can be used to obtain a Gamma probability distribution model that best fits the distribution pattern of historical deviation data.

[0043] Based on the quantile characteristics of the Gamma probability distribution model, evaluation threshold ranges for the deviation index are set. For example, according to the 13.4%, 21.8%, 25.8%, and 33.3% quantiles of the Gamma distribution, the deviation values ​​are divided into different numerical ranges, each corresponding to a preset landing operation quality level. Ultimately, the Gamma probability distribution model and its corresponding evaluation threshold ranges together constitute the landing operation quality evaluation model.

[0044] 107. Input the deviation index of the actual landing trajectory of a single landing into the evaluation model, and output the landing operation quality evaluation results.

[0045] Using the deviation index (or comprehensive deviation) of the single actual landing trajectory calculated in step 105 as input data, substitute it into the established evaluation model. The evaluation model matches the input deviation value to a pre-set evaluation threshold interval, and determines the grade interval to which the value belongs.

[0046] According to the matching result, the evaluation model directly outputs the corresponding landing operation quality evaluation result. The landing operation quality evaluation result includes five grades, and the division of the grades is based on the pre-set threshold intervals of the Gamma probability distribution quantiles of the comprehensive deviation index D calculated from historical data, specifically: When D ≤ 13.4%, the evaluation grade is "excellent"; When 13.4% < D ≤ 21.8%, the evaluation grade is "good"; When 21.8% < D ≤ 25.8%, the evaluation grade is "medium"; When 25.8% < D ≤ 33.3%, the evaluation grade is "qualified"; When D > 33.3%, the evaluation grade is "unqualified".

[0047] As Figure 11 shown, Figure 11 it intuitively presents a comparative analysis diagram of the ideal landing reference curve and the aircraft landing trajectory, which is used to visually demonstrate the deviation degree between the actual landing trajectory and the ideal landing trajectory.

[0048] As shown in Table 1, multiple sets of area deviation, angle deviation, and the comprehensive deviation index obtained through the area deviation and the angle deviation are provided. The comprehensive deviation index is input into the evaluation model, and the evaluation model outputs the landing operation quality evaluation result.

[0049]

[0050] Table 1 As Figure 8 shown, Figure 8 it is a comparative analysis diagram of multiple batches of actual landing trajectories and the ideal reference trend, which is used to evaluate the trajectory consistency and deviation degree of multiple landing processes.

[0051] Multiple colored curves represent several different actual landing trajectories. The overall trend of the multiple colored curves is consistent, all gradually descending from high altitude to the ground contact point (0,0), reflecting the typical descent characteristics of the landing process. The red curve is the highest and gentlest descent among all landing trajectories, representing a more conservative landing trajectory in this comparison. The difference between the red curve and other curves intuitively reflects the maximum deviation between the trajectories. The light blue curve is the lowest and steepest descent among all trajectories, which can be regarded as a more aggressive landing trajectory. Its altitude is significantly lower than other curves in the early stage, and it gradually converges in the later stage.

[0052] All curves converge to 0 altitude as they approach the touchdown point (horizontal distance approaches 0), indicating that all landing processes eventually achieve a smooth landing, meeting the basic requirements for a safe landing. During the critical descent phase from -400m to -100m horizontal distance, there are significant altitude differences between the curves, especially the red and light blue curves, which reflect trajectory fluctuations between different landings.

[0053] In this embodiment, the trajectory coordinate function is obtained by preprocessing the raw avionics data, and a confidence interval model is constructed based on the statistical distribution of altitude data at preset feature positions. A subset of trajectory data representing conventional landing is selected, and then this subset is averaged and fitted with a polynomial to generate an ideal landing reference curve based on real conventional landing data. This method abandons the fixed theoretical standard trajectory setting method in the prior art, making the reference curve more in line with the actual landing scenario and more representative and applicable. The area deviation and angle deviation are used together as deviation evaluation indicators, which can not only represent the positional difference between the actual trajectory and the ideal reference curve, but also reflect the trajectory attitude difference. Compared with the single position deviation indicator in the prior art, the evaluation dimension is more comprehensive and can more completely and accurately represent the overall deviation of the landing trajectory. An evaluation model is established based on the statistical distribution of deviation indicators of historical landing trajectory data. The evaluation threshold is objectively determined by the statistical distribution law of the data, without relying on human experience to set it. This reduces the subjectivity of the evaluation standard and improves the universality and objectivity of the evaluation model. It can truly reflect the reasonable landing trajectory deviation level under different scenarios, and finally achieve a more accurate and convincing quantitative evaluation of the landing operation quality.

[0054] Please see Figure 2 , Figure 2 An embodiment of the confidence interval model provided in this application includes: 201. Select the height values ​​of all trajectories at the starting feature position before the grounding point to form a sample set; First, the starting point of the leveling phase before the touchdown point was defined as 193 meters before the touchdown point, corresponding to an altitude of 6 meters. This is a typical position for the start of leveling during landing and a critical node for the aircraft to transition from the glide phase to the leveling phase during landing. Combined with standard aviation flight operating procedures, the altitude data at this position can effectively reflect the routine nature of the landing operation. All landing trajectory data to be analyzed were collected, and the altitude value of each trajectory at 193 meters before the touchdown point (corresponding to an altitude of 6 meters, i.e., the typical starting point of leveling) was extracted. All extracted altitude values ​​were integrated to form a sample set for subsequent statistical analysis.

[0055] 202. Verify that the sample set conforms to or approximately conforms to a normal distribution using a QQ graph; like Figure 9 As shown, the altitude value sample set obtained in step 201 is sorted in ascending order, resulting in ordered sample data. Using professional data processing methods, the standard normal quantile corresponding to each sample point after sorting is calculated. The standard normal quantile is determined based on the standard normal distribution, conforming to the conventional requirements of statistical analysis in the aviation field. A QQ plot is drawn with the quantiles of the sample set (i.e., the sorted altitude values) as the ordinate and the corresponding standard normal quantiles as the abscissa, ensuring consistent coordinate scales and clear data point markings during the plotting process. By observing the distribution trend of all data points in the QQ plot, it is determined whether the sample set conforms to or approximately conforms to a normal distribution. If the data points basically fit the fitted line without significant deviation, the verification is successful.

[0056] 203. Calculate the mean and standard deviation of the sample set; All altitude values ​​in the qualified altitude value sample set were uniformly calculated to obtain two core statistical parameters: the mean and the standard deviation. The mean, calculated by averaging all altitude values ​​in the sample set, characterizes the average altitude level of the sample set, corresponding to the baseline altitude level of a conventional landing trajectory 193 meters before the touchdown point (corresponding to an altitude of 6 meters, i.e., the typical starting position for leveling). The standard deviation, calculated by the deviation of all altitude values ​​from the mean, characterizes the dispersion of altitude values ​​in the sample set, reflecting the range of altitude fluctuations at this characteristic position during a conventional landing.

[0057] 204. Construct a confidence interval model based on the mean and standard deviation.

[0058] Based on the conventional requirements and practical application scenarios of landing trajectory analysis in the aviation field, an appropriate confidence level is determined. This confidence level balances the rigor and rationality of trajectory selection and conforms to the statistical regularity of conventional landing trajectories. According to the selected confidence level, the standard normal distribution quantile table is consulted to determine the corresponding standard normal distribution quantiles. Combining the two core parameters of sample mean and standard deviation, the corresponding statistical construction method is used to construct the confidence interval. This confidence interval represents the reasonable range of altitude values ​​at 193 meters before the touchdown point (corresponding to a height of 6 meters, i.e., the typical starting position after flattening). Figure 10 As shown, a confidence interval diagram is drawn, clearly marking the upper and lower limits and the mean position of the confidence interval, intuitively presenting the reasonable fluctuation range of the height value of this feature position. This confidence interval directly constitutes a confidence interval model for screening conventional landing trajectories.

[0059] In this embodiment, key nodes that fit the actual flight scenario are used as the screening criteria. Confidence intervals are constructed based on scientific normal distribution verification and objective statistical parameters. This avoids the subjectivity of manually setting thresholds and ensures the reliability and stability of the screening criteria. It can accurately eliminate abnormal trajectories such as hard landings and excessive stretching. At the same time, the whole process is clear and reproducible, and has good engineering practicality and scenario adaptability.

[0060] Please see Figure 3 , Figure 3 One embodiment of generating an ideal landing reference curve provided in this application includes: 301. Preset the fitting error convergence threshold; The fitting error convergence threshold is the core criterion for judging whether the fitting accuracy meets the requirements and whether to stop increasing the fitting order. Its setting needs to be combined with the actual application scenario and preset according to the accuracy requirements of the landing equipment and the dispersion of the trajectory data. The value range can be 0.01~0.1, and the preferred value is 0.05.

[0061] The convergence threshold setting can be flexibly adapted to different landing scenarios. When the landing accuracy requirement is high and the trajectory data dispersion is small, a smaller convergence threshold can be selected to further improve the fitting accuracy. When the landing accuracy requirement is moderate and the trajectory data dispersion is large, a larger convergence threshold can be selected to improve the fitting efficiency while ensuring that the fitting accuracy meets the requirements, thus ensuring a balance between fitting accuracy and computational efficiency.

[0062] 302. Perform polynomial fitting on the average trajectory data points from low to high order, and calculate the sum of squared residuals for each fitting. First, the basic data for fitting is obtained—average trajectory data points. The method is to average the height y corresponding to each integer x-coordinate, and finally obtain 401 average trajectory points. This process can eliminate random errors in single data collection.

[0063] The fitting process adopts a stepwise increasing method from low order to high order. The initial order of the low order is set to 3, and the increment step of the high order is 1. The 3rd, 4th and 5th order polynomials are fitted in sequence. The least squares method is used in the fitting process until the preset fitting termination condition is met.

[0064] After each fitting operation, the sum of squared residuals is calculated. The method for calculating the sum of squared residuals is as follows: for each polynomial curve obtained from the fitting operation... The difference between the actual height value and the fitted height value of the 401 average trajectory data points is calculated, and the squares of all differences are summed to obtain the residual sum of squares for this fitting. The residual sum of squares can intuitively reflect the degree of deviation between the fitted curve and the actual trajectory data points. The smaller the residual sum of squares, the higher the fit between the fitted curve and the actual trajectory, and the higher the fitting accuracy.

[0065] 303. When the ratio of the sum of squared residuals of the subsequent fit to the previous fit is less than the convergence threshold, stop increasing the order and determine the optimal fitting order. During the fitting process, a preset fitting error convergence threshold (e.g., 0.05) is used as the core criterion for determining the optimal fitting order.

[0066] The optimal fitting order is determined using a polynomial fitting formula: ; Where x is the independent variable, representing the horizontal coordinate of the landing trajectory; The coefficients of the polynomial are obtained by least squares fitting and will be summarized into the specific values ​​in Table 1. m is the highest order of the polynomial (i.e. the fitting degree, for example, m=4). n is the total number of average trajectory points involved in the fitting (e.g., n=401). N is the set of natural numbers, which means that m must be a positive integer.

[0067] Compare the ratio of the residual sum of squares obtained in each calculation with the preset convergence threshold. If the ratio of the residual sum of squares is less than the convergence threshold, it means that continuing to increase the fitting order will not significantly improve the fitting effect, and the fitting has reached a stable state. At this point, stop increasing the order and determine the optimal number of fitting iterations, such as 4.

[0068] 304. Based on the optimal fitting order, complete the polynomial curve fitting to generate the ideal landing reference curve.

[0069] Based on the optimal fitting order obtained in step 303, the ideal landing curves with coefficients shown in Table 2 are finally obtained. Optionally, the glide slope of the ideal landing curve above 6 meters is approximately 3.016°, which is highly consistent with the theoretical value.

[0070]

[0071] Table 2 In this embodiment, by gradually increasing the fitting order and comparing the ratio of the sum of squared residuals of previous and subsequent fittings with the convergence threshold, adaptive determination of the fitting order is achieved. This eliminates the need for manual subjective setting of the order, reducing operational difficulty and improving fitting efficiency. Using the ratio of the sum of squared residuals as the criterion ensures controllable fitting accuracy, and the convergence threshold can be flexibly adjusted according to data accuracy requirements, adapting to the fitting needs of different scenarios. Ultimately, the optimal fitting order is determined, avoiding both the excessive fitting deviation and inaccurate data reflection caused by low-order fitting, and the drawbacks of high-order fitting such as model complexity, overfitting, and decreased generalization ability.

[0072] Please see Figure 4 , Figure 4 One embodiment of the calculation deviation index provided in this application includes: 401. Determine the corresponding data points of the actual landing trajectory coordinate function and the ideal landing reference curve within the analysis interval; First, a predetermined analysis interval is defined. This interval represents the trajectory segment that needs to be evaluated during the landing process, covering the critical stages of landing. Then, based on the trajectory data collected during the actual landing, a coordinate function for a single actual landing trajectory is fitted. Simultaneously, the ideal landing reference curve obtained in step 304 is retrieved. Within the defined analysis interval, coordinate points at corresponding positions are extracted from both the actual trajectory coordinate function and the ideal landing reference curve at uniform intervals. These points are then identified as the corresponding data points for the single actual landing trajectory coordinate function and the ideal landing reference curve within the analysis interval, ensuring a one-to-one correspondence between the two sets of data points.

[0073] 402. Based on the corresponding data points, calculate the area deviation and angle deviation respectively. The area deviation is used to characterize the difference in the spatial position of the trajectory, and the angle deviation is used to characterize the difference in the flight attitude of the trajectory. Based on the corresponding data points determined in step 401, the area deviation and angle deviation are calculated respectively. The area deviation characterizes the spatial difference between the actual landing trajectory and the ideal landing reference curve. It reflects the degree of spatial offset of the actual trajectory relative to the ideal trajectory through the characteristics of the area enclosed by the corresponding data points; the greater the offset, the higher the area deviation value. The angle deviation characterizes the flight attitude difference between the actual landing trajectory and the ideal landing reference curve. It reflects the deviation of the actual flight attitude from the ideal attitude through the difference in the tangent direction of the trajectory at the corresponding data points; the greater the attitude deviation, the higher the angle deviation value.

[0074] The formula for calculating area deviation is as follows: ; in, For the ideal landing reference curve function, This is the trajectory coordinate function corresponding to the actual landing trajectory. To analyze the x-coordinate of the starting point, and = -400 meters, where 0 is the x-coordinate of the grounding point.

[0075] The formula for calculating the angular deviation is as follows: ; in, For the ideal landing reference curve function, This is the trajectory coordinate function corresponding to the actual landing trajectory. To analyze the x-coordinate of the starting point, and = -400 meters, where 0 is the x-coordinate of the grounding point. and The height and distance of the ideal landing curve.

[0076] 403. Calculate the overall deviation based on the area deviation and the angle deviation.

[0077] Based on the calculated area and angle deviations, a comprehensive deviation is calculated. Taking into account the landing safety assessment requirements and the impact of spatial location and flight attitude differences, the area and angle deviations are rationally integrated to obtain a comprehensive deviation that fully reflects the overall deviation of the actual landing trajectory from the ideal baseline curve. This comprehensive deviation directly reflects the compliance and safety of the actual landing trajectory, providing a reliable basis for optimizing and adjusting the landing trajectory.

[0078] In this embodiment, by extracting the corresponding data points of the actual and ideal trajectories, the basic reliability of the deviation calculation is ensured, avoiding evaluation errors caused by data deviations. The area and angle deviations are calculated separately, achieving a more detailed evaluation of the deviation dimensions, which is more targeted than a single evaluation method. The comprehensive deviation, integrating these two types of deviation indicators, can comprehensively reflect the overall trajectory deviation, intuitively demonstrating landing safety and compliance, providing a reliable basis for trajectory optimization and adjustment, improving the stability and safety of the landing process, and adapting to the evaluation needs of various landing scenarios.

[0079] Please see Figure 5 , Figure 5 An embodiment of the evaluation model provided in this application includes: 501. The maximum likelihood estimation method is used to fit the Gamma distribution to the comprehensive deviation, and the Gamma probability distribution model is obtained. In this step, the maximum likelihood estimation method is used to fit the Gamma distribution to the overall deviation, resulting in a Gamma probability distribution model. First, overall deviation data calculated during multiple landings are collected to ensure sufficient sample size and coverage of different landing scenarios, guaranteeing the reliability of the fitting results. Then, the maximum likelihood estimation method is used to fit and analyze the collected overall deviation sample data. This method identifies the parameters that best reflect the distribution characteristics of the sample data, determines the relevant parameters of the Gamma distribution, and thus constructs a Gamma probability distribution model that closely matches the distribution pattern of the overall deviation data.

[0080] 502. Based on the quantile of the Gamma probability distribution model, set the evaluation threshold range for the comprehensive deviation. After obtaining the Gamma probability distribution model, and considering the actual needs of landing safety assessment, quantile standards for classifying evaluation levels were determined. By analyzing the quantile characteristics of the Gamma probability distribution model, multiple continuous intervals were divided according to the probability requirements corresponding to different safety levels. Each interval corresponds to a different comprehensive deviation evaluation level. These intervals are the evaluation threshold intervals for comprehensive deviation, which can clearly distinguish the safety levels corresponding to different degrees of deviation.

[0081] 503. An evaluation model is constructed based on the Gamma probability distribution model and the evaluation threshold interval.

[0082] The Gamma probability distribution model obtained in step 501 is integrated with the evaluation threshold range set in step 502 to form a complete comprehensive deviation evaluation model. This evaluation model can use the Gamma probability distribution model to analyze the probability distribution position of the comprehensive deviation data to be evaluated, and then combine it with the evaluation threshold range to determine the evaluation level corresponding to the comprehensive deviation. This enables an accurate evaluation of the comprehensive deviation of a single landing trajectory, providing scientific and reliable support for landing safety judgment and trajectory optimization.

[0083] In this embodiment, the Gamma distribution is fitted to the overall deviation using the maximum likelihood estimation method, which can fully reflect the distribution pattern of the actual deviation data and improve the accuracy and applicability of the probability model. The evaluation threshold range is set based on the quantiles of the distribution model, which makes the threshold setting more scientific and objective and avoids the subjectivity and arbitrariness brought about by human experience judgment. The evaluation model is constructed by combining the distribution model and the threshold range, which can realize the quantitative classification and accurate judgment of the overall deviation. It can not only intuitively reflect the deviation risk level of the landing trajectory, but also provide a stable and reliable evaluation basis for subsequent safety assessment and trajectory optimization.

[0084] Please see Figure 6 This application provides an embodiment of a landing trajectory evaluation device based on avionics data, comprising: Extraction unit 601 is used to extract the data sequence during the landing phase and preprocess the data sequence to obtain the trajectory coordinate function, where the horizontal coordinate function is the horizontal distance and the vertical coordinate is the vertical height. The construction unit 602 is used to construct a confidence interval model based on the statistical distribution of altitude data at preset feature positions of the trajectory coordinate function, and to select a subset of trajectory data representing conventional landing from the confidence interval model; Optionally, building unit 602 is also used for: The height values ​​of all trajectories at the starting feature position before the grounding point are selected to form a sample set; Verify that the sample set conforms to or approximately conforms to a normal distribution using QQ graphs; Calculate the mean and standard deviation of the sample set; A confidence interval model is constructed based on the mean and standard deviation.

[0085] The processing unit 603 is used to perform averaging processing on a subset of trajectory data to obtain average trajectory data points; Fitting unit 604 is used to perform polynomial curve fitting on the average trajectory data points using the optimal approximation algorithm to generate an ideal landing reference curve. Optionally, the fitting unit 604 is also used for: The least squares-based optimal approximation algorithm performs polynomial curve fitting on the average trajectory data points to generate an ideal landing reference curve.

[0086] Optionally, the fitting unit 604 is also used for: Preset the fitting error convergence threshold; Polynomial fitting is performed on the average trajectory data points sequentially from low to high order, and the sum of squared residuals for each fitting is calculated. When the ratio of the sum of squared residuals of the subsequent fit to the previous fit is less than the convergence threshold, the order of the fit is stopped and the optimal fit order is determined. Polynomial curve fitting is performed based on the optimal fitting order to generate an ideal landing reference curve.

[0087] The first calculation unit 605 is used to calculate the deviation index based on the coordinate function of the actual landing trajectory and the ideal landing reference curve. The deviation index includes area deviation and angle deviation. Optionally, the first computing unit 605 is also used for: Determine the corresponding data points of the actual landing trajectory coordinate function and the ideal landing reference curve within the analysis interval; Based on the corresponding data points, the area deviation and angle deviation are calculated respectively. The area deviation is used to characterize the spatial position difference of the trajectory, and the angle deviation is used to characterize the flight attitude difference of the trajectory.

[0088] Optionally, a second computing unit 606 is also included, for: The overall deviation is calculated based on the area deviation and the angle deviation.

[0089] Unit 607 is established to build an evaluation model based on the statistical distribution of deviation index corresponding to historical landing trajectory data; Optionally, unit 607 is also used for: The maximum likelihood estimation method is used to fit the Gamma distribution to the comprehensive deviation, and a Gamma probability distribution model is obtained. The evaluation threshold range for comprehensive deviation is set based on the quantile of the Gamma probability distribution model. The evaluation model is constructed based on the Gamma probability distribution model and the evaluation threshold range.

[0090] Output unit 608 is used to input the deviation index of a single actual landing trajectory into the evaluation model and output the landing operation quality evaluation result.

[0091] For detailed implementation methods, please refer to... Figures 1 to 5 Examples are not detailed here.

[0092] Please see Figure 7 This application provides a landing trajectory evaluation device based on avionics data, comprising: Processor 701, memory 702, input / output unit 703, and bus 704.

[0093] The processor 701 is connected to the memory 702, the input / output unit 703, and the bus 704.

[0094] The memory 702 stores a program, and the processor 701 calls the program to execute it, such as... Figure 1 , Figure 2 , Figure 3 , Figure 4 as well as Figure 5 The method in the middle.

[0095] This application provides a computer-readable storage medium on which a program is stored, and when the program is executed on a computer, it performs the following... Figure 1 , Figure 2 , Figure 3 , Figure 4 as well as Figure 5 The method in the middle.

[0096] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0097] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0098] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0099] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0100] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A landing trajectory evaluation method based on avionics data, characterized in that, include: Extract the data sequence of the landing phase and preprocess the data sequence to construct a grounding point identification model and obtain the trajectory coordinate function, wherein the x-axis of the trajectory coordinate function is the horizontal distance and the y-axis is the vertical height; A confidence interval model is constructed based on the statistical distribution of altitude data at preset feature locations of the trajectory coordinate function, and a subset of trajectory data representing normal landing is selected from the confidence interval model. The average trajectory data points are obtained by averaging the subset of trajectory data. The average trajectory data points are fitted with a polynomial curve using the optimal approximation algorithm to generate an ideal landing reference curve. The deviation index of the landing is calculated based on the coordinate function of the actual landing trajectory and the ideal landing reference curve. The deviation index includes area deviation and angle deviation. An evaluation model is established based on the statistical distribution of deviation indices corresponding to a sufficient number of similar landing trajectory data. The deviation index of the single actual landing trajectory is input into the evaluation model, and the landing operation quality evaluation result is output.

2. The landing trajectory evaluation method according to claim 1, characterized by, The construction of a confidence interval model based on the statistical distribution of height data at preset feature locations using the trajectory coordinate function includes: The height values ​​of all trajectories at the starting feature position before the grounding point are selected to form a sample set; The sample set conforms to or approximately conforms to a normal distribution using a QQ graph. Calculate the mean and standard deviation of the sample set; A confidence interval model is constructed based on the mean and the standard deviation.

3. The landing trajectory evaluation method according to claim 1, characterized by, The step of using an optimal approximation algorithm to perform polynomial curve fitting on the average trajectory data points to generate an ideal landing reference curve includes: The least squares-based optimal approximation algorithm is used to perform polynomial curve fitting on the average trajectory data points to generate an ideal landing reference curve.

4. The landing trajectory evaluation method according to claim 3, characterized by, The least squares-based optimal approximation algorithm performs polynomial curve fitting on the average trajectory data points to generate an ideal landing reference curve, including: Preset the fitting error convergence threshold; The average trajectory data points are fitted with polynomials sequentially from low to high order, and the sum of squared residuals for each fit is calculated. When the ratio of the sum of squared residuals of the subsequent fit to the previous fit is less than the convergence threshold, the order of the fit is stopped and the optimal fit order is determined. Based on the optimal fitting order, a polynomial curve is fitted to generate the ideal landing reference curve.

5. The landing trajectory evaluation method according to claim 1, characterized by, The deviation index is calculated based on the coordinate function of the actual landing trajectory and the ideal landing reference curve. The deviation index includes area deviation and angle deviation, including: Determine the corresponding data points of the single actual landing trajectory coordinate function and the ideal landing reference curve within the analysis interval; Based on the corresponding data points, the area deviation and angle deviation are calculated respectively. The area deviation is used to characterize the spatial position difference of the trajectory, and the angle deviation is used to characterize the flight attitude difference of the trajectory.

6. The landing trajectory evaluation method according to claim 1, characterized by, After calculating the deviation index based on the coordinate function of a single actual landing trajectory and the ideal landing reference curve, wherein the deviation index includes area deviation and angle deviation, and before establishing an evaluation model based on the statistical distribution of the deviation index corresponding to historical landing trajectory data, the following steps are also included: The overall deviation is calculated based on the area deviation and angle deviation.

7. The landing trajectory evaluation method according to any one of claims 1 to 6, characterized by, The evaluation model is established based on the statistical distribution of the deviation index corresponding to historical landing trajectory data, including: The maximum likelihood estimation method is used to fit the comprehensive deviation to a Gamma distribution to obtain a Gamma probability distribution model; The evaluation threshold range for the comprehensive deviation is set based on the quantiles of the Gamma probability distribution model. The evaluation model is constructed based on the Gamma probability distribution model and the evaluation threshold interval.

8. An apparatus for evaluating a landing trajectory based on avionics data, characterized in that, The device includes: An extraction unit is used to extract the data sequence during the landing phase and preprocess the data sequence to construct a grounding point identification model and obtain a trajectory coordinate function, wherein the horizontal coordinate of the trajectory coordinate function is the horizontal distance and the vertical coordinate is the vertical height. The construction unit is used to construct a confidence interval model based on the statistical distribution of altitude data at preset feature positions of the trajectory coordinate function, and to select a subset of trajectory data representing a normal landing from the confidence interval model; The processing unit is used to perform averaging processing on the subset of trajectory data to obtain average trajectory data points; The fitting unit is used to perform polynomial curve fitting on the average trajectory data points using the optimal approximation algorithm to generate an ideal landing reference curve. The first calculation unit is used to calculate the deviation index of the landing based on the coordinate function of the actual landing trajectory and the ideal landing reference curve. The deviation index includes area deviation and angle deviation. Establish a unit to build an evaluation model based on the statistical distribution of deviation indices corresponding to a sufficient number of similar landing trajectory data; The output unit is used to input the deviation index of the single actual landing trajectory into the evaluation model and output the landing operation quality evaluation result.

9. An apparatus for evaluating a landing trajectory based on avionics data, characterized in that, The device includes: Processor, memory, input / output units, and bus; The processor is connected to the memory, the input / output unit, and the bus; The memory stores a program, which the processor calls to execute the landing trajectory evaluation method based on avionics data as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a program stored thereon, the program executing, when executed on a computer, the landing trajectory evaluation method based on avionics data as described in any one of claims 1 to 7.