Method and system for evaluating complexity of airport surface traffic situation

Abstract

The present application relates to the field of airport operation management and intelligent traffic technology, in particular to an airport surface traffic situation complexity enhancement evaluation method and system, comprising obtaining airport surface operation data, and constructing complexity evaluation key indicators for representing airport surface traffic situation according to the airport surface operation data; calculating the traffic complexity corresponding to the target time slot according to the complexity evaluation key indicators; obtaining the historical operation data of a plurality of historical period time slots corresponding to the target time slot, and constructing a multi-period input structure according to the operation data of the target time slot and the historical operation data of the plurality of historical period time slots; and calculating the predicted traffic complexity corresponding to the target time slot according to the prediction key indicators. The present application is used to realize the complexity evaluation indicators that comprehensively reflect the airport surface traffic operation state, and to enhance the evaluation of the traffic complexity of the target time slot in combination with the correlation between the target time slot and the historical operation scene.

Description

Technical Field

[0001] This invention relates to the fields of airport operation management and intelligent transportation technology, specifically to a method and system for assessing the increased complexity of airport surface traffic conditions. Background Technology

[0002] With the continuous expansion of air transport and the increasing density of airport operations, airport surface operations such as aircraft pushback, taxiing, holding, takeoff, landing, and gate support are intertwined, resulting in a highly dynamic and coupled airport surface traffic situation. Airport operations management departments typically need to analyze and make decisions regarding aircraft taxiing organization, runway resource allocation, gate scheduling, and surface operational status based on the complexity of airport surface traffic. Therefore, accurate assessment of the complexity of airport surface traffic has become an important aspect of intelligent airport operations management.

[0003] Currently, airport surface traffic complexity assessments typically rely on operational data such as taxiing time, number of aircraft, traffic flow density, and delays, combined with multiple operational indicators to determine the overall traffic complexity of the airport surface. While these methods can reflect the operational status of the airport surface to some extent, different operational indicators have different physical meanings and patterns of change. Using only a single operational indicator or simply combining multiple indicators cannot comprehensively reflect the correlation between aircraft operational status, resource demands, and operational efficiency during airport surface traffic operations, thus affecting the ability of traffic complexity assessment results to represent the overall operational situation of the airport.

[0004] On the other hand, most existing airport surface traffic complexity assessments analyze the current operational status at a target time or time slot, obtaining traffic complexity results for that specific moment. They lack comprehensive utilization of historical operational patterns and fail to fully consider the evolutionary characteristics of airport surface traffic conditions over time. When airports are affected by weather changes, flight fluctuations, or unforeseen events, traffic complexity obtained solely from current operational data is insufficient to reflect subsequent traffic trends, thus impacting airport surface traffic situation analysis and operational decisions. Summary of the Invention

[0005] This invention aims to solve the above-mentioned technical problems by providing a method and system for enhancing the assessment of the complexity of airport surface traffic situation. It realizes a complexity evaluation index that comprehensively reflects the traffic operation status of the airport surface, and enhances the assessment of traffic complexity of the target time slot by combining the correlation between the target time slot and historical operation scenarios.

[0006] The objective of this invention can be achieved through the following technical solutions: A method for assessing the increased complexity of airport surface traffic situations includes the following steps: S1. Obtain airport surface operation data and construct key indicators for evaluating the complexity of airport surface traffic based on the airport surface operation data. S2. Calculate the traffic complexity corresponding to the target time slot based on the key indicators of complexity evaluation. S3. For the target time slot, obtain the historical operation data of multiple historical period time slots corresponding to the target time slot, and construct a multi-period input structure based on the operation data of the target time slot and the historical operation data of multiple historical period time slots. S4. Based on the multi-period input structure, determine the similarity weights between multiple historical period time slots and the target time slot, and perform weighted processing on the key indicators corresponding to multiple historical period time slots according to the similarity weights to obtain the predicted key indicators of the target time slot; calculate the predicted traffic complexity corresponding to the target time slot based on the predicted key indicators.

[0007] Furthermore, key indicators for complexity evaluation include instantaneous surface traffic flow, cumulative surface traffic flow, aircraft queue length, time slot resource demand, and relative taxiing time. Instantaneous flow indicators include the number of aircraft taxiing at the start of the target time slot. The number of aircraft taxiing at the end of the target time slot. ; The cumulative traffic flow indicator includes the number of departing aircraft whose taxiing process overlaps with the target time slot. The number of arriving aircraft located on the airport surface within the target time slot. ; The aircraft queue length indicator includes the number of aircraft taking off within the target time slot. and the number of aircraft landing within the target time slot ; The time slot resource demand indicator includes the number of departing aircraft launched within the target time slot. The number of approaching aircraft that taxi into the target time slot ; The relative taxiing time metric is the ratio of the average taxiing time within the target time slot to the historical average taxiing time for the corresponding time slot under the same flight schedule. .

[0008] Furthermore, the key performance indicators (KPIs) are defined as follows: in, This indicates the assembly point for departing flights. Indicates the assembly point for inbound flights; and They represent the first The start and end times of each time slot; and These indicate the pushback time and departure time of departing flights, respectively. and These represent the arrival and landing times of inbound flights, respectively. Indicates the first The average taxiing time of all flights within a time slot This represents the historical average taxiing time of flights within the corresponding time slot under different dates and the same flight schedule. Based on the start and end times of the target time slot, the time overlap of flight operations in the departing and arriving flight assemblies is determined, and key indicators of complexity are evaluated based on the determination results; among them, This is an indicator function, and the return value of an indicator function is a binary result; when the condition in the indicator function is satisfied, ;otherwise, .

[0009] Furthermore, step S2 includes: Principal component analysis was performed on the key indicators for complexity evaluation to obtain the contribution rate of each key indicator; let... [ [ ] is the weight matrix, Represents n key indicators The weight vector satisfies: The results of pairwise comparisons are quantified using scaling methods to construct a... pairwise comparison matrix , where matrix elements Indicators With indicators The ratio of the scores between them, and satisfying , ; Calculate the largest eigenvalue of each pairwise comparison matrix and the corresponding weight vector w; solve the weight vector equation to obtain the largest eigenvalue of each pairwise comparison matrix. in, Pairwise comparison matrix The largest eigenvalue; Calculate the consistency ratio of pairwise comparison matrices based on the largest eigenvalue. ,in The number of key indicators, It is a random consistency index of the same order, and the weight vector is used as the index weight when the consistency ratio result meets the preset consistency condition.

[0010] Standardized indicators are obtained by standardizing the key indicators for complexity evaluation: in, Indicates the first The original values ​​of each indicator and Representing indicators The mean and standard deviation; Calculate the first based on standardized indicators and weight vectors. Weighted index for each sample: ; Determining the positive ideal solution based on weighted indices and negative ideal solution ; Calculate the first distance between the weighted index and the positive ideal solution. And calculate the second distance between the weighted index and the negative ideal solution. ;in, and They represent the first The distance from each sample to the positive and negative ideal solutions; and They represent the first and second ideal solutions, respectively. The possible values ​​of each element, For sample index, The total number of samples; The proximity coefficient is calculated based on the first distance and the second distance, and this proximity coefficient is used as the traffic complexity. The traffic complexity is expressed as: in, The proximity coefficient represents the sample scenario and is used to reflect the complexity of the traffic situation at the airport.

[0011] Furthermore, the multiple historical periodic time slots include the time slot corresponding to the day before the target time slot, the time slot corresponding to the seven days before the target time slot, and the time slot corresponding to the twenty-eight days before the target time slot; Step S3 includes: Extract the target time slot weight vector; extract the historical time slot weight vectors corresponding to the time slot one day before the target time slot, the time slot seven days before the target time slot, and the time slot twenty-eight days before the target time slot respectively; concatenate the target time slot weight vector and the historical time slot weight vector to obtain the multi-cycle input structure.

[0012] Furthermore, step S4 includes: Input the multi-period input structure into the multi-period weighted network; The multi-period input structure is standardized by using a normalization layer in a multi-period weighted network to obtain standardized input features. ,in , Indicates the number of candidate similar time slots. This represents the concatenation dimension between the target time slot features and the candidate similar time slot features. Represents multi-period feature dimensions. For normalization layer, This is the result of normalization; The standardized input features are convolved by the first convolutional layer in the multi-period weighted network to extract primary local contrast features between the target time slot and candidate similar time slots, thus obtaining the first feature map. ,in and These are the network parameters for the first convolutional layer. Mapping to the first convolutional layer network; The channel dimension or feature dimension of the first feature map is compressed by the compression layer in the multi-period weighted network to obtain compressed features. ,in Indicates compression transformation; By performing convolutional processing on the compressed features through the second convolutional layer in the multi-period weighted network, a higher-level similarity representation is obtained, resulting in the second feature map. ,in and These are the parameters for the second convolutional layer. Mapping to the second convolutional layer network; The second feature map is downsampled by pooling layers in a multi-period weighted network to reduce feature dimensionality and enhance the response of the main features, resulting in pooled features. ,in This represents a pooling operation, used to aggregate features within a local region; The residual features are obtained by performing residual mapping on the pooling features through residual blocks in a multi-period weighted network. ,in Represents the residual mapping function. This represents the trainable parameters in the residual block.

[0013] Furthermore, step S4 also includes: The residual features are flattened by a flattening layer in a multi-period weighted network, converting the multidimensional feature maps output by the residual blocks into one-dimensional vectors to obtain the flattened features. ; Indicates the flattening operation; The flattened features are mapped using fully connected layers in a multi-period weighted network, and the initial similarity scores for each candidate similar time slot are output based on the flattened feature vectors. ,in ; Indicates the first The initial similarity scores corresponding to each candidate similar time slot; This indicates a fully connected mapping operation; The initial similarity score is nonnegated by using a nonnegative activation layer in a multi-period weighted network to obtain a nonnegative similarity score. ;in This represents the modified linear unit activation function; The non-negative similarity scores are normalized by the normalized output layer in the multi-period weighted network to obtain the similarity weights.

[0014] Furthermore, step S4 also includes; Obtain the historical relative taxiing time index corresponding to each historical period time slot; Based on the similarity weights corresponding to each historical time slot, the historical relative taxiing time indices are weighted and summed to obtain the predicted relative taxiing time indices corresponding to the target time slot. The predicted relative taxiing time index is used as the relative taxiing time index corresponding to the target time slot; The predicted traffic complexity of the target time slot is calculated based on the predicted relative coasting time index and other key complexity evaluation indicators of the target time slot.

[0015] An assessment system for the increased complexity of airport surface traffic situations includes: The complexity evaluation key indicator construction module is used to acquire airport surface operation data and construct complexity evaluation key indicators to characterize the airport surface traffic situation based on the airport surface operation data. The traffic complexity calculation module is used to calculate the traffic complexity corresponding to the target time slot based on key complexity evaluation indicators. The multi-cycle input construction module is used to obtain historical operation data of multiple historical cycle time slots corresponding to the target time slot, and construct a multi-cycle input structure based on the operation data of the target time slot and the historical operation data of multiple historical cycle time slots. The similarity weight determination module is used to determine the similarity weights between multiple historical period time slots and the target time slot based on the multi-period input structure, and to perform weighted processing on the key indicators corresponding to the multiple historical period time slots based on the similarity weights to obtain the predicted key indicators of the target time slot. The traffic complexity prediction module is used to calculate the predicted traffic complexity corresponding to the target time slot based on key prediction indicators.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention constructs key indicators for evaluating the complexity of airport surface traffic conditions, and quantifies the airport surface traffic status based on airport surface operation data, so that the calculation of traffic complexity can comprehensively reflect the operational status of the airport surface.

[0017] 2. This invention calculates the traffic complexity corresponding to the target time slot based on key indicators of complexity evaluation, and constructs a multi-period input structure by combining the target time slot with multiple historical periodic time slots, so that the traffic complexity assessment can not only reflect the current operating status of the target time slot, but also analyze the target time slot in combination with historical operating patterns.

[0018] 3. This invention determines the similarity weights between multiple historical period time slots and the target time slot based on the multi-period input structure, and performs weighted processing on the key indicators corresponding to the historical period time slots according to the similarity weights to obtain the predicted key indicators of the target time slot, so that the prediction of the key indicators of the target time slot can comprehensively consider the operating characteristics of multiple historical period time slots.

[0019] 4. This invention calculates the predicted traffic complexity corresponding to the target time slot based on the predicted key indicators, so that the traffic complexity assessment includes the current traffic complexity and the predicted traffic complexity, providing traffic complexity assessment results for airport surface traffic situation analysis. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the process of implementing the airport surface traffic situation complexity assessment method of the present invention; Figure 2 This is a schematic diagram illustrating the process of constructing key indicators for complexity evaluation provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of the traffic complexity calculation process provided in an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the construction of a multi-cycle input structure provided in an embodiment of the present invention. Detailed Implementation

[0021] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0022] like Figures 1 to 4 The method for assessing the increased complexity of airport surface traffic situations, as shown, includes the following steps: S1. Obtain airport surface operation data and construct key indicators for evaluating the complexity of airport surface traffic based on the airport surface operation data; extract indicators of airport surface operation from a macro level and design them in combination with the topological relationship between aircraft and parking positions, taxiing efficiency and other dimensions to form a comprehensive indicator system for characterizing surface traffic.

[0023] Specifically, this embodiment selects airport flight plan data, airport surface operation data, and historical operation data as airport surface operation data. Key indicators for complexity evaluation include instantaneous surface traffic flow, cumulative surface traffic flow, aircraft queue length, time slot resource demand, and relative taxiing time. Instantaneous flow indicators include the number of aircraft taxiing at the start of the target time slot. The number of aircraft taxiing at the end of the target time slot. The cumulative traffic flow indicator includes the number of departing aircraft whose taxiing process overlaps with the target time slot. The number of arriving aircraft located on the airport surface within the target time slot. The aircraft queue length indicator includes the number of aircraft taking off within the target time slot. and the number of aircraft landing within the target time slot The time slot resource demand indicator includes the number of departing aircraft launched within the target time slot. The number of approaching aircraft that taxi into the target time slot The relative taxiing time metric is the ratio of the average taxiing time within the target time slot to the historical average taxiing time for the corresponding time slot under the same flight schedule. The aforementioned instantaneous traffic flow indicators, cumulative traffic flow indicators, aircraft queue length indicators, time slot resource demand indicators, and relative taxiing time indicators together constitute the set of key indicators for evaluating the complexity of the target time slot, and serve as input data for traffic complexity calculation in step S2.

[0024] Specifically, a key indicator system for complexity evaluation is constructed based on airport surface operation data. This system is built from dimensions such as airport surface operational status, aircraft taxiing status, and airport surface resource requirements. It includes instantaneous surface traffic flow indicators, cumulative surface traffic flow indicators, aircraft queue length indicators, time slot resource requirements indicators, and relative taxiing time indicators. These indicators collectively characterize the airport surface traffic situation within the target time slot. The key indicators are defined as follows: in, This indicates the assembly point for departing flights. Indicates the assembly point for inbound flights; and They represent the first The start and end times of each time slot; and These indicate the pushback time and departure time of departing flights, respectively. and These represent the arrival and landing times of inbound flights, respectively. Indicates the first The average taxiing time of all flights within a time slot This represents the historical average taxiing time of flights within the corresponding time slot under different dates and the same flight schedule; based on the start and end times of the target time slot, the time overlap of flight operations in the departing and arriving flight sets is determined, and the complexity is evaluated based on the determination results; among them, This is an indicator function, and the return value of an indicator function is a binary result; when the condition in the indicator function is satisfied, ;otherwise, .

[0025] Each key indicator for complexity evaluation quantifies the airport surface traffic situation in the target time slot from multiple dimensions, such as airport surface traffic flow scale, aircraft operating status, airport resource demand, and taxiing efficiency. Together, they form a set of key indicators for complexity evaluation corresponding to the target time slot, providing basic data for subsequent traffic complexity calculation.

[0026] S2. Calculate the traffic complexity corresponding to the target time slot based on the key indicators of complexity evaluation; construct a traffic complexity TSC calculation method based on the PCA-AHP-TOPSIS method; use the PCA-AHP-TOPSIS framework to fuse indicators, and propose the concept of relative coasting index as a component of complexity to measure the degree of congestion in the scene.

[0027] Specifically, step S2 includes: Principal component analysis (PCA) was performed on key complexity evaluation indicators to obtain their contribution rates. After obtaining the key complexity evaluation indicators, PCA was used to perform dimensionality reduction analysis on each indicator, extracting the main components that characterize the airport surface traffic situation. The contribution rate of each key complexity evaluation indicator was calculated based on the variance contribution rate of each principal component. By analyzing the contribution rates of each key complexity evaluation indicator, the relative contribution of each indicator to the complexity of the airport surface traffic situation can be obtained, providing an objective basis for comparing the importance of indicators in the subsequent analytic hierarchy process (AHP).

[0028] set up [ [ ] is the weight matrix, Represents n key indicators The weight vector satisfies: Based on the contribution rate of the indicators, the importance of any two key indicators for complexity evaluation is compared to obtain pairwise comparison results; The results of pairwise comparisons are quantified using scaling methods to construct a... pairwise comparison matrix , where matrix elements Indicators With indicators The ratio of the scores between them, and satisfying , ; Calculate the largest eigenvalue of each pairwise comparison matrix and the corresponding weight vector w; solve the weight vector equation to obtain the largest eigenvalue of each pairwise comparison matrix. in, Pairwise comparison matrix The largest eigenvalue; Calculate the consistency ratio of pairwise comparison matrices based on the largest eigenvalue. ,in The number of key indicators, It is a random consistency index of the same order, and the weight vector is used as the index weight when the consistency ratio result meets the preset consistency condition.

[0029] The importance of each key complexity evaluation indicator is compared pairwise, and a judgment matrix is ​​constructed based on the comparison results. The indicator weights corresponding to each key complexity evaluation indicator are obtained by solving the judgment matrix. A consistency check is performed on the constructed judgment matrix. When the consistency check result meets a preset consistency condition, i.e., the consistency ratio, the result is considered satisfactory. If the obtained index weights are used as the final weights of each complexity evaluation key index, then the importance relationships between each complexity evaluation key index are readjusted and the judgment matrix is ​​reconstructed until the consistency requirements are met, thereby ensuring the reliability and rationality of the index weight allocation results.

[0030] Standardized indicators are obtained by standardizing the key indicators for complexity evaluation: in, Indicates the first The original values ​​of each indicator and Representing indicators The mean and standard deviation; Calculate the first based on standardized indicators and weight vectors. Weighted index for each sample: ; Determining the positive ideal solution based on weighted indices and negative ideal solution ; Calculate the first distance between the weighted index and the positive ideal solution. And calculate the second distance between the weighted index and the negative ideal solution. ;in, and They represent the first The distance from each sample to the positive and negative ideal solutions; and They represent the first and second ideal solutions, respectively. The possible values ​​of each element, For sample index, The total number of samples; The proximity coefficient is calculated based on the first distance and the second distance, and this proximity coefficient is used as the traffic complexity. The traffic complexity is expressed as: in, The proximity coefficient represents the sample scenario and is used to reflect the complexity of the traffic situation at the airport.

[0031] After obtaining the weights of each key complexity evaluation indicator, the indicators are standardized to eliminate the impact of differences in indicator dimensions on the comprehensive evaluation results. Based on the standardized key complexity evaluation indicators and their corresponding weights, weighted indices are calculated for each key complexity evaluation indicator, and positive and negative ideal solutions are determined according to these weighted indices. Furthermore, the distances between each sample and the positive and negative ideal solutions are calculated, and the proximity coefficients for each sample are calculated based on these distances. These proximity coefficients are used as the traffic complexity corresponding to the target time slot, characterizing the complexity of the airport surface traffic situation corresponding to the target time slot.

[0032] S3. For the target time slot, acquire historical operation data of multiple historical period time slots corresponding to the target time slot, and construct a multi-period input structure based on the operation data of the target time slot and the historical operation data of multiple historical period time slots; adopt a multi-period comparison input mode and combine it with graph convolutional neural network to construct a prediction framework; through the design of multi-period input structure, reduce the interference of abnormal data on the prediction results, and at the same time explore the potential evolution law of airport situation. Specifically, a predictive feature aggregation set is designed. Multidimensional features are extracted for real-time taxiing time index prediction for the target time slot. These features include overall surface flight characteristics, overall surface traffic characteristics, and overall surface weather characteristics. Among these, the overall surface flight characteristics include average taxiing distance features. And the characteristics of aircraft type distribution, including the proportion of light aircraft. Medium-sized machines account for a certain percentage and heavy machinery proportion The overall traffic situation characteristics include the first eight situation indicators used to characterize the macroscopic traffic operation status within the target time slot. to The overall meteorological characteristics include meteorological assessment features calculated based on meteorological reports corresponding to the target time slot. .Will to Together, they constitute the input feature set of the target time slot and serve as the subsequent real-time taxiing time indicator. The input to the estimation model.

[0033] Specifically, in order to make full use of the periodic changes in airport traffic conditions at different time scales, multiple historical periodic time slots include the time slot corresponding to the day before the target time slot, the time slot corresponding to the seven days before the target time slot, and the time slot corresponding to the twenty-eight days before the target time slot.

[0034] Specifically, the target time slot feature vector and the historical time slot feature vectors corresponding to the previous day, the previous seven days, and the previous twenty-eight days are extracted around the target time slot. The historical features corresponding to each historical period time slot are summarized and statistically analyzed to construct a multi-period feature set that can characterize the dynamic relationship of different time scales.

[0035] Furthermore, the target time slot is paired with its corresponding historical time slots, and the feature vectors of the target time slot and the corresponding historical time slots are concatenated in a preset order to form a multi-period input structure. This multi-period input structure simultaneously includes the current operational features of the target time slot as well as the historical operational features corresponding to the daily, weekly, and monthly periods. This allows for the description of the scene association between the target time slot and historical time slots at different time scales, providing a unified input for subsequent similarity analysis.

[0036] By introducing historical operational characteristics from different periods, the impact of short-term disturbances, sudden events, and changes in the external environment on historical scene matching results can be reduced, improving the data integrity of similarity analysis between the target time slot and historical time slots, and providing basic data for subsequent similarity weight calculation and prediction of key indicators.

[0037] S4. Based on the multi-cycle input structure, determine the similarity weights between multiple historical cycle time slots and the target time slot, and perform weighted processing on the key indicators corresponding to multiple historical cycle time slots according to the similarity weights to obtain the predicted key indicators of the target time slot. Specifically, step S4 includes: The multi-period input structure is input into the multi-period weighted network; the multi-period weighted network includes a normalization layer, a first convolutional layer, a compression layer, a second convolutional layer, a pooling layer, a residual block, a flattening layer, a fully connected layer, a non-negative activation layer, and a normalized output layer.

[0038] The multi-period input structure is standardized by using a normalization layer in a multi-period weighted network to obtain standardized input features. ,in , Indicates the number of candidate similar time slots. This represents the concatenation dimension between the target time slot features and the candidate similar time slot features. Represents multi-period feature dimensions. For normalization layer, To normalize the results; to reduce the impact of differences in the dimensions of different features on the network training process.

[0039] The standardized input features are convolved by the first convolutional layer in the multi-period weighted network to extract primary local contrast features between the target time slot and candidate similar time slots, thus obtaining the first feature map. ,in and These are the network parameters for the first convolutional layer. Mapping to the first convolutional layer network; The channel dimension or feature dimension of the first feature map is compressed by the compression layer in the multi-period weighted network to obtain compressed features. ,in This represents a compression transformation; it reduces redundant information while retaining the main characteristic responses.

[0040] By performing convolutional processing on the compressed features through the second convolutional layer in the multi-period weighted network, a higher-level similarity representation is obtained, resulting in the second feature map. ,in and These are the parameters for the second convolutional layer. This is used to map the second convolutional layer network, thereby obtaining high-level similarity features between the target time slot and the historical periodic time slot at different time scales.

[0041] The second feature map is downsampled by pooling layers in a multi-period weighted network to reduce feature dimensionality and enhance the response of the main features, resulting in pooled features. ,in This represents a pooling operation, used to aggregate features within a local region to reduce feature dimensionality and enhance the response of the main features. The residual features are obtained by performing residual mapping on the pooling features through residual blocks in a multi-period weighted network. ,in Represents the residual mapping function. This represents the trainable parameters in the residual block. It enhances the network's ability to represent complex similarity relationships and improves gradient propagation during deep network training.

[0042] The residual features are flattened by a flattening layer in a multi-period weighted network, converting the multidimensional feature maps output by the residual blocks into one-dimensional vectors to obtain the flattened features. ; This indicates a flattening operation, enabling the similarity features obtained from the aforementioned convolution, pooling, and residual processing to be input into the subsequent weight generation process.

[0043] The flattened features are mapped using fully connected layers in a multi-period weighted network, and the initial similarity scores for each candidate similar time slot are output based on the flattened feature vectors. ,in ; Indicates the first The initial similarity scores corresponding to each candidate similar time slot; This represents a fully connected mapping operation; the fully connected layer is used to generate the initial matching degree between each candidate similar time slot and the target time slot based on the flattening features, so that each candidate similar time slot obtains a corresponding initial similarity score.

[0044] The initial similarity score is nonnegated by using a nonnegative activation layer in a multi-period weighted network to obtain a nonnegative similarity score. ;in This indicates a modified linear unit activation function; it removes negative values ​​from the initial similarity score, ensuring that the subsequent similarity score used for weighted calculation is non-negative, thus satisfying the calculation constraint that the weighting coefficients must be non-negative.

[0045] The non-negative similarity scores are normalized using a normalized output layer in a multi-period weighted network to obtain similarity weights. This normalized output layer converts the non-negative similarity scores corresponding to each candidate similar time slot into normalized weights, ensuring a uniform scale for the similarity weights across all candidate similar time slots. These weights are then used for subsequent weighting of historical relative coasting time indicators.

[0046] After obtaining the similarity weights for each historical period time slot, the historical relative taxiing time index for each historical period time slot is obtained. The historical relative taxiing time index is derived from airport surface operation data for each historical period time slot and is used to characterize the change in aircraft taxiing time within the corresponding historical period time slot relative to the historical average taxiing time.

[0047] The similarity weights corresponding to each historical time slot are used as weighting coefficients to sum the historical relative taxiing time indices corresponding to each historical time slot, thus obtaining the predicted relative taxiing time index for the target time slot. The higher the similarity weight, the greater the proportion of the historical relative taxiing time index of the corresponding historical time slot in the predicted relative taxiing time index.

[0048] The predicted relative taxiing time index is used as the relative taxiing time index corresponding to the target time slot, and is input into the traffic complexity calculation process along with other key complexity evaluation indicators for the target time slot to calculate the predicted traffic complexity corresponding to the target time slot. The predicted traffic complexity is used to characterize the complexity of the airport surface traffic situation after the predicted relative taxiing time index is used in the calculation of the target time slot.

[0049] Furthermore, the parameters of the target network in this embodiment are shown in Table 1. The prediction accuracy of the final calculated key indicators is shown in Table 2. It can be seen from the figure that the method proposed in this study exhibits higher prediction accuracy compared with common machine learning prediction models, and the complexity also shows a significant correlation with other indicators. This result can help the field to perform complexity analysis and forward-looking prediction.

[0050] Table 1 Table 2 An assessment system for the increased complexity of airport surface traffic situations includes: The complexity evaluation key indicator construction module is used to acquire airport surface operation data and construct complexity evaluation key indicators to characterize the airport surface traffic situation based on the airport surface operation data. The traffic complexity calculation module is used to calculate the traffic complexity corresponding to the target time slot based on key complexity evaluation indicators. The multi-cycle input construction module is used to obtain historical operation data of multiple historical cycle time slots corresponding to the target time slot, and construct a multi-cycle input structure based on the operation data of the target time slot and the historical operation data of multiple historical cycle time slots. The similarity weight determination module is used to determine the similarity weights between multiple historical period time slots and the target time slot based on the multi-period input structure, and to perform weighted processing on the key indicators corresponding to the multiple historical period time slots based on the similarity weights to obtain the predicted key indicators of the target time slot. The traffic complexity prediction module is used to calculate the predicted traffic complexity corresponding to the target time slot based on key prediction indicators.

[0051] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0052] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in sequences other than those illustrated or described herein.

[0053] The present invention has been further described above with reference to specific embodiments. However, it should be understood that the specific description herein should not be construed as limiting the nature and scope of the present invention. Various modifications made to the above embodiments by those skilled in the art after reading this specification are all within the scope of protection of the present invention.

Claims

1. An assessment method for increased complexity of airport surface traffic situation, characterized in that, Includes the following steps: S1. Obtain airport surface operation data, and construct key indicators for evaluating the complexity of airport surface traffic situation based on the airport surface operation data. S2. Calculate the traffic complexity corresponding to the target time slot based on the aforementioned key complexity evaluation indicators; S3. For the target time slot, obtain the historical operation data of multiple historical period time slots corresponding to the target time slot, and construct a multi-period input structure based on the operation data of the target time slot and the historical operation data of the multiple historical period time slots; S4. Determine the similarity weights between the multiple historical period time slots and the target time slot based on the multi-period input structure, and perform weighted processing on the key indicators corresponding to the multiple historical period time slots based on the similarity weights to obtain the predicted key indicators of the target time slot; calculate the predicted traffic complexity corresponding to the target time slot based on the predicted key indicators.

2. The airport surface traffic situation complexity enhancement assessment method according to claim 1, characterized in that, The key indicators for complexity evaluation include instantaneous surface traffic flow, cumulative surface traffic flow, aircraft queue length, time slot resource demand, and relative taxiing time. The instantaneous flow rate indicator includes the number of aircraft taxiing at the start of the target time slot. The number of aircraft taxiing at the end of the target time slot. ; The cumulative traffic flow indicator includes the number of departing aircraft whose taxiing process overlaps with the target time slot. The number of arriving aircraft located on the airport surface within the target time slot. ; The aircraft queue length indicator includes the number of aircraft taking off within the target time slot. and the number of aircraft landing within the target time slot ; The time slot resource demand indicator includes the number of departing aircraft launched within the target time slot. The number of approaching aircraft that taxi into the target time slot ; The relative taxiing time metric is the ratio of the average taxiing time within the target time slot to the historical average taxiing time for the corresponding time slot under the same flight schedule. .

3. The airport surface traffic situation complexity enhancement assessment method according to claim 2, characterized in that, The key indicators are defined in the following form: in, This indicates the assembly point for departing flights. Indicates the assembly point for inbound flights; and They represent the first The start and end times of each time slot; and These indicate the pushback time and departure time of departing flights, respectively. and These represent the arrival and landing times of inbound flights, respectively. Indicates the first The average taxiing time of all flights within a time slot This represents the historical average taxiing time of flights within the corresponding time slot under different dates and the same flight schedule. Based on the start and end times of the target time slot, the time overlap of flight operations in the departing flight set and the arriving flight set is determined, and the key complexity evaluation indicators are statistically analyzed based on the determination results; wherein, Let be an indicator function, and let be the return value of a binary result; when the condition in the indicator function is satisfied, ;otherwise, .

4. The airport surface traffic situation complexity enhancement assessment method according to claim 1, characterized in that, Step S2 includes: Principal component analysis was performed on the key indicators of complexity evaluation to obtain the contribution rate of each key indicator of complexity evaluation. set up [ [ ] is the weight matrix, Represents n key indicators The weight vector satisfies: Based on the contribution rate of the aforementioned indicators, the importance of any two key indicators for complexity evaluation is compared to obtain pairwise comparison results; The results of pairwise comparisons are quantified using scaling methods to construct a... pairwise comparison matrix , where matrix elements Indicators With indicators The ratio of the scores between them, and satisfying , ; Calculate the largest eigenvalue of the pairwise comparison matrix and the weight vector w corresponding to the largest eigenvalue; solve the weight vector equation to obtain the largest eigenvalue of the pairwise comparison matrix: in, Pairwise comparison matrix The largest eigenvalue; Calculate the consistency ratio of the pairwise comparison matrices based on the largest eigenvalue. ,in The number of key indicators, The weight vector is used as the weight of the index when the consistency ratio result meets the preset consistency condition.

5. The airport surface traffic situation complexity enhancement assessment method according to claim 4, characterized in that, Step S2 further includes: The standardized indicators are obtained by standardizing the key indicators for complexity evaluation: in, Indicates the first The original values ​​of each indicator and Representing indicators The mean and standard deviation; Calculate the first based on the standardized index and the weight vector. Weighted index for each sample: ; The positive ideal solution is determined based on the weighted index. and negative ideal solution ; Calculate the first distance between the weighted index and the positive ideal solution. And calculate the second distance between the weighted index and the negative ideal solution. ;in, and They represent the first The distance from each sample to the positive and negative ideal solutions; and They represent the first and second ideal solutions, respectively. The possible values ​​of each element, For sample index, The total number of samples; A proximity coefficient is calculated based on the first distance and the second distance, and this proximity coefficient is used as the traffic complexity, which is expressed as follows: in, The proximity coefficient represents the sample scenario and is used to reflect the complexity of the traffic situation at the airport.

6. The airport surface traffic situation complexity enhancement assessment method according to claim 1, characterized in that, The multiple historical period slots include the slot corresponding to the day before the target slot, the slot corresponding to the seven days before the target slot, and the slot corresponding to the twenty-eight days before the target slot; Step S3 includes: Extract the target time slot weight vector of the target time slot; extract the historical time slot weight vectors of the time slot corresponding to the day before the target time slot, the time slot corresponding to the seven days before the target time slot, and the time slot corresponding to the twenty-eight days before the target time slot respectively; concatenate the target time slot weight vector and the historical time slot weight vector to obtain the multi-period input structure.

7. The airport surface traffic situation complexity enhancement assessment method according to claim 6, characterized in that, Step S4 includes: The multi-period input structure is input into a multi-period weighted network; The multi-period input structure is standardized by the normalization layer in the multi-period weighted network to obtain standardized input features. ,in , Indicates the number of candidate similar time slots. This represents the concatenation dimension between the target time slot features and the candidate similar time slot features. Represents multi-period feature dimensions. For normalization layer, This is the result of normalization; The standardized input features are convolved by the first convolutional layer in the multi-period weighted network to extract primary local contrast features between the target time slot and candidate similar time slots, thus obtaining the first feature map. ,in and These are the network parameters for the first convolutional layer. Mapping to the first convolutional layer network; The channel dimension or feature dimension of the first feature mapping is compressed through the compression layer in the multi-period weighted network to obtain compressed features. ,in Indicates compression transformation; The compressed features are convolved through the second convolutional layer in the multi-period weighted network to obtain a higher-level similarity representation, thus yielding the second feature map. ,in and These are the parameters for the second convolutional layer. Mapping to the second convolutional layer network; The second feature map is downsampled through the pooling layer in the multi-period weighted network to reduce feature dimensionality and enhance the main feature response, resulting in pooled features. ,in This represents a pooling operation, used to aggregate features within a local region; The pooling features are processed by residual mapping through residual blocks in the multi-period weighted network to obtain residual features. ,in Represents the residual mapping function. This represents the trainable parameters in the residual block.

8. The airport surface traffic situation complexity enhancement assessment method according to claim 1, characterized in that, Step S4 further includes: The residual features are flattened by the flattening layer in the multi-period weighted network, converting the multidimensional feature maps output by the residual blocks into one-dimensional vectors to obtain the flattened features. ; Indicates the flattening operation; The flattened features are mapped using the fully connected layers in the multi-period weighted network, and the initial similarity scores for each candidate similar time slot are output based on the flattened feature vectors. ,in ; Indicates the first The initial similarity scores corresponding to each candidate similar time slot; This indicates a fully connected mapping operation; The initial similarity score is nonnegated by the nonnegative activation layer in the multi-period weighted network to obtain a nonnegative similarity score. ;in This represents the modified linear unit activation function; The non-negative similarity scores are normalized by the normalized output layer in the multi-period weighted network to obtain the similarity weights.

9. The airport surface traffic situation complexity enhancement assessment method according to claim 1, characterized in that, Step S4 also includes: Obtain the historical relative taxiing time index corresponding to each historical period time slot; Based on the similarity weights corresponding to each historical time slot, the historical relative taxiing time index is weighted and summed to obtain the predicted relative taxiing time index corresponding to the target time slot. The predicted relative taxiing time index is used as the relative taxiing time index corresponding to the target time slot; The predicted traffic complexity of the target time slot is calculated based on the predicted relative coasting time index and the other key complexity evaluation indicators of the target time slot.

10. An airport surface traffic situation complexity enhancement assessment system, characterized in that: include: The complexity evaluation key indicator construction module is used to acquire airport surface operation data and construct complexity evaluation key indicators to characterize the airport surface traffic situation based on the airport surface operation data. The traffic complexity calculation module is used to calculate the traffic complexity corresponding to the target time slot based on the key complexity evaluation indicators. A multi-cycle input construction module is used to acquire historical operating data of multiple historical cycle time slots corresponding to the target time slot for the target time slot, and construct a multi-cycle input structure based on the operating data of the target time slot and the historical operating data of the multiple historical cycle time slots. The similarity weight determination module is used to determine the similarity weight between the multiple historical period time slots and the target time slot based on the multi-period input structure, and to perform weighted processing on the key indicators corresponding to the multiple historical period time slots based on the similarity weight to obtain the predicted key indicators of the target time slot. The traffic complexity prediction calculation module is used to calculate the predicted traffic complexity corresponding to the target time slot based on the predicted key indicators.

Images