Airplane ground taxi path generation method based on high-frequency operation semantic compression

Abstract

The application belongs to the technical field of airport operation management, and proposes an aircraft apron taxiing path generation method based on high-frequency operation semantic compression, which comprises: acquiring an airport digital map and an aircraft apron taxiing path sequence obtained through map matching, constructing an operation semantic primitive set based on the appearance frequency of continuous sub-path segments, and using a greedy matching method to compress the atomic path sequence into an optimized semantic trajectory composed of operation semantic primitives and atomic path segments; performing map-flight-weather multi-source operation information joint semantic coding on the airport digital map, flight origin-destination path information and METAR message to obtain joint semantic features; under the topological constraint of the airport digital map, generating an aircraft apron taxiing path based on the joint semantic features. The application can reduce the representation burden of the path generation model for long path sequences, retain the representability of low-frequency paths, and improve the consistency of the generated taxiing path and the static topological connection relationship of the airport apron.

Description

Technical Field

[0001] This invention belongs to the field of airport operation management technology, and in particular relates to a method for generating aircraft surface taxiing paths based on high-frequency operation semantic compression. Background Technology

[0002] Airport surface operations are a crucial link in ensuring the safety and efficiency of civil aviation operations. With the continuous growth of airport flight volume, the operational relationships of aircraft within the apron, taxiways, runway entrances and exits, and their connecting areas are becoming increasingly complex. The rational generation and rapid adjustment of surface taxi paths have become key issues in airport operations management. Aircraft surface taxi paths not only need to meet the physical connectivity requirements shown on the airport digital map, but also need to conform to flight origin and destination positions, operating directions, control practices, and path selection requirements under complex operating environments. Therefore, developing a method that can automatically generate aircraft surface taxi paths is of great significance for reducing the workload of controllers and operations management personnel and improving the efficiency of surface operations organization.

[0003] Existing aircraft surface trajectory modeling methods typically rely on continuous coordinate points, trajectory sampling points, or map-matched road segment sequences. While continuous coordinate points can describe aircraft motion, they involve large datasets, are susceptible to monitoring errors, sampling frequency differences, and trajectory noise, and struggle to directly reflect airport taxiway structure and air traffic control semantics. Map-matched road segment sequences, while reflecting aircraft paths on airport digital maps, often involve numerous repetitive road segment combinations in long-distance taxiways if atomic segments are used directly as modeling and generation units. This results in long sequences, a heavy generation burden, and hinders the extraction of high-frequency taxiway segments with stable operational implications.

[0004] In actual airport surface operations, aircraft taxiing paths are not arbitrary free-moving trajectories, but are influenced by a variety of factors, including airport surface topology, flight origin-destination routes, weather conditions, and historical operational experience. Some continuous route segments frequently appear in historical operations, reflecting relatively stable operational habits and route organization patterns. Simultaneously, low-frequency but practically significant route adjustments may occur under different operational environments. Relying solely on high-frequency route patterns easily overlooks low-frequency routes; conversely, relying entirely on atomic route sequence sequences makes it difficult to effectively utilize the semantic information contained in high-frequency routes. Therefore, a balance needs to be struck between compressed and complete representation, ensuring that high-frequency routes can be effectively reused while maintaining the representability of low-frequency routes.

[0005] The main challenges in generating aircraft surface taxiing paths are as follows: (1) Aircraft surface trajectory data is characterized by long sequences, numerous repetitive segments, and low semantic levels. Although the road segment sequence after map matching can express the specific travel path, the direct participation of a large number of atomic road segments in the modeling process increases the representation burden in the path generation process. How to extract high-frequency continuous taxiing segments from the atomic road segment sequence and construct them into computable operational semantic primitives is a key problem that needs to be solved to achieve efficient taxiing path generation.

[0006] (2) Aircraft taxiing paths are influenced by multiple sources of information, including airport maps, flight arrival and departure information, and meteorological operational information. Airport digital maps provide static topological constraints, flight arrival and departure segments reflect path mission requirements, and meteorological information such as METAR messages reflect changes in the operational environment. These information sources are different and have different semantic levels. How to uniformly encode them into joint semantic features that can be used for path generation is a major challenge in aircraft taxiing path generation.

[0007] (3) The aircraft surface taxiway path must satisfy the static topological connectivity in the airport digital map. If the connection direction and traversability between road segments are not explicitly introduced during the path generation process, the generated result may have problems such as adjacent road segments not being connected, violating the traversal direction, or crossing non-adjacent taxiways. How to constrain the topological connectivity between candidate generation units during the path generation process and restore it to an atomic road segment sequence that can be used for airport digital map representation after generation is a key issue to ensure the engineering usability of the path result. Summary of the Invention

[0008] To address the shortcomings of the existing technologies, this invention proposes an aircraft taxiway generation method based on high-frequency operational semantic compression. This method can extract high-frequency continuous segment fragments from historical aircraft taxiways into operational semantic primitives while preserving the complete expressive power of atomic segments, thereby reducing the representation burden of long segment sequences. Furthermore, it combines multi-source operational information from maps, flights, and weather to generate aircraft taxiways that satisfy the static topological connectivity of the airport surface.

[0009] The technical solution of the present invention is as follows: A method for generating aircraft surface taxiing paths based on high-frequency operational semantic compression includes the following steps: Step 1: Convert the aircraft surface taxiway sequence after map matching into an optimized semantic trajectory composed of running semantic primitives and atomic segments; Step 2: Convert the airport digital map, flight origin and destination information, and METAR messages into joint semantic features; Step 3: Under the topological constraints of the airport digital map, generate aircraft surface taxiing paths based on joint semantic features.

[0010] Preferably, step 1 specifically includes: Step 1-1: Construct the aircraft surface taxiway sequence and candidate continuous sub-segment set after map matching; Steps 1-2: Count the frequency of each candidate continuous sub-segment in the aircraft surface taxiway sequence, construct all candidate continuous sub-segments that meet the preset frequency conditions into a set of operational semantic primitives, and use all atomic segments as a set of generation units; Steps 1-3: Based on the set of operational semantic primitives, compress the aircraft surface taxiway sequence into an optimized semantic trajectory composed of operational semantic primitives and atomic segments.

[0011] Preferably, step 2 specifically includes: Step 2-1: Define the vectorized representation of airport digital map information, flight origin and destination route information, and METAR messages; Step 2-2: Use a dual-stream encoder to perform joint semantic coding on airport digital map information, flight origin and destination information, and METAR messages; Steps 2-3: Based on residual connections, fuse flight operation semantic features and meteorological operation semantic features to form joint semantic features for taxiway path generation.

[0012] Preferably, step 3 specifically includes: Step 3-1: Construct topological constraints for generating units using topological connection information from the airport digital map; Step 3-2: By restricting the connection relationship between adjacent generating units through the graph constraint attention mechanism, and using semantic path prefixes and joint semantic features, a semantic gliding path is obtained; Step 3-3: Expand each running semantic primitive in the semantic taxiing path into a corresponding atomic segment sequence to obtain the aircraft surface taxiing path.

[0013] Preferably, the set of aircraft surface taxiway sequence in step 1-1 The expression is:

[0014] in, Indicates the first Sequence of aircraft surface taxiways , Indicates the number of taxiways on the aircraft surface. express The Middle m Each atomic segment, , express The number of atomic segments included. , This represents the set of atomic road segments in the airport digital map; Candidate continuous sub-segment set The expression is:

[0015] in, Indicates candidate continuous sub-segments in The starting position in Indicates the length of the candidate continuous sub-segment. This represents the minimum length of a candidate continuous sub-segment. This indicates the maximum length of the candidate continuous sub-segment.

[0016] Preferably, steps 1-3 specifically include: Step 1-3-1: Execute the semantic primitive set according to frequency of occurrence or preset priority. Sorting yields an ordered set of semantic primitives. Step 1-3-2: For the taxiway sequence of each aircraft The scanning pointer is used to scan from the starting position, and the existence of a running semantic primitive is checked sequentially in the ordered set of running semantic primitives. Match with multiple consecutive atomic segments starting from the current position; if so, then run the semantic primitive. Added optimized semantic trajectories and moved the scan pointer backward. If the current location is not specified, the atomic segment corresponding to the current location is added to the optimized semantic trajectory, and the scanning pointer is moved one segment to the right.

[0017] Preferably, the dual-stream encoder in step 2-2 includes a first encoding stream and a second encoding stream, wherein: The first encoded stream is used to encode airport digital map information and flight origin-destination route information to obtain flight operation semantic features. : The second encoded stream is used to encode the METAR message to obtain meteorological operational semantic features. .

[0018] Preferably, the expression for residual connection fusion in steps 2-3 is:

[0019] in, Indicates joint semantic features, This indicates normalization processing. Indicates the fusion weight. A projection function representing the semantic features of meteorological operations.

[0020] Preferably, the element topology constraint expression generated in step 3-1 is:

[0021] in, Represents generating unit The corresponding exit atomic segment, Represents generating unit The corresponding entry atomic segment, This represents the atomic segment topology connectivity matrix in the airport digital map. Represents the set of generating units. Represents generating unit Then allow connection generation unit , Represents generating unit and generating unit The connections between them do not satisfy the static topological connection relationship in the airport digital map.

[0022] Preferably, step 3-2 specifically includes: Step 3-2-1: Based on semantic path prefixes and joint semantic features The generation score for the next position is calculated using an attention decoding function. ; Step 3-2-2: Unify the semantic primitives and atomic segments into candidate generation units, and constrain the candidate generation units according to the topological constraints of the generation units to obtain the generation score after graph constraints. ; Step 3-2-3: Repeat steps 3-2-1 and 3-2-2 until the target road segment is reached, a termination marker is generated, or the preset maximum generation length is reached, thus obtaining the semantic gliding path. .

[0023] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The aircraft surface taxiing path generation method based on high-frequency operational semantic compression proposed in this invention takes into account the long length of aircraft surface road segment sequences and the large number of repetitive segments. This invention improves continuous taxiing segments with stable operational meaning into reusable semantic generation units by mining high-frequency continuous sub-segments and constructing operational semantic primitives, thereby reducing the representation burden when directly using atomic road segment sequences for path generation.

[0024] 2. Considering that aircraft taxiing paths are influenced by airport maps, flight plans, and meteorological operating conditions, this invention adopts a joint semantic coding method of map-flight-meteorological multi-source operating information. By fusing operating information from different sources through a dual-stream encoder and residual connection, the generation process can simultaneously utilize airport topology constraints, flight origin and destination position constraints, and meteorological operating conditions reflected in METAR messages.

[0025] 3. Considering that the generated path needs to satisfy the static topological connection relationship in the airport digital map, the present invention introduces a graph constraint attention mechanism in the path generation process. It uses the road segment connection relationship in the airport digital map to restrict the adjacent connection between candidate generation units, thereby reducing the possibility of generating discontinuous paths, unreachable paths, or paths that violate the travel direction.

[0026] 4. Considering that the final path needs to serve airport surface operations management, this invention restores the generated operational semantic primitives to atomic road segment sequences, so that the output results can directly correspond to the road segment identifiers in the airport digital map, which facilitates subsequent path display, path verification, manual review or surface operations management module calls. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the embodiments will be briefly introduced below. The features and advantages of the present invention can be more clearly understood by referring to the accompanying drawings. The accompanying drawings are schematic and should not be construed as limiting the present invention in any way. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a flowchart of the aircraft surface taxiing path generation method based on high-frequency operational semantic compression according to the present invention.

[0029] Figure 2 This is a flowchart of the high-frequency compression optimization method for gliding trajectories oriented towards runtime semantic primitives in this invention.

[0030] Figure 3 This is a schematic diagram of the data flow of the aircraft surface taxiing segment sequence, operational semantic primitives, mapping relationships between atomic segments, and path generation method in the invention.

[0031] Figure 4 This is a schematic diagram of the joint semantic encoding and decoding method for multi-source operational information of maps, flights and weather in this invention. Detailed Implementation

[0032] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments of the present invention and the features thereof can be combined with each other.

[0033] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0034] This invention provides a method for generating aircraft surface taxiways based on high-frequency operational semantic compression. It acquires an airport digital map and an aircraft surface taxiway segment sequence obtained through map matching, statistically analyzes the frequency of occurrence of consecutive sub-segments, and constructs operational semantic primitives from consecutive sub-segments that meet preset frequency conditions. A greedy matching method is then used to compress the aircraft surface taxiway segment sequence into an optimized semantic trajectory composed of operational semantic primitives and atomic segments. Based on this, a dual-stream encoder is used to perform joint semantic encoding on airport digital map information, flight origin-destination segment information, and METAR messages. Residual connections are used to fuse flight operational semantic features and meteorological operational semantic features to form joint semantic features for taxiway path generation. Finally, during path generation, topological connectivity information from the airport digital map is used to construct graph constraints. A graph constraint attention mechanism is used to restrict the connection relationships between adjacent generation units, and the generated operational semantic primitives are expanded into corresponding atomic segment sequences to obtain the aircraft surface taxiway path.

[0035] like Figures 1-4 As shown, the technical solution of the present invention consists of three parts: high-frequency compression optimization of taxiing trajectory oriented towards operational semantic primitives; joint semantic encoding of multi-source operational information from maps, flights and weather; and taxiing continuous path generation and atomic segment recovery based on graph-constrained attention.

[0036] The aircraft surface taxiing path generation method of the present invention specifically includes the following steps: Step 1: Convert the aircraft surface taxiway sequence after map matching into an optimized semantic trajectory composed of running semantic primitives and atomic segments.

[0037] The input consists of a sequence of aircraft surface taxiways after map matching. This sequence is a sequence of taxiways that has been mapped onto the airport digital map's taxiway markers. The airport digital map includes at least taxiway markers and the connections between the taxiways.

[0038] (1) Map matching of road segment sequences and construction of candidate continuous sub-road segments.

[0039] Let the set of atomic road segments in the airport digital map be:

[0040] in, The first one in the airport digital map Each atomic segment, , Indicates the number of atomic segments.

[0041] Let the set of aircraft surface taxiway sequence sequences after map matching be:

[0042] in, Indicates the first Sequence of aircraft surface taxiways , Indicates the number of taxiways on the aircraft surface. express The Middle m Each atomic segment, , express The number of atomic segments included. , This represents the set of atomic road segments in the airport digital map.

[0043] To extract gliding segments with recurring characteristics from atomic segment sequences, a minimum length for candidate continuous sub-segments is set. and maximum length Enumerate consecutive sub-segments for all road segment sequences to obtain a set of candidate consecutive sub-segments:

[0044] in, Indicates candidate continuous sub-segments in The starting position in This indicates the length of the candidate continuous sub-segment.

[0045] (2) Run the semantic primitive set construction.

[0046] For any candidate continuous sub-segment Count the frequency of its occurrence in all map matching road segment sequences.

[0047] in, Indicates candidate continuous sub-segments The number of atomic segments included; This is an indicator function that takes a value of 1 when the condition within the parentheses is true, and a value of 0 otherwise. Let the preset frequency threshold be... When the frequency of occurrence of candidate continuous sub-segments meets a preset frequency condition, they are identified as operational semantic primitives, resulting in a set of operational semantic primitives:

[0048] in, This represents a semantic primitive consisting of one or more consecutive atomic segments. It should be noted that the preset frequency threshold... This is used only to determine whether candidate continuous sub-segments should be promoted to runtime semantic primitives, and not to delete low-frequency segments or paths. To ensure that low-frequency paths and unconventional paths can still be represented, this invention retains all atomic segments as catch-all generation units. Therefore, the set of generation units in this invention is represented as follows:

[0049] in, It also includes running semantic primitives and atomic segments.

[0050] (3) Semantic compression optimization based on greedy matching.

[0051] After obtaining the set of semantic primitives, the road segment sequence after map matching is greedily compressed. First, the set of semantic primitives is sorted according to frequency of occurrence or preset priority. Sorting yields an ordered set of semantic primitives. For example... Figure 2 As shown, for any road segment sequence Set the scan pointer and begin scanning from the beginning of the road segment sequence. When the scan pointer is at... At that time, sequentially determine whether there is a running semantic primitive in the ordered set of running semantic primitives. Match with consecutive road segments starting from the current location, i.e., determine whether the following conditions are met:

[0052] If a runtime semantic primitive that satisfies the above expression exists, then add that runtime semantic primitive to the optimized semantic trajectory and move the scan pointer forward. The first road segment position; if no semantic primitive satisfying the above formula exists, the atomic road segment corresponding to the current position is added to the optimized semantic trajectory, and the scanning pointer is moved one road segment position to the right. Through the above greedy matching process, the first... The sequence of aircraft surface taxiway segments is compressed into an optimized semantic trajectory:

[0053] in, It is composed of semantic primitives and atomic segments. This indicates an optimization of the semantic trajectory length. Indicates optimized semantic trajectory The first in Each generation unit. This step reduces repetitive expressions in long path sequence by replacing high-frequency continuous gliding segments with running semantic primitives; at the same time, by preserving atomic segments, uncompressed low-frequency segments can still directly participate in subsequent path generation.

[0054] Step 2: Convert the airport digital map, flight origin and destination information, and METAR messages into joint semantic features.

[0055] In this step, airport digital map information, flight origin and destination information, and METAR messages are converted into joint semantic features.

[0056] (1) Definition and vectorized representation of multi-source operation information.

[0057] For flights yet to be generated, obtain their origin and destination route information:

[0058] in, Indicates the starting segment. Indicates the target road segment.

[0059] Obtain the METAR message corresponding to the flight schedule to be generated, denoted as:

[0060] in, This refers to a field or field unit in a METAR message. The METAR message may include one or more of the following: message release time, wind direction, wind speed, visibility, weather phenomena, cloud cover, cloud height, temperature, dew point, and air pressure. This invention does not limit the specific number of fields or their organization. Let the vectorized representations of airport digital map information, flight origin-destination route information, and the METAR message be as follows:

[0061] in, This represents a digital map of the airport. , and These represent the vectorized representation functions of the corresponding information. These vectorized representation functions can be implemented using embedding tables, linear mappings, sequence encoding, or combinations thereof; this invention does not limit their specific network structure or parameter form.

[0062] (2) Construction of dual-stream encoder.

[0063] To avoid semantic interference caused by direct mixing of information from different sources in the original input layer, this invention employs a dual-stream encoder for joint semantic coding.

[0064] The first encoding stream is used to encode airport digital map information and flight origin-destination route information to obtain flight operation semantic features:

[0065] in, Indicates the first encoded stream, It represents the semantic features of flight operations formed by the combination of airport digital maps and flight origin and destination information.

[0066] The second encoded stream is used to encode the METAR message to obtain meteorological operational semantic features:

[0067] in, Indicates the second encoded stream, This represents the meteorological operational semantic features formed by METAR messages.

[0068] Through the aforementioned dual-flow structure, the path task features formed by flight origin-destination constraints and airport map information, and the meteorological operating environment features formed by METAR messages are extracted respectively, which is beneficial for subsequent fusion within a unified semantic space.

[0069] (3) Feature fusion and training objectives based on residual connections.

[0070] Since flight operation semantic features and meteorological operation semantic features come from different sources, they can first be mapped to a consistent feature space through a projection function, and then fused through residual connections:

[0071] in, This represents the joint semantic features used for taxiing path generation; A projection function representing the semantic features of meteorological operations; This represents the fusion weights, which can be learnable parameters or preset parameters; This indicates normalization processing. As can be seen from the above formula, the flight operation semantic features corresponding to the airport map and flight origin-destination route information serve as the main path generation conditions, while the meteorological operation semantic features corresponding to the METAR report participate in condition correction through residual terms.

[0072] During the training phase, the optimized semantic trajectory obtained in step 1 These can be used as supervisory labels to train the gliding path semantic generation model. The gliding path semantic generation model includes a multi-source information vectorization module, a two-stream encoder, a feature fusion module, and a graph-constrained attention decoder. Its parameters to be learned are denoted as... For the first The training samples include airport digital map information as input. Flight origin and destination information and the corresponding METAR message at runtime Its supervised output is an optimized semantic trajectory composed of running semantic primitives and atomic segments. The purpose of training is to enable the taxiing path semantic generation model to learn the mapping relationship between airport maps, flight missions, and weather conditions to optimize semantic trajectories, so that the model can optimize semantic trajectories given pre-generated semantic path prefixes. In this case, it is possible to predict the next generating unit.

[0073]

[0074] in, This indicates that when the model parameters are At that time, the taxiway semantic generation model is based on airport digital map information. , No. Information on the origin and destination segments of the sample flights METAR message and currently generated semantic path prefixes The next generating unit is predicted as The conditional probability is obtained by processing the generation score output by the graph-constrained attention decoder through topological constraint processing and normalization. Candidate generation units belong to the generation unit set. . This represents the training loss function, used to measure the semantic trajectories generated by the model relative to the supervision labels. The difference between them; the loss function is a negative log-likelihood loss, equivalent to accumulating the cross-entropy of the next generating unit predictions for each generating position in each training sample. By minimizing Update model parameters This improves the model's prediction probability for each generating unit in the true optimized semantic trajectory.

[0075] Step 3: Generation of continuous gliding paths and restoration of atomic segments based on graph-constrained attention.

[0076] This step generates aircraft surface taxiing paths under the topological constraints of the airport digital map. This step uses semantic primitives and atomic segments as candidate generation units and restricts the connectivity between adjacent generation units during the generation process.

[0077] (1) Generate unit topology constraints.

[0078] During the path generation phase, it is necessary to determine whether candidate generation units can be connected adjacently based on the road segment connectivity relationships in the airport digital map. For any generation unit... If it is an atomic segment, then the generation unit directly corresponds to one atomic segment; if it is a runtime semantic primitive, then the generation unit corresponds to a continuous subsequence composed of multiple atomic segments. To uniformly represent the connection relationships between generation units, we define... Represents generating unit The corresponding entry atomic segment, Represents generating unit The corresponding exit atomic segment. For an atomic segment, both the entry and exit atomic segments are themselves; for a runtime semantic primitive, the entry atomic segment is the first segment of its corresponding atomic segment subsequence, and the exit atomic segment is the last segment of its corresponding atomic segment subsequence. Let the atomic segment topology connection matrix in the airport digital map be... .according to Construct the topological constraint matrix between generating units:

[0079] in, Represents generating unit The corresponding exit atomic segment, Represents generating unit The corresponding entry atomic segment, Represents the set of generating units. Represents generating unit Then allow connection generation unit , Represents generating unit and generating unit The connections between them do not satisfy the static topological connection relationship in the airport digital map.

[0080] Through the above formula, the connection relationship of atomic road segments in the airport digital map is extended to the generation unit level composed of operational semantic primitives and atomic road segments, providing a constraint basis for subsequent graph-constrained attention generation.

[0081] (2) Semantic gliding path generation based on graph-constrained attention.

[0082] During the generation phase, let the generated semantic path prefix be:

[0083] in, Generate a unit for the end of the current path. Based on the semantic path prefix and the joint semantic features obtained in step 2. The generation score for the next position is calculated using the attention decoding function:

[0084] in, , This represents a graph-constrained attention decoding function. Represents candidate generation unit In the The unconstrained generation scores for each position are calculated. To ensure the static topological continuity of the generated path, candidate generation units are constrained according to the topological constraints of the generation units. When At that time, for candidate generation units The generated score after graph constraints for:

[0085] in, Indicates the first The current path end generation unit whose position has been determined Represents the set of generating units The first in One candidate next generation unit; Represents generating unit With candidate generation unit The topological connection constraint identifier between them. When When, it means The corresponding exit atomic segment and The corresponding entrance atomic segments satisfy the connectability relationship in the airport digital map topology connection matrix, and are candidate generation units. Allowed as the next generation unit; when When this occurs, it indicates that the two do not satisfy the static topological connection relationship in the airport digital map, and the candidate generation unit is... When the current location is blocked, its generated score is set to [value]. Or a sufficiently small invalid value. Based on the generated score after graph constraints, select the next generating unit. The selection method can be maximum probability selection, bundle search, or other candidate search methods; this invention is not limited to these methods. Repeat the above generation process until the target road segment is reached, a termination marker is generated, or the preset maximum generation length is reached, to obtain the semantic gliding path:

[0086] in, , Because the candidate generation units are restricted by a topological constraint matrix during the generation process, the semantic taxiing path satisfies the static connection relationship in the airport digital map at the generation unit level.

[0087] (3) Run the semantic primitives to restore the atomic segment sequence and output it.

[0088] Semantic glide path The generating units in the semantic taxiway can be either atomic segments or runtime semantic primitives. To obtain taxiways that directly correspond to the airport digital map, the semantic taxiway needs to be restored to a sequence of atomic segments. For any generating unit in the semantic taxiway, if it is a runtime semantic primitive, it is expanded into the corresponding atomic segment subsequence; if it is an atomic segment, it is directly retained. The atomic segment subsequences corresponding to each generating unit are concatenated according to the generation order to obtain the aircraft surface taxiway.

[0089] in, Represents generating unit The corresponding atomic segment or atomic segment subsequence, This indicates that the data is concatenated in the order it was generated. The final output... This is an aircraft surface taxiway in the form of an atomic segment sequence. This path can be used for airport surface map display, path continuity verification, controller review, candidate taxiway generation, or subsequent invocation by airport surface operations management modules.

[0090] In summary, this invention achieves high-frequency operational semantic compression of aircraft surface route sequences through step 1, joint semantic encoding of multi-source operational information from maps, flights, and weather through step 2, and semantic taxiway path generation and atomic route recovery under graph constraints through step 3. This reduces the representational burden of long route sequence sequences while preserving the representability of low-frequency paths and improving the consistency between the generated taxiway paths and the static topological connections of the airport digital map.

[0091] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0092] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" of the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0093] In this invention, the terms "first," "second," "third," and "fourth" are used for descriptive purposes only and should not be construed as indicating or implying relative importance. The term "multiple" refers to two or more unless otherwise expressly defined.

[0094] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for generating aircraft surface taxiing paths based on high-frequency operational semantic compression, characterized in that, Includes the following steps: Step 1: Convert the aircraft surface taxiway sequence after map matching into an optimized semantic trajectory composed of running semantic primitives and atomic segments; Step 2: Convert the airport digital map, flight origin and destination information, and METAR messages into joint semantic features; Step 3: Under the topological constraints of the airport digital map, generate aircraft surface taxiing paths based on joint semantic features.

2. The aircraft surface taxiing path generation method according to claim 1, characterized in that, Step 1 specifically includes: Step 1-1: Construct the aircraft surface taxiway sequence and candidate continuous sub-segment set after map matching; Steps 1-2: Count the frequency of each candidate continuous sub-segment in the aircraft surface taxiway sequence, construct all candidate continuous sub-segments that meet the preset frequency conditions into a set of operational semantic primitives, and use all atomic segments as a set of generation units; Steps 1-3: Based on the set of operational semantic primitives, compress the aircraft surface taxiway sequence into an optimized semantic trajectory composed of operational semantic primitives and atomic segments.

3. The aircraft surface taxiing path generation method according to claim 2, characterized in that, Step 2 specifically includes: Step 2-1: Define the vectorized representation of airport digital map information, flight origin and destination route information, and METAR messages; Step 2-2: Use a dual-stream encoder to perform joint semantic coding on airport digital map information, flight origin and destination information, and METAR messages; Steps 2-3: Based on residual connections, fuse flight operation semantic features and meteorological operation semantic features to form joint semantic features for taxiway path generation.

4. The aircraft surface taxiing path generation method according to claim 3, characterized in that, Step 3 specifically includes: Step 3-1: Construct topological constraints for generating units using topological connection information from the airport digital map; Step 3-2: By restricting the connection relationship between adjacent generating units through the graph constraint attention mechanism, and using semantic path prefixes and joint semantic features, a semantic gliding path is obtained; Step 3-3: Expand each running semantic primitive in the semantic taxiing path into a corresponding atomic segment sequence to obtain the aircraft surface taxiing path.

5. The aircraft surface taxiing path generation method according to claim 4, characterized in that, The set of aircraft surface taxiway sequence in step 1-1 The expression is: in, Indicates the first Sequence of aircraft surface taxiways , Indicates the number of taxiways on the aircraft surface. express The Middle m Each atomic segment, , express The number of atomic segments included. , This represents the set of atomic road segments in the airport digital map; Candidate continuous sub-segment set The expression is: in, Indicates candidate continuous sub-segments in The starting position in Indicates the length of the candidate continuous sub-segment. This represents the minimum length of a candidate continuous sub-segment. This indicates the maximum length of the candidate continuous sub-segment.

6. The aircraft surface taxiing path generation method according to claim 5, characterized in that, Steps 1-3 specifically include: Step 1-3-1: Execute the semantic primitive set according to frequency of occurrence or preset priority. Sorting yields an ordered set of semantic primitives. Step 1-3-2: For the taxiway sequence of each aircraft The scanning pointer is used to scan from the starting position, and the existence of a running semantic primitive is checked sequentially in the ordered set of running semantic primitives. Match with multiple consecutive atomic segments starting from the current position; if so, then run the semantic primitive. Added optimized semantic trajectories and moved the scan pointer backward. If the current location is not specified, the atomic segment corresponding to the current location is added to the optimized semantic trajectory, and the scanning pointer is moved one segment to the right.

7. The aircraft surface taxiing path generation method according to claim 6, characterized in that, The dual-stream encoder in step 2-2 includes a first encoding stream and a second encoding stream, wherein: The first encoded stream is used to encode airport digital map information and flight origin-destination route information to obtain flight operation semantic features. : The second encoded stream is used to encode the METAR message to obtain meteorological operational semantic features. .

8. The aircraft surface taxiing path generation method according to claim 7, characterized in that, The expression for residual connection fusion in steps 2-3 is: in, Indicates joint semantic features, This indicates normalization processing. Indicates the fusion weight. A projection function representing the semantic features of meteorological operations.

9. The aircraft surface taxiing path generation method according to claim 8, characterized in that, The expression for generating the element topology constraint in step 3-1 is as follows: in, Represents generating unit The corresponding exit atomic segment, Represents generating unit The corresponding entry atomic segment, This represents the atomic segment topology connectivity matrix in the airport digital map. Represents the set of generating units. Represents generating unit Then allow connection generation unit , Represents generating unit and generating unit The connections between them do not satisfy the static topological connection relationship in the airport digital map.

10. The aircraft surface taxiing path generation method according to claim 9, characterized in that, Step 3-2 specifically includes: Step 3-2-1: Based on semantic path prefixes and joint semantic features The generation score for the next position is calculated using an attention decoding function. ; Step 3-2-2: Unify the semantic primitives and atomic segments into candidate generation units, and constrain the candidate generation units according to the topological constraints of the generation units to obtain the generation score after graph constraints. ; Step 3-2-3: Repeat steps 3-2-1 and 3-2-2 until the target road segment is reached, a termination marker is generated, or the preset maximum generation length is reached, thus obtaining the semantic gliding path. .

Images