Ship lockage behavior rule analysis method based on trajectory data driving

Abstract

The invention discloses a ship lockage behavior rule analysis method based on trajectory data driving, and relates to the technical field of ship lockage behavior analysis, and the method comprises the steps: carrying out the ship lockage link analysis through employing a ship trajectory and a ship stay area, and obtaining a ship lockage link time consumption analysis result; according to the ship trajectory and the channel center line, fusing multi-parameter constraint and grid division to carry out feature analysis to obtain a ship stay feature analysis result; performing multi-parameter coupled ship following behavior analysis according to the ship trajectory data and the ship lockage record to obtain a ship following behavior analysis result; and obtaining a ship lockage link time consumption analysis result, a ship staying characteristic analysis result and a ship following behavior analysis result as ship lockage behavior rule analysis results. According to the method, the regions with abnormal motion characteristics in the ship lockage process can be quickly identified, and the ship following characteristic parameter time sequence analysis is adopted, so that the abnormal conditions influencing the ship lock lockage efficiency can be efficiently and intuitively analyzed.

Description

Technical Field

[0001] The present invention relates to the technical field of ship lock-passing behavior analysis, and in particular to a ship lock-passing behavior law analysis method driven by trajectory data. Background Art

[0002] With the widespread use of AIS (Automatic Identification System) and Beidou satellite navigation system, more and more ship trajectories are recorded. Through in-depth analysis and semantic mining of massive trajectory data, we can better understand the movement characteristics of ships and reveal the behavior laws of ships, which has important research significance for the safe management of ship navigation. As a box-shaped navigation structure used to overcome the concentrated water level difference and ensure the smooth passage of ships, the ship lock not only changes the passage mode and speed of ships in the waterway, but also forms a highly dense phenomenon of ships upstream and downstream, increasing the complexity of ship navigation and safety management. The ship lock capacity is a key indicator to measure the performance of the ship lock, which is affected by many factors such as line layout, scale design, draft depth, loading status, etc. Mining the implicit ship movement laws from trajectory data and identifying the characteristics of ship passing behavior can provide a solid theoretical basis and scientific decision-making support for optimizing ship scheduling organization, improving the ship lock capacity, and ensuring the safe passage of ships.

[0003] At present, a lot of research has been conducted on the analysis of ship motion patterns and the identification of ship behavior characteristics using methods such as data analysis, pattern recognition, and artificial intelligence. A large number of studies can achieve the extraction of key trajectory features and the effective detection of abnormal trajectories, which helps to understand the laws of ship motion and improve the understanding of ship behavior. However, most current research focuses on ships in waters such as sea, nearshore, ports, and ferries. There are relatively few analyses of ship trajectories in locks and specific waters upstream and downstream. Existing methods are difficult to directly apply to the analysis of ship behavior patterns when passing through locks. This is specifically manifested in the following aspects:

[0004] (1) Ships generally need to go through multiple staged dispatches when passing through the lock. Affected by the lock dispatch mechanism, ships will stop at anchorages, waiting areas, berthing piers and lock chambers. The movement trajectory of ships is significantly different from the stop behavior of ships when they are sailing normally in the channel and berthing at the port.

[0005] (2) The special structure of the lock has changed the navigation mode of ships, so that ships can only sail in narrow waters with limited width. At the same time, due to the constraints of specific navigation rules and human intervention, the ship's route, speed, heading and berthing position will change significantly. These changes are obviously different from when the ship is sailing in wide seas or specific channels;

[0006] (3) When navigating in narrow and restricted waterways, ships need to take into account both safe navigation and efficient passage through locks. Therefore, ships will closely follow the ship in front while ensuring that they maintain a safe distance from the ship in front, which is a typical characteristic of following behavior.

[0007] In summary, the behavior of ships passing through locks has its own unique rules. How to analyze and mine the motion trajectory data of ships passing through locks to reveal their behavior rules is a key issue that has not yet been solved. Therefore, a ship passing lock behavior law analysis method driven by trajectory data is urgently needed to solve the deficiencies in the existing technology. Summary of the invention

[0008] The purpose of this invention is to propose a method for analyzing the laws of ship passing lock behavior based on trajectory data drive, revealing the inherent laws of ship passing lock behavior from the two dimensions of microscopic ship motion and macroscopic trajectory aggregation, analyzing the links that restrict the efficiency of passing locks and abnormal ship behavior in the operation of large locks, and providing theoretical support and practical guidance for optimizing lock scheduling strategies and improving lock passing capacity.

[0009] To achieve the above object, the present invention provides a method for analyzing the behavior rules of ships passing through locks based on trajectory data, comprising the following steps:

[0010] S1. Analyze the ship passing through the lock using the ship trajectory and the ship stop area to obtain the time consumption analysis result of the ship passing through the lock;

[0011] S2, performing feature analysis based on the ship trajectory and the channel centerline, integrating multi-parameter constraints and grid division, to obtain a ship stay feature analysis result;

[0012] S3, performing a multi-parameter coupled ship following behavior analysis based on the ship trajectory data and the ship passing record to obtain a ship following behavior analysis result;

[0013] S4. Obtain the time consumption analysis result of the ship passing through the lock, the ship stay characteristic analysis result and the ship following behavior analysis result as the ship passing through the lock behavior regularity analysis result.

[0014] Optionally, the ship track and the ship stop area are used to analyze the ship passing through the lock to obtain the time consumption analysis results of the ship passing through the lock, including:

[0015] S1-1, obtaining a set of ship trajectory points using the ship trajectory and the ship stop area;

[0016] S1-2, obtaining a ship crossing point set according to the ship trajectory point set;

[0017] S1-3, obtaining a set of candidate critical points according to the set of ship crossing points;

[0018] S1-4, using the ship stop information to filter and screen the candidate critical point set in turn to obtain the ship lock process node set;

[0019] S1-5, analyzing the ship passing through the lock according to the ship passing through lock process node set to obtain the time consumption pattern of the ship in the stay area and the time consumption pattern of the ship passing through the lock as the time consumption analysis result of the ship passing through the lock;

[0020] The ship stop information includes the ship stop duration and the average speed of the ship during the stop period.

[0021] Optionally, obtaining a set of ship trajectory points using the ship trajectory and the ship stop area includes:

[0022] S1-1-1, performing a buffer operation according to the ship stay area to obtain a ship extended stay area;

[0023] S1-1-2, performing a spatial intersection operation on the ship trajectory and the extended ship stay area to obtain a set of trajectory points to be analyzed in the ship stay area;

[0024] S1-1-3, obtaining a set of trajectory points to be analyzed in the ship's stop area as the ship's trajectory point set;

[0025] The calculation formula for the extended ship stay area is as follows:

[0026] D'=ST_Buffer(D,dis)

[0027] The calculation formula of the ship trajectory point set is as follows:

[0028] P'=ST_Intersects(D',P)

[0029] Where D' is the extended stop area of ​​the ship, ST_Buffer(·) is the buffer operation, D is the ship stop area, dis is the buffer distance, P' is the set of trajectory points to be analyzed in the ship stop area, ST_Intersects(·) is the spatial intersection operation, and P is the ship trajectory.

[0030] Optionally, obtaining a ship crossing point set according to the ship trajectory point set includes:

[0031] S1-2-1, using the ship trajectory point set to obtain an initial ship trajectory point;

[0032] S1-2-2, according to the initial ship trajectory point, obtaining the corresponding forward center of gravity point and the rearward center of gravity point as the center of gravity point of the ship trajectory;

[0033] S1-2-3, obtaining the ship crossing point set according to the center of gravity of the ship trajectory and the ship stop area;

[0034] The forward center of gravity is calculated as follows:

[0035] O(i,k)=ST_Centroid(p i ' -k+1 ,p i ' -k+2 ,…,p i ')

[0036] The calculation formula of the rearward center of gravity is as follows:

[0037] O'(i,k)=ST_Centroid(p i ',p i ' +1 ,…,p i ' +k-1 )

[0038] Where D' is the extended stop area of ​​the ship, ST_Buffer(■) is the buffer operation, D is the ship stop area, dis is the buffer distance, P' is the set of trajectory points to be analyzed in the ship stop area, ST_Intersects(■) is the spatial intersection operation, and P is the ship trajectory.

[0039] Optionally, acquiring the ship crossing point set according to the center of gravity of the ship trajectory and the ship stop area includes:

[0040] S1-2-3-1. Obtain the spatial relationship between the forward center of gravity and the ship's stop area and the spatial relationship between the backward center of gravity and the ship's stop area according to the center of gravity of the ship's trajectory and the ship's stop area;

[0041] S1-2-3-2, judging whether the spatial relationship between the forward center of gravity point and the ship stop area changes from not included to included according to the first judgment condition, if so, obtaining the forward center of gravity point as the ship entry point, and executing S1-2-3-3, otherwise, returning to S1-2-1;

[0042] S1-2-3-3, judging whether the spatial relationship between the rearward center of gravity point and the ship stay area changes from inclusion to non-inclusion according to the second judgment condition, if so, obtaining the rearward center of gravity point as the ship departure point, and executing S1-2-3-4, otherwise, returning to S1-2-1;

[0043] S1-2-3-4, using the ship entry point and the ship exit point as the ship crossing point set;

[0044] The calculation formula of the first judgment condition is as follows:

[0045]

[0046] The calculation formula of the second judgment condition is as follows:

[0047]

[0048] Where ST_Within(·) is the spatial inclusion operation function, O(a,k) is the ath forward centroid point, O(a-1,k) is the a-1th forward centroid point, and d j is the j-th ship stop area, O'(b,k) is the b-th rearward center of gravity, O'(b+1,k) is the b+1-th rearward center of gravity, True is correct, False is wrong, a, b, j are all constants.

[0049] Optionally, the ship stay feature analysis is performed based on the ship trajectory and the channel centerline by fusing multi-parameter constraints and grid division to obtain the ship stop feature analysis results, including:

[0050] S2-1, obtaining a ship channel by segmenting the channel centerline;

[0051] S2-2, using the ship channel and the ship track to perform track point channel mapping to obtain the ship track to be processed;

[0052] S2-3, assigning attributes to the ship track points of the ship track to obtain ship track point attributes;

[0053] S2-4, performing data cleaning and trajectory smoothing processing in sequence according to the to-be-processed ship trajectory and the attributes of the ship trajectory points to obtain the target ship trajectory;

[0054] S2-5, determining whether the target ship trajectory is a single ship trajectory, if so, executing S2-6, otherwise, directly executing S2-7;

[0055] S2-6, acquiring ship stay characteristics according to the target ship trajectory by fusing multi-parameter constraints;

[0056] S2-7, obtaining the ship stay characteristics by using network division according to the target ship trajectory and the channel centerline;

[0057] S2-8, performing feature analysis according to the ship stay feature to obtain ship stay point distribution information and ship top flow index information as the ship stay feature analysis result;

[0058] Wherein, the ship trajectory includes a plurality of ship trajectory points.

[0059] Optionally, acquiring ship stay features by fusing multi-parameter constraints according to the target ship trajectory includes:

[0060] S2-6-1. Set the ship running distance threshold, ship running time threshold and ship running speed threshold as multi-parameter constraint conditions;

[0061] S2-6-2. Based on the Stop / Move model, the target ship trajectory is detected using the multi-parameter constraint condition to obtain a set of candidate stop points;

[0062] S2-6-3. Perform spatial intersection operation on the candidate stay point set and the ship stay area to obtain normal ship stay points and abnormal ship stay points as the ship stay features.

[0063] Optionally, obtaining the ship stay feature by using network division according to the target ship trajectory and the channel centerline includes:

[0064] S2-7-1, grid division is performed according to the center line of the waterway to obtain a basic unit for density calculation;

[0065] S2-7-2, obtaining a set of ship trajectory points contained in the grid cell according to the ship trajectory mapping points of the target ship trajectory and the grid cell of the density calculation;

[0066] S2-7-3, obtaining a density weight parameter of the grid cell according to the set of ship trajectory points contained in the grid cell;

[0067] S2-7-4, calculating the stay index of the grid unit using the density weight parameter of the grid unit;

[0068] S2-7-5, obtaining a density threshold according to the residence index of the grid unit;

[0069] S2-7-6, determining whether the stay index of the grid unit is lower than the density threshold, if so, obtaining the grid unit as a ship smooth point, otherwise, obtaining the grid unit as a ship congestion point;

[0070] S2-7-7, obtaining a thematic map of the retention index and a thematic map of the patency index according to the retention index of the grid unit by using a graded coloring method;

[0071] S2-7-8, using the ship smooth spot, the ship congestion spot, the stop index thematic map and the smooth index thematic map as the ship stop feature;

[0072] The basic unit of the density calculation is a plurality of equally spaced grid units.

[0073] Optionally, the calculation formula of the stay index of the grid unit is as follows:

[0074]

[0075] Where σ(m) represents the residence index of the mth grid unit, d is the length of the grid unit, and w i is the density weight parameter of the i-th grid unit, and k is a constant.

[0076] Optionally, the ship following behavior analysis is performed based on the ship trajectory data and the ship passing record to obtain the ship following behavior analysis results, including:

[0077] S3-1, using the ship lock passing record to obtain the lock passing record;

[0078] S3-2, obtaining a corresponding ship trajectory line set as the ship trajectory at the lock according to the ship passing record at the lock and the ship trajectory data;

[0079] S3-3, preprocessing is performed according to the ship trajectory of the lock to obtain corresponding ship trajectory mapping points;

[0080] S3-4, performing linear interpolation and mean resampling on the trajectory attributes of the ship trajectory mapping points to obtain a resampled time series and the trajectory of the ship at the same lock;

[0081] S3-5, performing a time series analysis of the following characteristic parameters according to the ship trajectory mapping points to obtain a time series analysis result of the following characteristic parameters of the ships in the same lock;

[0082] S3-6, performing a ship following distance time series analysis based on the resampled time series and the trajectory of the ship at the same lock to obtain a ship following distance time series analysis result;

[0083] S3-7, performing a coupling relationship analysis of the following characteristic parameters based on the time series analysis result of the following characteristic parameters and the time series analysis result of the ship following distance, to obtain a following distance-speed coupling relationship and a following distance-acceleration coupling relationship as the following characteristic parameter coupling relationship analysis result;

[0084] S3-8. Obtain the time series analysis result of the car-following characteristic parameter and the coupling relationship analysis result of the car-following characteristic parameter as the ship's car-following behavior analysis result.

[0085] Compared with the closest prior art, the present invention has the following beneficial effects:

[0086] The present invention adopts a trajectory stop point extraction method with multi-parameter constraints, which can quickly identify areas with abnormal motion characteristics during the ship passing through the lock, and provide a more targeted basis for the ship lock scheduling management; the present invention can analyze the spatiotemporal aggregation of ship trajectories through stop index analysis and congestion index analysis, and intuitively display the operating status of the lock and the surrounding waterway; the ship following characteristic parameter time series analysis adopted by the present invention, by analyzing the change trend of characteristic parameters such as position, speed, acceleration, and following distance of ships at the same lock, is helpful to efficiently and intuitively analyze abnormal conditions that affect the efficiency of the lock, such as untimely ship start-up, slow sailing speed, and disorderly and loose entry and exit; the present invention combines the coupling relationship curve and data density distribution to intuitively display the abnormal data distribution in the ship passing through the lock, and can provide important data support for improving the lock passing capacity; starting from the two dimensions of microscopic ship motion and macroscopic trajectory aggregation, the present invention reveals the inherent law of ship passing through the lock, analyzes the links that restrict the efficiency of passing through the lock and the abnormal behavior of ships in the operation of large locks, and provides theoretical support and practical guidance for the optimization of lock scheduling strategy and the improvement of lock passing capacity. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0088] Figure 1 This is a flow chart of a method for analyzing ship lock-passing behavior rules based on trajectory data driven by an embodiment of the present invention;

[0089] Figure 2 Flowchart of the stop area crossing identification algorithm proposed in the embodiment of the present invention, wherein (a) is a flowchart of selecting a set of ship trajectory points based on the ship stop area, (b) is a flowchart of identifying a set of crossing points based on the center of gravity point, and (c) is a flowchart of merging candidate crossing points based on multi-condition judgment;

[0090] Figure 3 A schematic diagram of the time consumption analysis result of the ship stay area proposed in an embodiment of the present invention;

[0091] Figure 4 A schematic diagram of the time consumption analysis results of each link of gate-passing proposed in an embodiment of the present invention;

[0092] Figure 5 A flow chart of a method for extracting stop features of ships passing through a lock proposed in an embodiment of the present invention;

[0093] Figure 6Schematic diagram of the distribution of ship stopover points proposed in an embodiment of the present invention, wherein (a) is the temporal and spatial distribution of the stopover points, (b) is the mooring situation at the upstream anchorage, (c) is the mooring situation at the downstream waiting area, and (d) is the mooring situation at the downstream anchorage;

[0094] Figure 7 The results of extracting trajectory stop features using the LINE-STING algorithm proposed in an embodiment of the present invention, where (a) is a stop index thematic map with the number of trajectory points as the weight, (b) is a congestion index thematic map with the average speed (including stationary points) as the weight, and (c) is a congestion index thematic map with the average speed (excluding stationary points) as the weight;

[0095] Figure 8 A flow chart of a car-following behavior analysis algorithm proposed in an embodiment of the present invention;

[0096] Fig. 9 This is a time series analysis diagram of the following characteristic parameters of the ship at the same lock proposed in an embodiment of the present invention, where (a) is the change in ship position, (b) is the change in ship speed, (c) is the change in ship acceleration, and (d) is the change in ship following distance;

[0097] Fig.10 Graph showing the coupling relationship analysis results of the following characteristic parameters proposed in an embodiment of the present invention, wherein (a) is the following distance-speed coupling relationship, and (b) is the following distance-acceleration coupling relationship. DETAILED DESCRIPTION

[0098] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in combination with the specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0099] The terms used in the embodiments of the present invention are only used to explain the specific embodiments of the present invention and are not intended to limit the present invention.

[0100] At present, a lot of research has been carried out on the analysis of ship motion laws and the identification of ship behavior characteristics using methods such as data analysis, pattern recognition and artificial intelligence.

[0101] Trajectory stop points divide continuous trajectories into several trajectory segments with similar shapes, which can reflect the purpose and intention of ship activities and become a hot topic in studying the laws of ship movement. At present, stop point analysis algorithms can be roughly divided into three categories: algorithms based on motion features, methods based on geometric features, and methods based on density clustering. The Stop / Move model is a classic method for extracting stop points based on trajectory motion features. It abstracts the trajectory into a sequence of Stop / Move objects and realizes the analysis of trajectory stops according to the motion state, but ignores the influence of speed parameters on trajectory stops. Many studies have expanded on the basis of the Stop / Move model and proposed stop point algorithms that can adapt to different scenarios by integrating motion parameters such as distance, time, speed, and topological relationships. In addition, methods based on geometric features are widely used in scenarios where motion trajectories have obvious spatial distribution forms. These methods are often used in combination with algorithms based on motion features. For example, from the perspectives of trajectory motion features and geometric patterns, the trajectory characteristics of individual and group refueling behaviors can be analyzed to establish a stop behavior detection model suitable for multi-source vehicle trajectories. Based on large-scale ship trajectory data, the alpha-shape algorithm is used to extract the channel boundary. The density clustering method starts from the temporal and spatial distribution characteristics of the trajectory, and clusters the dense areas of trajectory points to discover the temporal and spatial clustered stay areas. Common clustering methods include DBSCAN algorithm, minimum covering circle, stay index analysis, and kernel density estimation. These algorithms have achieved remarkable results in identifying trajectory stay points, providing a data basis for the subsequent identification of ship behavior characteristics. However, given the diversity of actual application scenarios and the complexity of trajectory data characteristics, especially the uniqueness of ship lock trajectories, the universality of existing algorithms is limited to a certain extent, and it is difficult to fully adapt to the complex and changing trajectory analysis needs.

[0102] At present, methods for detecting anomalies in ship trajectories include those based on statistical analysis, prediction models, and machine learning. For example, the kernel density estimation method is used to conduct statistical analysis and data mining on the behavior patterns of inland ferries, and then an abnormal ferry behavior detection algorithm based on position anomalies and speed anomalies is established. By using the ship behavior semantic cognitive computing model, the ship behavior with different motion characteristics and spatial topological characteristics in the port waters is modeled and expressed by integrating the ship's spatiotemporal trajectory data and navigation environment information. On the basis of extracting trajectory semantic information, a ship activity knowledge graph is constructed based on the SAM ship activity ontology model, which relatively completely expresses the ship's activities at sea from ship behavior to navigation process. In summary, ship trajectory anomaly detection presents a trend of diversified technology development and is gradually developing towards intelligence.

[0103] The above research can realize the extraction of key features of trajectories and the effective detection of abnormal trajectories, which is helpful to understand the laws of ship movement and improve the cognition of ship behavior. However, most of the current research focuses on ships in waters such as sea, nearshore, ports and ferries. There are relatively few analyses of ship trajectories in locks and specific waters upstream and downstream. Existing methods are difficult to directly apply to the analysis of ship behavior patterns when passing through locks. In addition, compared with ships in water environments such as sea, nearshore, inland rivers and ports, ships passing through locks have their own unique behavior patterns. Although many studies have focused on the analysis of ship behavior in various waters, no directly related research has been found to analyze the behavior patterns of ships passing through locks from massive trajectory data.

[0104] Therefore, the present invention focuses on the exploration of the behavior law of ships passing through the locks. Based on the in-depth analysis of the movement characteristics and navigation laws of ships passing through the locks, a method for analyzing the behavior law of ships passing through the locks based on trajectory data is proposed. First, a stop area crossing identification method taking into account the center of gravity drift is proposed, which realizes the accurate identification and time consumption analysis of each link of the ship passing through the lock; secondly, a stop feature extraction method integrating multi-feature constraints and grid division is proposed, and the trajectory stop and spatiotemporal distribution characteristics of ships passing through the locks are analyzed from two dimensions: individual stop behavior and group aggregation phenomenon; then, a multi-parameter coupled ship following behavior analysis method is proposed to analyze the temporal change law and coupling relationship between each following feature parameter.

[0105] Changzhou Ship Lock is located in the Xunjiang River section of the Xijiang River. It is the last cascade ship lock of the Xijiang River shipping trunk line and is known as the "throat" of the Xijiang River waterway. Changzhou Ship Lock has a total of four-line bidirectional single-stage ship locks, with two types of anchorages in the upstream and downstream for dry seasons and flood seasons. In order to facilitate scheduling, waiting areas are arranged near the upstream and downstream entrance areas, and mooring piers are arranged on both sides of the pilot channel.

[0106] This embodiment is based on the analysis of the ship passing records and ship movement trajectories of the Changzhou Ship Lock in June 2021. The specific data content is shown in Table 1. Among them, the ship movement trajectory is collected by the ship-borne Beidou terminal with a collection frequency of 30 seconds. Compared with AIS data, the data provided by the Beidou terminal has a higher collection frequency and position accuracy.

[0107] Table 1 Main data content

[0108]

[0109] Data preprocessing steps: First, convert the ship trajectory point data into trajectory line data; then, use the expected minimum duration of the trajectory to discard the shorter trajectory; at the same time, use the time gap between consecutive trajectory points to split the trajectory into multiple independent trajectories. Trajectory motion parameters include speed, direction and angle, among which speed, direction and their changes play an important role in stop extraction and semantic recognition. In order to facilitate trajectory analysis, auxiliary information such as speed, acceleration, direction, direction difference, time increment and distance increment are added to each trajectory according to the numerical changes of consecutive trajectory points.

[0110] like Figure 1 As shown, an embodiment of the present invention provides a method for analyzing ship lock passing behavior rules based on trajectory data, comprising the following steps:

[0111] S1. Analyze the ship passing through the lock using the ship trajectory and the ship stop area to obtain the time consumption analysis results of the ship passing through the lock, that is, use the stop area crossing identification method taking into account the center of gravity drift to analyze the ship passing through the lock, and obtain the time consumption rules of the ship in each stop area and the time consumption rules of each link of the ship passing through the lock;

[0112] Ships passing through a lock usually include five main links, namely, from the anchorage to the waiting area, from the waiting area to the pier, from the pier to the lock chamber, the lock operation phase, and the lock chamber exit phase. Accurately identifying the lock passing links is the prerequisite for discovering the time-consuming patterns of each link of the ship passing through the lock from massive ship trajectory data.

[0113] The process of identifying the ship passing through the lock is essentially to identify the crossing behavior of the ship at the key lock nodes, and cut the ship trajectory into multiple independent trajectory segments with the crossing points as the boundaries. Key lock nodes refer to specific stop areas such as waiting areas, berthing piers, and lock chambers. They are relatively fixed and require ships to follow certain navigation rules. In order to effectively identify these lock links, geo-fencing technology can be used to determine the crossing position and crossing time of ships entering or leaving the lock nodes. Among them, ray method, convex hull polygon, triangulation network and other methods are commonly used geo-fencing crossing identification algorithms. However, these algorithms have limitations when applied to the identification of ship lock nodes. First, ships are often accompanied by complex berthing and unberthing behaviors at the lock nodes, which makes the trajectory data prone to drift, especially at the boundaries of the docking area, the trajectory points may jump repeatedly, thus affecting the accuracy of the algorithm; secondly, the area where the ship chooses to dock is not fixed. For example, the docking position may be different when going up and down, and ships will use different waiting areas during the dry season and flood season. In addition, the docking area is directly connected to the waterway, which increases the risk of misjudgment. Therefore, the determination of geo-fence crossing is relatively complex, and the crossing behavior cannot be identified solely based on spatial relationship judgment.

[0114] To solve the above problems, a method for identifying crossing points in a stop area taking into account the center of gravity drift is proposed. The method includes three steps: selecting a set of ship trajectory points based on the ship stop area, identifying a set of crossing points based on the center of gravity point, and merging candidate crossing points based on multi-condition judgment. Figure 2 As shown, specifically:

[0115] S1-1, using the ship trajectory and the ship stop area to obtain a set of ship trajectory points, such as Figure 2 (a) shows the steps of selecting a set of ship trajectory points based on the ship stop area, where the ship trajectory is represented as a set of n continuous trajectory points sorted in time P = {p 1 ,p 2 ,…,p n}, the ship stop area is represented by a set of m polygons D = {d 1 ,d 2 ,…,d m}.

[0116] S1-1-1, performing a buffer operation according to the ship stay area to obtain a ship extended stay area;

[0117] In order to reduce the impact of trajectory point drift at the boundary of the ship's stay area and ensure the integrity of the trajectory at the boundary as much as possible, a buffer operation is performed on the ship's stay area to obtain the ship's extended stay area D' = {d′ 1 ,d' 2 ,…,d' m}, the calculation method is:

[0118] D'=ST_Buffer(D,dis)(1)

[0119] Among them, ST_Buffer(·) is the buffer operation, and dis is the buffer distance.

[0120] S1-1-2, performing a spatial intersection operation on the ship trajectory and the extended ship stay area to obtain a set of trajectory points to be analyzed in the ship stay area;

[0121] The ship trajectory P is spatially intersected with the ship extended stop area D', thereby obtaining a set of trajectory points to be analyzed that enter the ship stop area P' = {p' 1 ,p' 2 ,…,p' n}, the calculation method is:

[0122] P'=ST_Intersects(D',P)(2)

[0123] Among them, ST_Intersects(·) is a spatial intersection operation.

[0124] S1-1-3. Obtain a set of trajectory points to be analyzed in the ship's stop area as the ship's trajectory point set.

[0125] S1-2, obtaining the center of gravity of the ship trajectory according to the ship trajectory point set, Figure 2 (b) The steps of identifying the set of crossing points based on the center of gravity are shown, and the change of the center position intuitively shows the judgment process of the ship entry point and the ship departure point. The boundary points are judged by the method of converging within the center of gravity of the trajectory, which effectively eliminates the trajectory drift error caused by speed change;

[0126] S1-2-1. Obtaining an initial ship trajectory point p using the ship trajectory point set i ',p i ' is the i-th trajectory point to be analyzed in the ship's stop area;

[0127] S1-2-2, according to the initial ship trajectory point, obtaining the corresponding forward center of gravity point and the rearward center of gravity point as the center of gravity point of the ship trajectory;

[0128] Define p i The centroids of the k consecutive trajectory points before and after the point are the forward centroid O(i, k) and the backward centroid O'(i, k) respectively. Traverse the set of trajectory points to be analyzed P' and calculate p' in turn. i The corresponding forward center of gravity point O(i, k) and backward center of gravity point O'(i, k) are calculated as follows:

[0129] O(i,k)=ST_Centroid(p′ i-k+1 ,p′ i-k+2 ,…,p′ i ) (3)

[0130] O'(i,k)=ST_Centroid(p′ i ,p′ i+1 ,…,p′ i+k-1 ) (4)

[0131] Where, ST_Centroid(·) is the geometric centroid operation function, which is used to calculate the geometric centroid of the input trajectory point set, and k is a constant;

[0132] Traverse the set of trajectory points to be analyzed P' and calculate p' in sequence i The corresponding forward center of gravity point O(i,k) and backward center of gravity point O'(i,k).

[0133] S1-2-3, obtaining the ship crossing point set according to the center of gravity of the ship trajectory and the ship stop area;

[0134] S1-2-3-1. Obtain the spatial relationship between the forward center of gravity and the ship's stop area and the spatial relationship between the backward center of gravity and the ship's stop area according to the center of gravity of the ship's trajectory and the ship's stop area;

[0135] S1-2-3-2, according to the first judgment condition, determine whether the spatial relationship between the forward center of gravity point and the ship stay area changes from not included to included. If so, obtain the forward center of gravity point as the ship entry point and execute S1-2-3-3. Otherwise, return to S1-2-1. Specifically:

[0136] When the spatial relationship between the forward centroid and the stop area changes from "Without" to "Within", the trajectory point p' a Defined as the ship entry point, that is, the forward center point is taken as the ship entry point. Conversely, return to S1-2-1 to traverse the set of trajectory points to be analyzed P', and calculate p' in turn i The corresponding forward center of gravity point O(i,k) is obtained, and the spatial relationship between the forward center of gravity point and the stop area is determined, where a is a constant;

[0137] S1-2-3-3, according to the second judgment condition, determine whether the spatial relationship between the rearward center of gravity point and the ship stay area changes from inclusion to non-inclusion. If so, obtain the rearward center of gravity point as the ship departure point and execute S1-2-3-4. Otherwise, return to S1-2-1. Specifically:

[0138] When the spatial relationship between the backward centroid and the stop area changes from within to without, the trajectory point p' b Defined as the departure point of the ship, that is, the rearward center of gravity is taken as the departure point of the ship. Otherwise, return to S1-2-1 to traverse the set of trajectory points to be analyzed P', and calculate p' in sequence i The corresponding backward centroid point O'(i,k), and determine the spatial relationship between the backward centroid point and the stay area, where b is a constant;

[0139] The calculation formula of the first judgment condition is as follows:

[0140]

[0141] The calculation formula of the second judgment condition is as follows:

[0142]

[0143] Where ST_Within(·) is a spatial inclusion operation function used to determine whether the center point is completely inside the stop area, O(a,k) is the ath forward centroid point, O(a-1,k) is the a-1th forward centroid point, and d jis the j-th ship stop area, O'(b,k) is the b-th rearward center of gravity, O'(b+1,k) is the b+1-th rearward center of gravity, True is correct, False is wrong, a, b, j are all constants.

[0144] S1-2-3-4, using the ship entry point and the ship exit point as the ship crossing point set;

[0145] S1-3, obtaining a set of candidate critical points according to the set of ship crossing points, Figure 2 (c) shows the steps of merging candidate crossing points based on multi-condition judgment, and the candidate critical point set is obtained by calculating the ship entry point and the ship departure point.

[0146] S = {p' a1 ,p' b1 ,p' a2 ,p' b2 ,…,p' an ,p' bn};

[0147] S1-4. Filter and screen the candidate critical point set in sequence using the ship stay information to obtain the ship passing process node set, wherein the ship stay information is the ship stay duration and the average speed of the ship stay period, specifically:

[0148] Since the ship needs to repeatedly adjust its posture when berthing at the boundary of the stop area to accurately dock at the target position, the trajectory points may jump back and forth at the boundary, resulting in redundant crossing points. In order to solve this problem, the candidate critical point set is filtered and screened in turn according to the ship's stay time and the average speed of the ship's stay period, and finally the accurate node set of the lock passing process S' = {p' a ,p' b}.

[0149] S1-5, analyzing the ship passing through the lock according to the ship passing through lock process node set to obtain the time consumption pattern of the ship in the stay area and the time consumption pattern of the ship passing through the lock as the time consumption analysis result of the ship passing through the lock, specifically:

[0150] The ship passing through the lock process is analyzed according to the node set of the ship passing through the lock process, and the time consumption rules of the ship in each stop area and the time consumption rules of each link of the ship passing through the lock are obtained, such as Figure 3-4 As shown, the ships with high grey values ​​are upstream ships, and the ships with low grey values ​​are downstream ships.

[0151] Depend on Figure 3It can be seen that the ship stopped in the stop area before passing through the lock, and entered a free driving state after passing through the lock. The time spent in the stop areas is from most to least in the order of berthing piers (A1, B1), lock chambers (C), waiting areas (A2, B2), and anchorages (C1, C2), and the stop time is mainly concentrated in berthing piers and lock chambers. The ship's stay time at the berthing pier is the longest, the data changes the most, and the upstream time is significantly higher than the downstream time. In contrast, the overall time spent by the ship in the lock chamber does not change much, and the downstream time is slightly higher than the upstream time.

[0152] Depend on Figure 4 It can be seen that in the time distribution of upstream ships, S5, S6, and S7 occupy the main part, and the average time consumption of these three links is similar. Although the average time consumption is close, it is worth noting that there are significant differences in the channel mileage of these three links. The channel mileage of S7 is much higher than that of S5 and S6. Therefore, S7 cannot be simply regarded as the main link affecting the time consumption of ships passing through the lock. For downstream ships, S2 and S3 links occupy the main part, which is lower than that of upstream ships as a whole. In addition, the time fluctuation of upstream ships in the S7 link is significantly higher than that of downstream ships in the S1 link.

[0153] Based on the above results, it can be concluded that the time spent on ship passing the lock is mainly from the ship entering the waiting area to passing through the lock, especially from the berthing pier to the lock chamber, which is short in distance and takes a long time, and is the link that takes the most time for ships to pass through the lock. If the influence of the waterway mileage factor is taken into account, the influence of each link on the ship passing through the lock is ranked in order: from the berthing pier to the lock chamber, from the waiting area to the berthing pier, the lock chamber operation, from the downstream anchorage to the waiting area, and from the upstream anchorage to the waiting area.

[0154] S2. Performing feature analysis based on the fusion of multi-parameter constraints and grid division on the ship trajectory and the channel centerline to obtain the ship stay feature analysis result, that is, analyzing the ship stay features based on the ship trajectory and the ship channel using the stay feature extraction method that combines multi-parameter constraints and grid division to obtain the distribution of ship stay points and the thematic map of ship stay index;

[0155] The speed of ships passing through the lock changes steadily and their motion trajectories are relatively fixed. The stop feature analysis methods based on motion features and geometric features are difficult to comprehensively and accurately summarize the stop behaviors of ships passing through the lock. The density clustering-based method performs well in extracting stop features of freely moving vehicles or ships in an open environment, but its advantages are not obvious in the task of extracting features that are clustered along the motion trajectory. Therefore, in view of the data characteristics of the trajectory of ships passing through the lock, this embodiment proposes a stop feature extraction method that integrates multi-parameter constraints and grid division. This method is based on the motion parameters and position distribution of the ship, and by integrating multiple parameter constraints, it analyzes the individual stop behaviors and group aggregation phenomena of ships passing through the lock, thereby realizing the extraction of ship stop features. The specific process is as follows: Figure 5 As shown, specifically:

[0156] S2-1. The ship channel is obtained by segmenting the channel centerline, that is, the channel centerline is divided into five parts according to the lock passing process: upstream channel, upstream pilot channel, lock chamber, downstream pilot channel and downstream channel.

[0157] S2-2, using the ship channel and the ship trajectory to perform trajectory point channel mapping to obtain the ship trajectory to be processed, that is, establishing a matching relationship between the ship trajectory point and the ship channel according to the ship channel and the ship trajectory point, mapping the ship trajectory point to the nearest segment, the mapping point is the vertical projection point of the ship trajectory point to the center line of the channel, and then obtaining the ship trajectory to be processed, wherein the ship trajectory is composed of a number of ship trajectory points.

[0158] S2-3, assigning attributes to the ship trajectory points of the ship trajectory to obtain the ship trajectory point attributes, that is, assigning new attributes to the ship trajectory points: ① segment number, that is, the segment identifier corresponding to the trajectory point; ② mileage value, that is, the actual distance from the mapping point to the starting point of the segment; ③ offset, that is, the straight-line distance between the trajectory point and its mapping point.

[0159] S2-4, performing data cleaning and trajectory smoothing processing in sequence according to the to-be-processed ship trajectory and the attributes of the ship trajectory points to obtain the target ship trajectory, specifically:

[0160] Data cleaning: ① When the speed exceeds the preset speed threshold, a data cleaning method based on speed anomaly is used to eliminate the spikes in the trajectory; ② When the offset exceeds the preset distance threshold, a data cleaning method based on distance anomaly is used to eliminate the drift points in the trajectory;

[0161] Trajectory smoothing: The trajectories are filtered and smoothed using a Kalman filter with a constant velocity model and the missing positions in the trajectory are fitted.

[0162] S2-5, determining whether the target ship trajectory is a single ship trajectory, if so, executing S2-6, otherwise, directly executing S2-7;

[0163] S2-6, according to the target ship trajectory fused with multi-parameter constraints to obtain ship stop features, that is, individual stop feature extraction, for a single ship trajectory, based on the Stop / Move model, introduces a multi-parameter constraint trajectory stop point extraction method, by setting the ship's running distance, duration and speed threshold, detect the target ship trajectory in the specified area for a certain duration of static or low-speed trajectory points, and obtain a set of candidate stop points; then, the stop point and the stop area are judged for spatial intersection, and the abnormal ship stop point and the normal ship stop point are identified. Specifically:

[0164] S2-6-1. Set the ship running distance threshold, ship running time threshold and ship running speed threshold as multi-parameter constraint conditions, namely, distance constraint, time constraint and speed constraint;

[0165] S2-6-2, based on the Stop / Move model, using the multi-parameter constraint condition to detect the target ship trajectory to obtain the stop center point, the stop duration and the stop point set, and then obtain the candidate stop point set;

[0166] S2-6-3. Perform spatial intersection operation on the candidate stay point set and the ship stay area to obtain normal ship stay points and abnormal ship stay points as the ship stay features.

[0167] S2-7. The ship stop feature is obtained by network division according to the target ship trajectory and the channel centerline, that is, group aggregation feature extraction. For the trajectory point set, the LINE-STING algorithm is proposed for trajectory stop feature extraction based on the STING (Statistical Information Grid) algorithm idea. The algorithm calculates statistical information on each linear grid unit and performs stop feature analysis based on this information. Specifically:

[0168] S2-7-1, grid division is performed according to the center line of the waterway to obtain a basic unit for density calculation, that is, the center line of the waterway is divided into n equally spaced grid units as the basic unit for density calculation;

[0169] S2-7-2, obtaining a set of ship trajectory points contained in the grid unit according to the ship trajectory mapping points of the target ship trajectory and the grid unit of the density calculation, that is, determining a set of ship trajectory points contained in the grid unit according to the spatial relationship between the ship trajectory mapping points of the target ship trajectory and the grid unit;

[0170] S2-7-3, obtaining a density weight parameter of the grid cell according to the set of ship trajectory points contained in the grid cell, that is, selecting the number of trajectory points or the average speed as the density weight parameter of the grid cell according to the set of ship trajectory points contained in the grid cell;

[0171] S2-7-4. Calculate the stay index of the grid unit using the density weight parameter of the grid unit, the calculation method is:

[0172]

[0173] Where σ(m) represents the residence index of the mth grid unit, d is the length of the grid unit, and w iis the density weight parameter of the ith grid cell, k is a constant, and k consecutive grid cells are taken from the mth grid cell forward or backward to participate in the calculation of the residence index. Increasing the k value can reduce the data error, but it may also reduce the data accuracy.

[0174] S2-7-5, obtaining a density threshold according to the residence index of the grid unit, that is, setting a density threshold according to the residence index of the grid unit, and different weight parameters correspond to different density thresholds;

[0175] S2-7-6, determining whether the stay index of the grid unit is lower than the density threshold, if so, obtaining the grid unit as a ship smooth point, otherwise, obtaining the grid unit as a ship congestion point;

[0176] S2-7-7, obtaining a thematic map of the stay index and a thematic map of the patency index by using a graded coloring method according to the stay index of the grid unit, that is, drawing a thematic map of the stay index and a thematic map of the patency index by using a graded coloring method based on the stay index of the grid unit, so as to intuitively display the stay characteristics of the ship;

[0177] S2-7-8, taking the ship stop features as the ship stop points, the ship congestion points, the stop index thematic map and the smoothness index thematic map.

[0178] S2-8, according to the ship stay characteristics, characteristic analysis is performed to obtain the ship stay point distribution information and the ship top flow index information as the ship stay characteristic analysis result, that is, the propagation stay characteristics are analyzed to obtain the ship stay point distribution and the ship stay index thematic map, such as Figure 6-7 shown.

[0179] Figure 6 The distribution of stop points on the waterway and their stop time are intuitively displayed. Figure 6 In (a)-(d), the triangle marks represent abnormal stay points, and the circle marks represent normal stay points. It can be seen that the abnormal stay points are mainly distributed in the surrounding areas of the upstream and downstream anchorages (① and ③) and the adjacent area of ​​the downstream waiting area (②). Most of these stay points are short-term stays, with a stay time of about half an hour. The stay points with a stay time of more than 1 hour are mainly concentrated near the downstream waiting area (②).

[0180] Figure 7 (a) is a thematic map of the stay index calculated and drawn with the number of trajectory points as the weight parameter. The results show that the trajectory points show obvious clustering in the stay areas such as the lock chamber, pier, waiting area, and anchorage area. This result is highly consistent with the analysis results of the stay point distribution ( Figure 6), verifying the effectiveness of the method. Specifically, the ship's stay index at the downstream pier is slightly higher than that in the lock chamber and upstream pier, and the stay index in the downstream waiting area is slightly higher than that in the upstream waiting area, which is consistent with the time consumption analysis results of the stay area ( Figure 4 ). The time consumed in the upstream is generally higher than that consumed in the downstream. The ship stay index at the downstream pier is high and the stay time is long, which greatly affects the efficiency of ship passing through the lock.

[0181] Figure 7 (b) and Figure 7 (c) is a congestion index thematic map calculated and drawn with average speed as the weight parameter, where Figure 7 (b) is the result calculated based on all trajectory points. Figure 7 (c) is a congestion index thematic map calculated and drawn based on the trajectory points after filtering out the stationary points and low-speed points. The results show that the congestion index of the lock and its surrounding areas is the highest, followed by the upstream anchorage. In the waterway without a mooring area, there are many congestions of varying degrees between the downstream pilot channel and the downstream anchorage. In contrast, the ship speed change in the upstream pilot channel is relatively stable. It can be seen that the downstream waterway is long, the ship's navigation speed changes unevenly, and there are abnormal stop behaviors. Strengthening the downstream ship navigation management is conducive to improving the efficiency of ship passing through the lock. Compared with the congestion index thematic map containing stationary points ( Figure 7 (b)), the congestion index thematic map excluding stationary points ( Figure 7 (c)) can more accurately reflect the movement law of ships when they are not waiting for dispatch. For ships in motion, the upstream and downstream channels are generally smooth, with an average speed of more than 8km / h, but the congestion index of the upstream is higher than that of the downstream. The congestion level of the lock and the pilot channel is relatively high, with an average speed of less than 5km / h, and the average speed near the lock chamber is as low as about 3km / h. According to the relevant provisions of the "General Design Specifications for Locks" (JTJ 305-2001), when the lock is operated in both directions, the reference average speed of a motorized single ship entering the lock is 3.6m / s, and the reference average speed of exiting the lock is 5.04m / s. Therefore, in the process of ships entering and exiting the lock, while ensuring safety, increasing the average speed of the ship can effectively shorten the ship's lock passing time and further improve the ship's passing capacity.

[0182] S3, performing a ship following behavior analysis based on a multi-parameter coupling method according to the ship trajectory data and the ship passing the lock to obtain a ship following behavior analysis result, that is, using a multi-parameter coupling ship following behavior analysis method according to the ship trajectory data and the ship passing the lock to analyze the ship following behavior, and obtain a time series analysis result of the following characteristic parameters and a coupling relationship analysis result of the following characteristic parameters;

[0183] The following model studies the dynamic process of the change of motion state between the leading object and the following object, while the ship following model essentially studies the change law of the distance, speed and acceleration between the leading ship and the following ship under certain constraints and their interaction mechanism. It is worth noting that there are significant differences between the ship following model and the vehicle following model. The vehicle following model usually assumes that the vehicle is traveling on a one-way road with restricted overtaking. However, ships need to go through multiple dispatches to pass through the lock. Although ships are also traveling in a one-way direction with restricted overtaking in the waterway, the following order of ships may change at the lock-passing nodes such as the waiting area, berthing piers, and lock chambers. Figure 8 As shown in the figure, the scheduling gear changes the first-in-first-out order of ships, making the ship following behavior more complicated. In addition, compared with vehicles, ships have longer starting distances and braking distances, and usually maintain low speeds during ship passage, with relatively small acceleration and basically unchanged following distance, which is in sharp contrast to the characteristics of vehicles with short braking distances, frequent acceleration fluctuations, and large speed changes.

[0184] Therefore, this embodiment takes lock-passing ships as the research object and proposes a multi-parameter coupled ship following behavior analysis method. This method reveals the characteristics of ship following behavior through locks by studying the change law and coupling relationship of ship following characteristic parameters such as following distance, speed, acceleration, etc. The specific process is as follows: Figure 8 As shown, specifically:

[0185] S3-1, using the ship passing record to obtain the ship passing record of the lock, wherein the ship passing time in the ship passing record is based on the gate closing time, and the gate closing time of the ships at the same lock is the same, based on which the ship passing record of each lock can be extracted;

[0186] S3-2, according to the ship passing record of the lock and the ship trajectory data, a corresponding ship trajectory line set is obtained as the ship trajectory of the lock, that is, according to the ship list of the same lock, a set of ship motion trajectory points is retrieved from the ship trajectory data, and a corresponding ship trajectory line set is generated to obtain the ship trajectory of the lock;

[0187] S3-3, performing preprocessing according to the ship trajectory of the lock to obtain corresponding ship trajectory mapping points, the preprocessing including the steps of trajectory point channel mapping, data cleaning and trajectory smoothing;

[0188] S3-4, linear interpolation and mean resampling are performed on the trajectory attributes of the ship trajectory mapping points to obtain a resampled time series and the ship trajectory of the same lock, that is, a unified time series, and linear interpolation and mean resampling are performed on the trajectory attributes (such as point speed and acceleration) of the ship trajectory mapping points to ensure the consistency of data acquisition frequency;

[0189] S3-5. Performing a time series analysis of the following characteristic parameters according to the ship trajectory mapping points to obtain the time series analysis results of the following characteristic parameters of the ships at the same lock, i.e., the following characteristic parameter time series analysis, specifically includes: ① Based on the spatial position of the trajectory mapping points, comparing and analyzing the position distribution and change trend of the ships at the same lock. ② Combining the speed and acceleration of the mapped points, comparing and analyzing the change trend of the speed and acceleration of the ships at the same lock;

[0190] S3-6, performing a ship following distance time series analysis based on the resampled time series and the trajectory of the ship at the same lock to obtain a ship following distance time series analysis result, i.e., a ship following distance time series analysis, including: in view of the change in the following order of the ship after passing the lock node, retrieving the ship position distribution information within the sampling interval based on the resampled time series; then, reordering the ships in order of precedence along the navigation direction of the ships to correct the ship following order; finally, calculating the straight-line distance between the front and rear ships, and comparing and analyzing the change trend of the following distance of the ships at the same lock;

[0191] S3-7, performing a coupling relationship analysis of the following characteristic parameters based on the time series analysis result of the following characteristic parameters and the time series analysis result of the ship following distance, obtaining a following distance-speed coupling relationship and a following distance-acceleration coupling relationship as the following characteristic parameter coupling relationship analysis result, i.e., a following characteristic parameter coupling relationship analysis, including: based on the results of the following characteristic parameter time series analysis and the ship following distance time series analysis of the ships at the same lock, analyzing the coupling relationship between the ship following characteristic parameters and the ship following distance in a unified time series, and fitting the following distance-speed coupling relationship function under different headings and different routes;

[0192] S3-8, obtaining the time series analysis result of the following characteristic parameter and the coupling relationship analysis result of the following characteristic parameter as the ship following behavior analysis result, such as Figure 9-10 shown.

[0193] Fig. 9 (a)-(d) show the comparative analysis results of different following characteristic parameters in the same time series. The order of ship entry is ①②③④⑤. After the waiting area scheduling, berthing pier scheduling, and ship exit, the following order has changed, and the exit order has become ②③⑤①④. This result shows that the ship gear will affect the analysis of the driving order. The waiting time of the ship at the berthing pier is the longest, which is much longer than the normal waiting time. When the ship enters the lock, the ship travels at a low speed, but the speed and acceleration at this time fluctuate greatly, and the channel mileage from the berthing pier to the lock chamber is short, which means that when the front ship has entered the lock chamber, the rear ship has just started. Because the ship did not start in time in the waiting area, the process of entering the lock is not orderly and compact, resulting in a short mileage and long time in the berthing pier scheduling stage, which reduces the efficiency of ship passing the lock.

[0194] Compared with the scheduling in the waiting area, the speed and acceleration fluctuations during the lock exit scheduling stage are more stable. Especially when the ship enters the berthing pier from the waiting area, the speed drops rapidly. By continuously adjusting the speed and position, the ship finally docks smoothly at the berthing position. This process takes a long time. Although the following distance of ships basically remains within the normal range, the following distances of some ships greatly exceed the normal range, which greatly affects the lock passage efficiency of ships. Therefore, strengthening the organization and scheduling of ships in the waiting area and at the berthing pier to ensure more orderly and compact ship passage through the lock is crucial for further tapping the lock passage capacity.

[0195] Fig.10 It shows the coupling relationship between the following characteristic parameters and the following distance of ships with different headings and on different routes. The horizontal axis is the following distance, and the vertical axis is the speed or acceleration. The smaller the gray value in the figure, the more concentrated the data distribution, and the stronger the coupling between variables. From the fitted coupling relationship function, it can be seen that the relationship between the following distance and the speed conforms to a power function distribution with a power exponent between 0 and 1, that is, f(x) = x a (0 < a < 1), and the size of the power exponent fluctuates due to different navigation directions and lock routes. Under the same route condition, the power exponent value for upstream is less than that for downstream, which means that the following distance of upstream ships is larger or the traveling speed is slower. Under the condition of the same navigation direction, the power exponent values for the first and second lines are greater than those for the third and fourth lines, which means that the following distance of ships on the first and second lines is smaller or the traveling speed is faster.

[0196] Fig.10 (a) shows that when the speed exceeds 3 m / s, the ship's traveling shows significant following characteristics. As the following distance increases, the speed also increases accordingly, but the growth rate of the speed gradually decreases and tends to the normal traveling speed. Based on the coupling relationship function and the data density distribution, the reference speed of ship traveling can be estimated, providing a reference for tapping the lock passage capacity. Fig.10 In (a), the situation of larger following distances is common, especially those data pairs with large following distances and low traveling speeds have a greater impact on the lock passage efficiency of ships.

[0197] Fig.10 (b) shows the coupling relationship between the following distance and the acceleration, presenting a conical divergence distribution centered on the origin. Under the same route condition, the conical shape for upstream is sharper than that for downstream, which means that the acceleration of upstream ships is more stable. Under the condition of the same navigation direction, the conical shape for the first and second lines is sharper than that for the third and fourth lines, which means that the acceleration of ships on the first and second lines is more stable.

[0198] By analyzing the coupling relationship between the speed, acceleration, and following distance, the abnormal data distribution in the ship lock following behavior can be intuitively shown. By fitting the power function, the coupling relationship between ship characteristic parameters can be analyzed, and combined with the data density distribution, a quantitative evaluation of the coupling degree can be achieved.

[0199] S4. Obtain the time consumption analysis result of the ship passing through the lock, the ship stay characteristic analysis result and the ship following behavior analysis result as the ship passing through the lock behavior regularity analysis result.

[0200] This example is based on the analysis of ship passing records and ship movement trajectories at the Changzhou Ship Lock in June 2021. The data comes from the Guangxi Xijiang Ship Lock Joint Dispatching Center. After experiments and analysis at the Changzhou Ship Lock, the effectiveness of this method was verified, and the following conclusions were drawn based on the experimental results:

[0201] (1) The time taken to pass the lock from the waiting area to the lock chamber accounts for the majority of the time taken by ships. Among them, the longest time is spent at the mooring pier. Whether it is the lock passing link or the stop area, the time taken to go up is generally longer than the time taken to go down. In order of the degree of influence on the efficiency of ship passing the lock, the links of the Changzhou Ship Lock are: from the mooring pier to the lock chamber, from the waiting area to the mooring pier, the lock chamber operation, from the downstream anchorage to the waiting area, and from the upstream anchorage to the waiting area.

[0202] (2) The trajectory stop point extraction method with multi-parameter constraints can quickly identify areas with abnormal motion characteristics during the ship's passage through the lock, providing a more targeted basis for lock scheduling and management. The LINE-STING trajectory stop feature extraction method can analyze the spatiotemporal aggregation of ship trajectories through stop index analysis and congestion index analysis, and intuitively display the operating status of the lock and surrounding waterways. Using the method proposed in this embodiment, it was found that the capacity of the waiting anchorage originally designed for the Changzhou Lock is seriously insufficient, and it is recommended to strengthen the reconstruction and expansion of the anchorage area and long-term planning; there are abnormal stop behaviors in the downstream waterway, and navigation management needs to be further strengthened; although the relative congestion level of the lock and the pilot channel is relatively high, the average speed of ships in and out of the lock is still lower than the design reference speed, which shows that the passing capacity of the Changzhou Lock still has room for improvement.

[0203] (3) Time series analysis of ship following characteristic parameters. By analyzing the changing trends of characteristic parameters such as position, speed, acceleration, and following distance of ships at the same lock, it is helpful to efficiently and intuitively analyze abnormal conditions that affect the efficiency of ship lock passage, such as untimely ship start-up, slow sailing speed, and disorderly entry and exit. The coupling relationship between the ship's following distance and speed shows a power function distribution characteristic with a power index between 0 and 1, and the power index fluctuates due to different sailing directions and lock lines. Combining the coupling relationship curve and data density distribution, the abnormal data distribution in the ship's following behavior can be intuitively displayed, which can provide important data support for improving the ship lock's passing capacity.

[0204] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0205] The present invention is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0206] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0207] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0208] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A ship passing lock behavior analysis method based on trajectory data, characterized in that: The specific steps include: S1. Analyze the ship passing through the lock using the ship trajectory and the ship stop area to obtain the time consumption analysis result of the ship passing through the lock; S2, performing feature analysis based on the ship trajectory and the channel centerline, integrating multi-parameter constraints and grid division, to obtain a ship stay feature analysis result; S3, performing a multi-parameter coupled ship following behavior analysis based on the ship trajectory data and the ship passing record to obtain a ship following behavior analysis result; S4. Obtain the time consumption analysis result of the ship passing through the lock, the ship stay characteristic analysis result and the ship following behavior analysis result as the ship passing through the lock behavior regularity analysis result.

2. The method for analyzing ship lock-passing behavior rules based on trajectory data drive according to claim 1 is characterized in that: The ship's track and the ship's stop area are used to analyze the ship's passage through the lock to obtain the time consumption analysis results of the ship's passage through the lock, including: S1-1, obtaining a set of ship trajectory points using the ship trajectory and the ship stop area; S1-2, obtaining a ship crossing point set according to the ship trajectory point set; S1-3, obtaining a set of candidate critical points according to the set of ship crossing points; S1-4, using the ship stop information to filter and screen the candidate critical point set in turn to obtain the ship lock process node set; S1-5, analyzing the ship passing through the lock according to the ship passing through lock process node set to obtain the time consumption pattern of the ship in the stay area and the time consumption pattern of the ship passing through the lock as the time consumption analysis result of the ship passing through the lock; The ship stop information includes the ship stop duration and the average speed of the ship during the stop period.

3. The ship passing lock behavior law analysis method based on trajectory data drive according to claim 2 is characterized in that: Acquiring a set of ship trajectory points using the ship trajectory and the ship stop area includes: S1-1-1, performing a buffer operation according to the ship stay area to obtain a ship extended stay area; S1-1-2, performing a spatial intersection operation on the ship trajectory and the extended ship stay area to obtain a set of trajectory points to be analyzed in the ship stay area; S1-1-3, obtaining a set of trajectory points to be analyzed in the ship's stop area as the ship's trajectory point set; The calculation formula for the extended ship stay area is as follows: D'=ST_Buffer(D,dis) The calculation formula of the ship trajectory point set is as follows: P'=ST_Intersects(D',P) Where D' is the extended stop area of ​​the ship, ST_Buffer(·) is the buffer operation, D is the ship stop area, dis is the buffer distance, P' is the set of trajectory points to be analyzed in the ship stop area, ST_Intersects(·) is the spatial intersection operation, and P is the ship trajectory.

4. The ship passing lock behavior law analysis method based on trajectory data drive according to claim 2 is characterized in that: Acquiring a ship crossing point set according to the ship trajectory point set includes: S1-2-1, using the ship trajectory point set to obtain an initial ship trajectory point; S1-2-2, according to the initial ship trajectory point, obtaining the corresponding forward center of gravity point and the rearward center of gravity point as the center of gravity point of the ship trajectory; S1-2-3, obtaining the ship crossing point set according to the center of gravity of the ship trajectory and the ship stop area; The forward center of gravity is calculated as follows: O(i,k)=ST_Centroid(p i ' -k+1 ,p i ' -k+2 ,…,p i ') The calculation formula of the rearward center of gravity is as follows: O'(i,k)=ST_Centroid(p i ',p i ' +1 ,…,p i ' +k-1 ) Where O(i,k) is the forward centroid, ST_Centroid(·) is the geometric centroid calculation function, and p i ' is the i-th trajectory point to be analyzed in the ship's stop area, k is a constant, and O'(i,k) is the backward center of gravity.

5. The method for analyzing ship passing lock behavior rules based on trajectory data drive according to claim 4 is characterized in that: Acquiring the ship crossing point set according to the center of gravity of the ship trajectory and the ship stop area includes: S1-2-3-1. Obtain the spatial relationship between the forward center of gravity and the ship's stop area and the spatial relationship between the backward center of gravity and the ship's stop area according to the center of gravity of the ship's trajectory and the ship's stop area; S1-2-3-2, judging whether the spatial relationship between the forward center of gravity point and the ship stop area changes from not included to included according to the first judgment condition, if so, obtaining the forward center of gravity point as the ship entry point, and executing S1-2-3-3, otherwise, returning to S1-2-1; S1-2-3-3, judging whether the spatial relationship between the rearward center of gravity point and the ship stay area changes from inclusion to non-inclusion according to the second judgment condition, if so, obtaining the rearward center of gravity point as the ship departure point, and executing S1-2-3-4, otherwise, returning to S1-2-1; S1-2-3-4, using the ship entry point and the ship exit point as the ship crossing point set; The calculation formula of the first judgment condition is as follows: The calculation formula of the second judgment condition is as follows: Where ST_Within(·) is the spatial inclusion operation function, O(a,k) is the ath forward centroid point, O(a-1,k) is the a-1th forward centroid point, and d j is the j-th ship stop area, O'(b,k) is the b-th rearward center of gravity, O'(b+1,k) is the b+1-th rearward center of gravity, True is correct, False is wrong, a, b, j are all constants.

6. The method for analyzing ship lock-passing behavior rules based on trajectory data drive according to claim 1 is characterized in that: The ship's stay characteristics are obtained by integrating multi-parameter constraints and grid division based on the ship's trajectory and channel centerline. The analysis results include: S2-1, obtaining a ship channel by segmenting the channel centerline; S2-2, using the ship channel and the ship track to perform track point channel mapping to obtain the ship track to be processed; S2-3, assigning attributes to the ship track points of the ship track to obtain ship track point attributes; S2-4, performing data cleaning and trajectory smoothing processing in sequence according to the to-be-processed ship trajectory and the attributes of the ship trajectory points to obtain the target ship trajectory; S2-5, determining whether the target ship trajectory is a single ship trajectory, if so, executing S2-6, otherwise, directly executing S2-7; S2-6, acquiring ship stay characteristics according to the target ship trajectory by fusing multi-parameter constraints; S2-7, obtaining the ship stay characteristics by using network division according to the target ship trajectory and the channel centerline; S2-8, performing feature analysis according to the ship stay feature to obtain ship stay point distribution information and ship top flow index information as the ship stay feature analysis result; Wherein, the ship trajectory includes a plurality of ship trajectory points.

7. The method for analyzing ship passing lock behavior based on trajectory data drive according to claim 6 is characterized in that: According to the target ship trajectory, fusion of multi-parameter constraints to obtain ship stay characteristics includes: S2-6-1. Set the ship running distance threshold, ship running time threshold and ship running speed threshold as multi-parameter constraint conditions; S2-6-2. Based on the Stop / Move model, the target ship trajectory is detected using the multi-parameter constraint condition to obtain a set of candidate stop points; S2-6-3. Perform spatial intersection operation on the candidate stay point set and the ship stay area to obtain normal ship stay points and abnormal ship stay points as the ship stay features.

8. The method for analyzing ship passing lock behavior rules based on trajectory data drive according to claim 6 is characterized in that: Acquiring the ship stay feature by using network division according to the target ship trajectory and the channel centerline includes: S2-7-1, grid division is performed according to the center line of the waterway to obtain a basic unit for density calculation; S2-7-2, obtaining a set of ship trajectory points contained in the grid cell according to the ship trajectory mapping points of the target ship trajectory and the grid cell of the density calculation; S2-7-3, obtaining a density weight parameter of the grid cell according to the set of ship trajectory points contained in the grid cell; S2-7-4, calculating the stay index of the grid unit using the density weight parameter of the grid unit; S2-7-5, obtaining a density threshold according to the residence index of the grid unit; S2-7-6, determining whether the stay index of the grid unit is lower than the density threshold, if so, obtaining the grid unit as a ship smooth point, otherwise, obtaining the grid unit as a ship congestion point; S2-7-7, obtaining a thematic map of the retention index and a thematic map of the patency index according to the retention index of the grid unit by using a graded coloring method; S2-7-8, using the ship smooth spot, the ship congestion spot, the stop index thematic map and the smooth index thematic map as the ship stop feature; The basic unit of density calculation is a plurality of equally spaced grid units.

9. The method for analyzing ship lock-passing behavior rules based on trajectory data drive according to claim 8, characterized in that: The calculation formula of the stay index of the grid unit is as follows: Where σ(m) represents the residence index of the mth grid unit, d is the length of the grid unit, and w i is the density weight parameter of the i-th grid unit, and k is a constant.

10. The method for analyzing ship passing lock behavior rules based on trajectory data drive according to claim 1, characterized in that: The ship following behavior analysis is carried out based on the ship trajectory data and the ship passing record through multi-parameter coupling to obtain the ship following behavior analysis results including: S3-1, using the ship lock passing record to obtain the lock passing record; S3-2, obtaining a corresponding ship trajectory line set as the ship trajectory at the lock according to the ship passing record at the lock and the ship trajectory data; S3-3, preprocessing is performed according to the ship trajectory of the lock to obtain corresponding ship trajectory mapping points; S3-4, performing linear interpolation and mean resampling on the trajectory attributes of the ship trajectory mapping points to obtain a resampled time series and the trajectory of the ship at the same lock; S3-5, performing a time series analysis of the following characteristic parameters according to the ship trajectory mapping points to obtain a time series analysis result of the following characteristic parameters of the ships in the same lock; S3-6, performing a ship following distance time series analysis based on the resampled time series and the trajectory of the ship at the same lock to obtain a ship following distance time series analysis result; S3-7, performing a coupling relationship analysis of the following characteristic parameters based on the time series analysis result of the following characteristic parameters and the time series analysis result of the ship following distance, to obtain a following distance-speed coupling relationship and a following distance-acceleration coupling relationship as the following characteristic parameter coupling relationship analysis result; S3-8. Obtain the time series analysis result of the car-following characteristic parameter and the coupling relationship analysis result of the car-following characteristic parameter as the ship's car-following behavior analysis result.

Images