Water area period collision danger identification method based on dynamic time warping

Abstract

The invention relates to the technical field of ship traffic safety and risk assessment, in particular to a water area time period collision risk identification method and device based on dynamic time warping. Firstly, a mapping relation between collision avoidance parameters and collision dangers is established from the perspective of geometric encounter; meanwhile, the traffic environment pressure sensed by the ship is determined from the angles of traffic intensity and compactness, and then the traffic environment pressure is combined with the geometric encounter collision danger to establish a ship collision danger time sequence. And finally, mining the features shown by the time sequences through a dynamic time warping algorithm, identifying the similarity between the features and the danger sequences, characterizing the collision danger features through the accumulated distance between the time sequences, and calculating the collision danger features on the basis of considering the contribution of each ship to the collision danger of the water area. And the collision danger in the water area time period is quantified. According to the method, the future collision danger of the water area can be predicted more accurately, so that preventive measures are deployed in time, and the marine traffic safety is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of ship traffic safety and risk assessment, and in particular to a method for identifying collision risks in water areas based on dynamic time regularization. Background Art

[0002] With the rapid development of the global shipping industry, the prevention and management of ship collision accidents has become an important issue in the field of maritime safety. Ship collisions not only cause casualties and property losses, but may also cause serious damage to the marine environment and affect international trade and economic activities. Although the increase in ship traffic capacity has made a huge contribution to global economic growth, it cannot be ignored that it has posed more threats to maritime traffic safety and has also posed challenges to the supervision of ship traffic. The increase in ship traffic has led to an increase in the density of ships in the waters, making ship traffic in specific waters particularly crowded and increasing the possibility of collision accidents.

[0003] In view of this situation, a comprehensive quantitative analysis and prediction of the collision risk in a specific time period in the water area is crucial to ensure the safety of maritime traffic in the area. However, the current methods for assessing collision risks in water periods are mostly based on macro-statistics. Although these methods can assess risks to a certain extent, they often ignore the actual encounter risks between ships and the pressure of the traffic environment around the ships, resulting in incomplete and inaccurate assessment results. At the same time, these methods are difficult to establish a time series of collision risks and lack analysis of the dynamic changes of collision risks. Some studies starting from a micro perspective only consider the accumulation of collision risks at isolated discrete time points. It can be seen that the above studies are difficult to capture and explore the continuity attributes and change trend characteristics of collision risks over time, thereby limiting the accuracy of identifying collision risks in water periods. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a method for identifying collision risks in water periods based on dynamic time warping. The present invention mainly uses real-time ship traffic data to accurately identify collision risks in water periods from the perspective of ship collision risk time series, thereby improving the accuracy of maritime traffic safety analysis.

[0005] The technical means adopted by the present invention are as follows:

[0006] A method for identifying water period collision hazards based on dynamic time warping comprises the following steps:

[0007] Acquire navigation status information of each ship from the ship information set, wherein the navigation status information of the ship includes the ship position, ship heading, and ship speed; calculate the closest encounter distance and minimum encounter time between two ships based on the navigation status information of the ship;

[0008] Obtain the closest encounter distance data and the minimum encounter time data between all ships in the target waters and other ships, construct a spatial representation of the collision risk in the target waters based on the closest encounter distance data, and construct a temporal representation of the collision risk in the target waters based on the minimum encounter time data; construct a geometric encounter collision risk representation of ships based on the spatial representation of the collision risk in the target waters and the temporal representation of the collision risk in the target waters;

[0009] Acquire water state information of the target water area, wherein the water state information includes water average density information; construct a traffic intensity representation of collision risk in the target water area based on the navigation state information of the ship and the water average density information; construct a ship compactness representation of collision risk in the target water area based on the navigation state information of the ship; construct a ship traffic environment pressure collision risk representation based on the traffic intensity representation of collision risk in the target water area and the ship compactness representation of collision risk in the target water area;

[0010] Based on the ship geometry collision risk representation and the ship traffic environment pressure collision risk representation, the collision risk index faced by each ship at a certain moment is calculated, so as to construct the time series of ship collision risk;

[0011] Based on the time series of ship collision risks, a distance matrix between two ships is constructed; based on the elements of the distance matrix, a cumulative distance matrix between two ships is recursively constructed; collision risk characteristics between two ships are obtained from the cumulative distance matrix; based on the collision risk characteristics, the overall collision risk characteristics faced by each ship in the current time period are calculated; the overall collision risk characteristics faced by all ships in the current time period are weightedly combined to obtain a collision risk representation for the water period; and collision risks are identified based on the collision risk representation for the water period.

[0012] The present invention also discloses a water period collision risk identification device based on dynamic time warping, which is used to implement any of the above methods, including:

[0013] A basic data acquisition module, which is used to acquire the navigation status information of each ship from the ship information set, wherein the navigation status information of the ship includes the ship position, ship heading, and ship speed; and calculate the closest encounter distance and minimum encounter time between two ships based on the navigation status information of the ship;

[0014] A ship geometric encounter collision risk representation construction module is used to obtain the closest encounter distance data and the minimum encounter time data between all ships in the target waters and other ships, construct a spatial representation of the collision risk in the target waters based on the closest encounter distance data, and construct a temporal representation of the collision risk in the target waters based on the minimum encounter time data; construct a ship geometric encounter collision risk representation based on the spatial representation of the collision risk in the target waters and the temporal representation of the collision risk in the target waters;

[0015] A ship traffic environment pressure collision risk representation construction module is used to obtain water state information of a target water area, wherein the water state information includes water average density information; construct a traffic intensity representation of collision risk in the target water area based on the ship's navigation state information and the water average density information; construct a ship compactness representation of collision risk in the target water area based on the ship's navigation state information; construct a ship traffic environment pressure collision risk representation based on the traffic intensity representation of collision risk in the target water area and the ship compactness representation of collision risk in the target water area;

[0016] The time series construction module of ship collision risk is used to calculate the collision risk index faced by each ship at a certain moment based on the ship geometry encounter collision risk representation and the ship traffic environment pressure collision risk representation, so as to construct the time series of ship collision risk;

[0017] A collision risk identification module is used to construct a distance matrix between two ships based on the time series of ship collision risks, recursively construct a cumulative distance matrix between two ships based on the elements of the distance matrix, obtain collision risk characteristics between two ships from the cumulative distance matrix, calculate the overall collision risk characteristics faced by each ship in the current time period based on the collision risk characteristics; perform weighted combination of the overall collision risk characteristics faced by all ships in the current time period to obtain a water period collision risk representation, and perform collision risk identification based on the water period collision risk representation.

[0018] Compared with the prior art, the present invention has the following advantages:

[0019] 1. Compared with the traditional traffic flow and probability model, the method for identifying collision risks in water periods based on dynamic time regularization provided by the present invention comprehensively considers the geometric encounter relationship between ships and the pressure of the traffic environment around the ships. It not only focuses on the dynamic position and motion state of the ship in the time and space level, but also includes the impact of the surrounding traffic environment pressure on the collision risk of the ship. This multi-dimensional analysis can more comprehensively evaluate the collision risk faced by the ship.

[0020] 2. The method for identifying collision risks in water periods based on dynamic time regularization provided by the present invention deeply analyzes the dynamic change process of collision risks from the perspective of time series, and can explore the influence of the time attributes and characteristics of ship collision risks on water collision risks on the basis of considering the continuity and change trend of single-ship collision risks in time. In addition, the method can handle collision risk sequences with nonlinear time differences of different time steps, different change rates, and different sequence lengths, and has high flexibility and adaptability.

[0021] 3. The method for identifying collision risks in water periods based on dynamic time regularization provided by the present invention adopts a more scientific modeling method. From the perspective of single-ship navigation, it carefully considers the contribution of each ship to the overall water risk, and organically combines them on this basis. This micro-to-macro modeling method can not only more accurately assess the risk of each ship, but also characterize and reflect the collision risk of the period as a whole. On the basis of the obtained collision risk time series, combined with the new technology of time series prediction, it is possible to more accurately predict the future collision risk of the water area, providing stronger support for the safety of ship navigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0023] Figure 1 This is a flow chart of a method for identifying collision hazards in water areas based on dynamic time regularization according to the present invention.

[0024] Figure 2 Schematic diagram of the triangle of relative motion of ships in an embodiment of the present invention.

[0025] Figure 3 Schematic diagram of the ship traffic environment perception domain in an embodiment of the present invention.

[0026] Figure 4 It is a schematic diagram of the compactness measurement of ship traffic in an embodiment of the present invention.

[0027] Figure 5 Schematic diagram of the ship collision risk time series distance matrix in an embodiment of the present invention.

[0028] Figure 6 Schematic diagram of the cumulative distance matrix of the ship collision risk time series in an embodiment of the present invention.

[0029] Figure 7Schematic diagram of a method for determining the time series weight of ship collision risk in an embodiment of the present invention.

[0030] Figure 8 It is a schematic diagram representing the 12-hour collision risk in a certain water area in an embodiment of the present invention.

[0031] Fig. 9 It is a schematic diagram representing the distribution of water collision hazards in a certain water area in an embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0033] The present invention provides a method for identifying collision risks in water areas based on dynamic time regularization. The specific process is as follows: Figure 1 As shown, it mainly includes the following steps:

[0034] S1. Acquire the navigation status information of each ship from the ship information set, wherein the navigation status information of the ship includes the ship position, ship heading, and ship speed; calculate the closest encounter distance and minimum encounter time between two ships based on the navigation status information of the ship; obtain the closest encounter distance data and minimum encounter time data between all ships in the target waters and other ships, construct a spatial representation of the collision risk in the target waters based on the closest encounter distance data, and construct a temporal representation of the collision risk in the target waters based on the minimum encounter time data; construct a ship geometric encounter collision risk representation based on the spatial representation of the collision risk in the target waters and the temporal representation of the collision risk in the target waters.

[0035] This step is mainly used to identify the geometric encounter collision risk of ships. The DCPA and TCPA parameters between ships in the waters are calculated, and the collision risk caused by other ships in the study waters to each ship is described from the spatial and temporal perspective of geometric encounter. The negative exponential function is used to establish a mapping relationship between it and the collision risk.

[0036] From the perspective of geometric encounter, the space-time collision risk posed by the ship and the surrounding ship traffic is described by collision avoidance parameters, that is, by calculating the closest encounter distance DCPA and the minimum encounter time TCPA between ships, and by using them together, the collision risk between ships can be effectively quantitatively evaluated. Based on the water area ship data set, the DCPA and TCPA between two ships can be obtained by calculating the relative motion parameters between the two ships, such as the relative distance between the ships, the relative speed of the ships, the relative orientation of the ships, the relative heading of the ships, etc. Then, the DCPA and TCPA faced by each ship are combined to form an overall spatial and temporal urgency measure faced by the ship in the water area, and the negative exponential function is used to standardize the results, establish a mapping relationship between the geometric encounter collision avoidance parameters and the collision risk, and obtain the geometric encounter collision risk CR of each ship. geometry .

[0037] Specifically, the present invention constructs a water area ship information set by extracting information such as ship mobile service identification code, timestamp, ship longitude and latitude, ship speed, ship heading, ship length, etc. in the water area from various navigation instruments and equipment, such as AIS. The necessary information is extracted, the time-space collision risk between ships is calculated, and standardized mapping is performed based on this to obtain the time-space geometric encounter level collision risk of a certain ship.

[0038] Based on the ship AIS data, ship O and ship T are placed in the relative motion triangle, and their relative motion relationship is as follows: Figure 2 As shown, the DCPA and TCPA between the two ships can be calculated according to the following formula:

[0039] First, calculate the relative distance rd between the two ships:

[0040]

[0041] Calculate the relative position c of the two ships b :

[0042]

[0043] Calculate the relative heading c r :

[0044]

[0045] Therefore, the DCPA between the two ships can be calculated OT :

[0046] DCPA OT =rd*sin(c r -c b -π)

[0047] Among them, DCPAOT represents the closest encounter distance between ship O and ship T, rd represents the relative distance between ship O and ship T, c r represents the relative heading between ship O and ship T, c b Indicates the relative position between ship O and ship T.

[0048] Using the above parameters, the TCPA between ship O and ship T, which represent the time urgency, can also be calculated: OT :

[0049]

[0050] Among them, TCPA OT represents the minimum encounter time between ship O and ship T, v r represents the relative speed between ship O and ship T.

[0051] Using the same method, the DCPA and TCPA between any two ships in the waters can be calculated. Based on the results obtained, the average spatial urgency and temporal urgency of the collision risk caused by all other ships to each ship are characterized, and then the collision risk caused by other ships in the study waters is described from the perspective of geometric encounter time and space:

[0052]

[0053] Among them, DCPA ship i The spatial representation of the collision risk of ship i in the target waters, DCPA ship i-ship j represents the closest encounter distance between ship i and ship j, n represents the number of ships in the target waters; TCPA ship i TCPA is the time representation of the collision risk of ship i in the target waters. ship i-ship j represents the minimum encounter time between ship i and ship j.

[0054] According to the negative exponential function, the DCPA and TCPA values ​​are normalized so that they can characterize the spatiotemporal urgency of the danger on a unified scale and establish a mapping relationship between them and the collision danger:

[0055]

[0056]

[0057] Among them, CR geometry Represents the collision risk of the ship geometry, where α D and β D The adjustment coefficient of the negative exponential function representing the spatial representation of collision risk, α T and βT The adjustment coefficients of the negative exponential function representing the time of collision risk can be solved by presetting the values ​​corresponding to the more extreme situations at both ends of the described situation.

[0058] S2. Obtain the water state information of the target water area, the water state information includes the average water density information; construct the traffic intensity representation of the collision risk in the target water area based on the navigation state information of the ship and the average water density information; construct the ship compactness representation of the collision risk in the target water area based on the navigation state information of the ship; construct the ship traffic environment pressure collision risk representation based on the traffic intensity representation of the collision risk in the target water area and the ship compactness representation of the collision risk in the target water area. Among them, the average water density information refers to the average (instantaneous) ship density in the study water area. It can be obtained by dividing the sum of each instantaneous ship density value within a certain period of time (such as 1 day, 1 week, 1 month, etc.) by the number of instantaneous points.

[0059] This step is mainly used to identify the collision risk and comprehensive collision risk of the ship's traffic environment pressure. Within the scope of the traffic environment perception domain, the collision risk at the level of traffic environment pressure perceived by the ship is determined, including the traffic intensity measured from the perspective of ship density and the traffic compactness measured from the perspective of the spatial relationship of the ship field.

[0060] For a ship, in addition to the collision risk from other ships in terms of time and space, the overall traffic environment pressure around it will also affect the collision risk. The present invention defines the range of the impact of surrounding ship traffic on it as the traffic environment perception domain.

[0061] Based on the water area ship data set and combined with the average density of the water area, the upper limit of the number of ships around a given ship is determined, and then the collision risk at the level of traffic intensity perceived by the ship is determined. At the same time, the ship domain overlap index P-SDOI formed by a given ship and each ship in the perception domain of the surrounding traffic environment is used to represent the compactness of the ship arrangement around the ship, and the overall spatial distribution compactness of other ships around a certain ship is measured, and mapped to collision risk through a negative exponential function, and combined with the collision risk at the level of traffic intensity to form the ship collision risk CR under the influence of traffic environment pressure. traffic .

[0062] Specifically, the present invention extracts necessary information from the ship information set in the water area, and determines the collision risk at the traffic environment pressure level perceived by the ship from the two aspects of traffic intensity and traffic compactness within the scope of the traffic environment perception domain. The traffic intensity is obtained based on the ship density in the traffic environment perception domain, and the traffic compactness is obtained through the spatial relationship of the ship field in the traffic environment perception domain.

[0063] To more accurately measure the collision risk faced by a ship during navigation, it is also necessary to describe the impact from the perspective of the pressure of the surrounding traffic environment.

[0064] The present invention defines the range of influence of surrounding ship traffic on it as the traffic environment perception domain, such as Figure 3 The necessary information is extracted from the ship information set in the water area, and combined with the average density of the water area, the upper limit of the number of ships around a given ship is determined, and then the collision risk at the traffic intensity level perceived by the ship is determined by the following formula:

[0065]

[0066] Among them, CR around The traffic intensity representation of the collision risk in the target waters, N TED Represents the number of ships in the traffic environment perception domain, N TEDmax Indicates the maximum number of ships in the traffic environment perception domain.

[0067] The compactness of the arrangement of ships around a given ship is expressed by the average degree of overlap between the ship domain and the ships in the perception domain of its surrounding traffic environment, such as Figure 4 , the index is defined as P-SDOI, which can be calculated by the following formula:

[0068]

[0069] The P-SDOI is normalized to the collision risk result by the following formula:

[0070]

[0071] Among them, CR compact The compactness of ships representing the collision risk in the target waters, α s and β s The compactness of the ship collision risk is characterized by the adjustment parameter of the negative exponential function, dis ij represents the distance between the center of the ship area of ​​ship i and ship j, r shipi and r shipj Indicates the radius length of the ship area of ​​ship i and ship j.

[0072] The adjustment parameters can be obtained by setting the safest and most dangerous extreme case results. The collision risk of a ship caused by the overall traffic environment pressure around it is calculated by the following formula, and then combined with the collision risk of the geometric encounter angle to characterize the collision risk perceived by a ship at a certain moment:

[0073]

[0074] Among them, CR traffic Represents the collision hazard representation of the ship traffic environment.

[0075] S3. Calculate the collision risk index faced by each ship at a certain moment based on the ship geometry encounter collision risk representation and the ship traffic environment pressure collision risk representation, so as to construct a time series of ship collision risks; construct a distance matrix between two ships based on the time series of ship collision risks, recursively construct an accumulated distance matrix between two ships based on the elements of the distance matrix, obtain the collision risk characteristics between two ships from the accumulated distance matrix, and calculate the overall collision risk characteristics faced by each ship in the current time period based on the collision risk characteristics; perform weighted combination of the overall collision risk characteristics faced by all ships in the current time period, so as to obtain the collision risk representation of the water period, and identify the collision risk according to the collision risk representation of the water period.

[0076] This step is mainly used to identify the collision risk in water areas. The collision risk of the geometric encounter angle is combined with the collision risk of the traffic environment to characterize the collision risk perceived by a ship at a certain moment; based on the collision risk perceived by a ship at all times in the water area obtained above, a collision risk time series of the ship is generated. The characteristics of the constructed collision risk time series are mined through the dynamic time warping algorithm, and the similarity between it and the risk sequence is identified, and the collision risk characteristics are characterized by the cumulative distance between the time series. By organically combining the collision risk characteristics of all ships in this time period, the overall collision risk result of the water area in this time period is obtained.

[0077] Specifically, the collision risk at the traffic environment pressure level and the collision risk at the geometric encounter level are combined to calculate the collision risk faced by a certain ship at a certain moment.

[0078]

[0079] On this basis, the collision risk perceived by a ship at a certain moment is characterized to form the collision risk faced by the ship that changes over time, that is, the time series of the collision risk perceived by the ship:

[0080] CS i ={CR1,CR2,CR3,…,CR t}

[0081] Among them, CR i represents the collision risk faced by ship i at a certain moment, CR traffic Indicates the collision risk of ship traffic environment pressure, CR geometryIt represents the collision risk characterization of the ship geometry, and t represents the length of the time series, that is, the number of elements contained in the time series.

[0082] The time series data mining technology - dynamic time warping algorithm is used to mine the characteristics of the collision risk time series, analyze the fluctuation pattern of the ship collision risk time series, identify the similarity between it and the risk series, and characterize the collision risk through the cumulative distance between time series.

[0083] For two ship collision risk time series CS1 and CS2, the distance matrix M D The matrix is ​​constructed, where the rows and columns of the matrix correspond to the two time series respectively, and the elements of the matrix correspond to the point-to-point Euclidean distances of the CS1 sequence and the CS2 sequence in the rows and columns. The cumulative distance matrix M of the ship collision risk time series CS1 and CS2 is further generated. C .like Figure 5 , where the rows and columns of the matrix correspond to two time series respectively, and the elements of the matrix correspond to the point-to-point Euclidean distances of the CS1 sequence and the CS2 sequence in the rows and columns.

[0084] On the basis of the above distance matrix, the cumulative distance matrix M of the ship collision risk time series CS1 and CS2 is further generated. C ,like Figure 6 Starting from the lower left corner of the cumulative distance matrix, that is, the element corresponding to the first point of the two sequences, the cumulative distance is calculated in sequence to the upper right corner of the cumulative distance matrix.

[0085] Use the following starting conditions and recursive rules to complete the calculation of the cumulative distance from the starting point to the end point:

[0086] M C (1,1)=M D (1,1)

[0087] M C (i,j)=min{M C (i,j-1),M C (i-1,j),M C (i-1,j-1)}+M D (1,1)

[0088] Among them, M C represents the accumulated distance matrix, M D represents the distance matrix, M C(i,j) represents the element in the i-th row and j-th column of the cumulative distance matrix. i and j represent the lengths of the two time series, that is, the number of elements contained in each time series, and their value range is a positive integer greater than or equal to 1. The recursive rule of the above cumulative distance is to obtain the shortest path of the current element by adding the distance of the current element to the shortest path of the previous element. For a certain element, its previous element is one of the three points adjacent to it in the matrix arrangement, and the element of the shortest path among them is taken.

[0089] The above recursive rule is for the general case of the cumulative distance matrix. If for the elements on the boundary, since the previous element is only adjacent to it in the same column or row, the cumulative distance is calculated by the following rules:

[0090] M C (i,1)=M C (i-1,1)+M D (i,1)

[0091] M C (1,j)=M C (1,j-1)+M D (1,j)

[0092] The cumulative distance in the upper right corner of the cumulative distance matrix recursively derived by the above rules can characterize the similarity between the two time series and serve as an important basis for mining the collision risk characteristics of the series.

[0093] The final cumulative distance is the cumulative distance in the upper right corner of the cumulative distance matrix recursively derived through the above rules, that is, the cumulative distance of the end point, and this value can represent the similarity of the two time series. The smaller the value, the smaller the cost of one sequence to the other, or the smaller the distance, and the higher the similarity of the two time series. And this value is the collision risk characteristic of the time series:

[0094] distance cost (CS1,CS2)=M C (i max ,j max )

[0095] Among them, distance cost (CS1, CS2) represents the collision risk characteristics, M C (i max ,j max ) represents the upper right corner element of the accumulated distance matrix.

[0096] The most dangerous time series of collision risk between ships sailing in this water area was further constructed. After calculating the final cumulative distance between a certain ship collision risk time series and the ship using the dynamic time warping algorithm, the mapping relationship between the final cumulative distance and the collision risk was established through the negative exponential function:

[0097]

[0098] Among them, CR shipi It represents the overall collision risk faced by ship i in the current time. α and β are the adjustment coefficients of the negative exponential equation of the overall collision risk characteristics, which can be solved by presetting the values ​​corresponding to the extreme cases. CS i represents the collision risk time series of ship i, CS D Indicates the most dangerous sequence, CR ship i It represents the overall collision risk faced by ship i during this period. Specifically, CS D It can be obtained by constructing a virtual time series, in which all elements are extreme values ​​of collision risk, that is, all elements of the series are 1, and the length of the series depends on the division of the research period.

[0099] On the basis of the above results, by organically combining the overall collision risk faced by all ships in the time period, the overall collision risk result of the water area in the time period can be obtained. Since each ship contributes differently to the overall collision risk of the water area during this period, different weights should be given to different ships. For each ship, because of the different lengths of time it sails in the water area, the contribution time to the overall collision risk of the water area is different. The present invention uses this as the basis for assessing the contribution of each ship to the overall collision risk of the water area, such as Figure 7 According to the weight of each ship and the collision risk characteristics faced by each ship calculated according to the above dynamic time warping algorithm, the collision risk of waters in a period of time is calculated by the following formula:

[0100]

[0101] Among them, T i-initial and T i-end They represent the start time and end time of the voyage of ship i in the research waters within the research time range, ω ship i represents the weight of ship i for the collision risk in the water area, CR interval is the overall collision risk of the waters during the study period, and n represents the number of ships in the study waters during the study period.

[0102] The scheme and effect of the present invention are further described below through specific application examples.

[0103] Figure 8 It is the collision risk change of a certain water area in 12 hours calculated by the method provided by the present invention. The collision risk situation of 24 hours is composed of the collision risk results of 12 1-hour time periods. Fig. 9 It indicates the collision risk situation of different sub-areas in the studied water area in the same hour period. The results obtained through the above embodiments can accurately perceive the changes in the collision risk of the water area and the distribution of the collision risk in the water area. Therefore, the water area period collision risk result calculated by the method provided by the present invention can effectively characterize the water area collision risk within a specified time period, and the result can serve as a strong basis for the marine traffic safety analysis of the water area.

[0104] The present invention discloses a method for identifying collision risks in water period based on dynamic time warping. The method mainly mines the characteristics of time series through the dynamic time warping algorithm, organically combines the overall collision risks faced by all ships in the time period, and obtains the overall collision risk of the water area in the time period. First, from the perspective of geometric encounter, the collision risk of the space-time angle formed by the ship and the surrounding ships is described by the collision avoidance parameters, and a mapping relationship between the collision risk and the collision risk is established. At the same time, from the perspective of the traffic intensity and compactness around the ship, the traffic environment pressure perceived by the ship is determined, and then it is combined with the collision risk of the geometric encounter angle to establish the collision risk time series of each ship. Finally, the characteristics of the time series are mined through the dynamic time warping algorithm, the similarity between the time series and the risk sequence is identified, and the collision risk characteristics are characterized by the cumulative distance between the time series, and the collision risk of the water period is quantified on the basis of considering the contribution of each ship to the collision risk in the water area. On the one hand, the invention provides a theoretical basis for ship traffic management personnel to evaluate and analyze the geometric encounter relationship between ships and the pressure of the surrounding traffic environment, so as to more accurately identify and prevent potential collision accidents and ensure the safety and efficiency of maritime transportation; on the other hand, the invention can make use of the characteristics of time series to further explore the influence of time-varying attributes and characteristics on the risk of collision in waters, and more accurately identify the risk of collision in waters during certain periods. At the same time, the invention helps to more accurately predict the future risk of collision in waters, so as to timely deploy preventive measures and improve maritime traffic safety.

[0105] Another aspect of the embodiment of the present invention further provides a water period collision risk identification device based on dynamic time warping, which is used to implement any of the above methods, including:

[0106] A basic data acquisition module, which is used to acquire the navigation status information of each ship from the ship information set, wherein the navigation status information of the ship includes the ship position, ship heading, and ship speed; and calculate the closest encounter distance and minimum encounter time between two ships based on the navigation status information of the ship;

[0107] A ship geometric encounter collision risk representation construction module is used to obtain the closest encounter distance data and the minimum encounter time data between all ships in the target waters and other ships, construct a spatial representation of the collision risk in the target waters based on the closest encounter distance data, and construct a temporal representation of the collision risk in the target waters based on the minimum encounter time data; construct a ship geometric encounter collision risk representation based on the spatial representation of the collision risk in the target waters and the temporal representation of the collision risk in the target waters;

[0108] A ship traffic environment pressure collision risk representation construction module is used to obtain water state information of a target water area, wherein the water state information includes water average density information; construct a traffic intensity representation of collision risk in the target water area based on the ship's navigation state information and the water average density information; construct a ship compactness representation of collision risk in the target water area based on the ship's navigation state information; construct a ship traffic environment pressure collision risk representation based on the traffic intensity representation of collision risk in the target water area and the ship compactness representation of collision risk in the target water area;

[0109] The time series construction module of ship collision risk is used to calculate the collision risk index faced by each ship at a certain moment based on the ship geometry encounter collision risk representation and the ship traffic environment pressure collision risk representation, so as to construct the time series of ship collision risk;

[0110] A collision risk identification module is used to construct a distance matrix between two ships based on the time series of ship collision risks, recursively construct a cumulative distance matrix between two ships based on the elements of the distance matrix, obtain collision risk characteristics between two ships from the cumulative distance matrix, calculate the overall collision risk characteristics faced by each ship in the current time period based on the collision risk characteristics; perform weighted combination of the overall collision risk characteristics faced by all ships in the current time period to obtain a water period collision risk representation, and perform collision risk identification based on the water period collision risk representation.

[0111] As for an embodiment of a device for identifying collision hazards in water periods based on dynamic time regularization of the present invention, since it corresponds to an embodiment of a method for identifying collision hazards in water periods based on dynamic time regularization above, the description is relatively simple. For relevant similarities, please refer to the partial description in the embodiment of a method for identifying collision hazards in water periods based on dynamic time regularization above, which will not be described in detail here.

[0112] 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 aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying water period collision hazards based on dynamic time warping, characterized in that: The following steps are involved: Acquire navigation status information of each ship from the ship information set, wherein the navigation status information of the ship includes the ship position, ship heading, and ship speed; calculate the closest encounter distance and minimum encounter time between two ships based on the navigation status information of the ship; Obtain the closest encounter distance data and the minimum encounter time data between all ships in the target waters and other ships, construct a spatial representation of the collision risk in the target waters based on the closest encounter distance data, and construct a temporal representation of the collision risk in the target waters based on the minimum encounter time data; construct a geometric encounter collision risk representation of ships based on the spatial representation of the collision risk in the target waters and the temporal representation of the collision risk in the target waters; Acquire water state information of the target water area, wherein the water state information includes water average density information; construct a traffic intensity representation of collision risk in the target water area based on the navigation state information of the ship and the water average density information; construct a ship compactness representation of collision risk in the target water area based on the navigation state information of the ship; construct a ship traffic environment pressure collision risk representation based on the traffic intensity representation of collision risk in the target water area and the ship compactness representation of collision risk in the target water area; Based on the ship geometry collision risk representation and the ship traffic environment pressure collision risk representation, the collision risk index faced by each ship at a certain moment is calculated, so as to construct the time series of ship collision risk; Based on the time series of ship collision risks, a distance matrix between two ships is constructed; based on the elements of the distance matrix, a cumulative distance matrix between two ships is recursively constructed; collision risk characteristics between two ships are obtained from the cumulative distance matrix; based on the collision risk characteristics, the overall collision risk characteristics faced by each ship in the current time period are calculated; the overall collision risk characteristics faced by all ships in the current time period are weightedly combined to obtain a collision risk representation for the water period; and collision risks are identified based on the collision risk representation for the water period.

2. The method for identifying water area collision risk based on dynamic time warping according to claim 1 is characterized in that: The closest approach distance between two ships is calculated according to the following formula: DCPA OT =rd*sin(c r -c b -π) where DCPA OT represents the closest encounter distance between ship O and ship T, rd represents the relative distance between ship O and ship T, c r represents the relative heading between ship O and ship T, c b Indicates the relative position between ship O and ship T; The minimum encounter time is calculated according to the following formula: Among them, TCPA OT represents the minimum encounter time between ship O and ship T, v r represents the relative speed between ship O and ship T.

3. The method for identifying water area collision risk based on dynamic time warping according to claim 2 is characterized in that: The spatial representation of the collision risk in the target water area is obtained by the following calculation: Among them, DCPA shipi The spatial representation of the collision risk of ship i in the target waters, DCPA ship i-ship j represents the closest encounter distance between ship i and ship j, and n represents the number of ships in the target waters; The temporal representation of the collision risk in the target waters is obtained by the following calculation: Among them, TCPA shipi TCPA is the time representation of the collision risk of ship i in the target waters. ship i-ship j represents the minimum encounter time between ship i and ship j.

4. The method for identifying water area collision risk based on dynamic time warping according to claim 3 is characterized in that: The ship geometry encounters a collision risk representation based on the following calculations: Among them, CR geometry Represents the collision risk of the ship geometry, where α D and β D The adjustment coefficient of the negative exponential function representing the spatial representation of collision risk, α T and β T The adjustment coefficient of the negative exponential function representing the time of collision risk.

5. The method for identifying water area collision risk based on dynamic time warping according to claim 1 is characterized in that: The traffic intensity representation of collision risk in the target water area is obtained by the following calculation: Among them, CR around The traffic intensity representation of the collision risk in the target waters, N TED Represents the number of ships in the traffic environment perception domain, N TEDmax Indicates the maximum number of ships in the traffic environment perception domain; The compactness representation of the ship collision risk in the target waters is obtained by the following calculation: P-SDOI shipi Standardize to get: Among them, CR compact The compactness of ships representing the collision risk in the target waters, α s and β s The compactness of the ship collision risk is characterized by the adjustment parameter of the negative exponential function, dis i′j′ represents the distance between the center of the ship area of ​​ship i and ship j, r shipi and r shipj represents the radius length of the ship area of ​​ship i and ship j; The pressure collision hazard characterization of the ship traffic environment is obtained by the following calculation: Among them, CR traffic Represents the collision hazard representation of the ship traffic environment.

6. The method for identifying water area collision risk based on dynamic time warping according to claim 1 is characterized in that: The time series of ship collision risk is: CS i ={CR1,CR2,CR3,…,CR t } Among them, CR i represents the collision risk faced by ship i at a certain moment, CR traffic Indicates the collision risk of ship traffic environment pressure, CR geometry It represents the collision risk characterization of the ship geometry, and t represents the length of the time series.

7. The method for identifying water area collision risk based on dynamic time warping according to claim 1 is characterized in that: The elements of the distance matrix correspond to the point-to-point Euclidean distances of the time series of the risk of collision between two ships; The accumulated distance matrix is ​​constructed as follows: Starting from the lower left corner of the cumulative distance matrix, the cumulative distance is calculated towards the upper right corner of the cumulative distance matrix. The starting conditions and recursive rules are as follows: M C (1,1)=M D (1,1) M C (i,j)=min{M C (i,j-1),M C (i-1,j),M C (i-1,j-1)}+M D (1,1) Among them, M C (1,1) represents the lower left corner element of the cumulative distance matrix, M D (1,1) represents the lower left corner element of the distance matrix, For elements on the boundary, the cumulative distance is calculated using the following rules: M C (i,1)=M C (i-1,1)+M D (i,1) M C (1,j)=M C (1,j-1)+M D (1,j) Among them, M C represents the accumulated distance matrix, M D represents the distance matrix, M C (i,j) represents the element in the i-th row and j-th column of the cumulative distance matrix. i and j represent the lengths of the two time series, that is, the number of elements contained in each time series. Their value range is a positive integer greater than or equal to 1.

8. The method for identifying water area collision risk based on dynamic time warping according to claim 7 is characterized in that: The collision risk characteristics are obtained based on the following calculations: distance cos (CS1,CS2)=M C (i max ,j max ) Among them, distance cost (CS1, CS2) represents the collision risk characteristics, M C (i max ,j max ) represents the upper right corner element of the accumulated distance matrix; The overall collision risk characteristics faced by the ship in the current period are obtained according to the following calculation: Among them, CR shipi represents the overall collision risk faced by ship i at the current time, α and β are the adjustment coefficients of the overall collision risk characteristic negative exponential equation, CS i , represents the time series of collision risk of ship i, CS D represents the most dangerous sequence; The collision risk characterization for the water period is obtained by the following calculation: Among them, CR interval is the overall collision risk of the target waters in the current period, T i-initial and T i-end They represent the start time and end time of ship i sailing in the target waters in the current period, ω shipi It represents the weight of ship i for the collision risk in the water area, and n represents the number of ships in the study waters during the study period.

9. The method for identifying water area collision risk based on dynamic time warping according to claim 1, characterized in that: The collision risk identification is performed according to the collision risk characterization of the water period, including: sensing the change of the collision risk of the target water area and the distribution of the collision risk in the target water area according to the collision risk characterization of the water period.

10. A device for identifying water period collision hazards based on dynamic time warping, used to implement the method described in any one of claims 1 to 9, characterized in that: include: A basic data acquisition module, which is used to acquire the navigation status information of each ship from the ship information set, wherein the navigation status information of the ship includes the ship position, ship heading, and ship speed; and calculate the closest encounter distance and minimum encounter time between two ships based on the navigation status information of the ship; A ship geometric encounter collision risk representation construction module is used to obtain the closest encounter distance data and the minimum encounter time data between all ships in the target waters and other ships, construct a spatial representation of the collision risk in the target waters based on the closest encounter distance data, and construct a temporal representation of the collision risk in the target waters based on the minimum encounter time data; construct a ship geometric encounter collision risk representation based on the spatial representation of the collision risk in the target waters and the temporal representation of the collision risk in the target waters; A ship traffic environment pressure collision risk representation construction module is used to obtain water state information of a target water area, wherein the water state information includes water average density information; construct a traffic intensity representation of collision risk in the target water area based on the ship's navigation state information and the water average density information; construct a ship compactness representation of collision risk in the target water area based on the ship's navigation state information; construct a ship traffic environment pressure collision risk representation based on the traffic intensity representation of collision risk in the target water area and the ship compactness representation of collision risk in the target water area; The time series construction module of ship collision risk is used to calculate the collision risk index faced by each ship at a certain moment based on the ship geometry encounter collision risk representation and the ship traffic environment pressure collision risk representation, so as to construct the time series of ship collision risk; A collision risk identification module is used to construct a distance matrix between two ships based on the time series of ship collision risks, recursively construct a cumulative distance matrix between two ships based on the elements of the distance matrix, obtain collision risk characteristics between two ships from the cumulative distance matrix, calculate the overall collision risk characteristics faced by each ship in the current time period based on the collision risk characteristics; perform weighted combination of the overall collision risk characteristics faced by all ships in the current time period to obtain a water period collision risk representation, and perform collision risk identification based on the water period collision risk representation.

Images