A Method and System for Ship Collision Risk Assessment Based on Macro-Micro Risk

Abstract

This invention discloses a method and system for ship collision risk assessment based on macro-micro risk, relating to the field of maritime science and technology. The method includes the following steps: collecting historical ship collision accident characteristic data, including collision location, traffic flow density, signal and communication coverage, management response time, and collision frequency, and training a neural network model. Using location, traffic flow density, and communication coverage as inputs, the method outputs predicted values ​​for management response time and collision frequency, calculating a macro-risk assessment value. Real-time navigation parameters of the ship to be analyzed, parameters of encountered ships, and environmental impact parameters are collected to calculate the ship behavior state index, encounter risk coefficient, and environmental impact factor, obtaining a micro-risk assessment value. Finally, the collision risk assessment coefficient is calculated by combining the macro- and micro-risk assessment values ​​and compared with a set threshold to generate a risk assessment result. This method provides real-time and accurate risk assessment results, thereby effectively reducing the occurrence of collision accidents.

Description

Technical Field

[0001] This invention relates to the field of marine science and technology, specifically to a method and system for assessing ship collision risk based on macro-micro risks. Background Technology

[0002] Ship collisions are a significant issue affecting global shipping safety. With the continuous development of the global shipping industry and the increasing number of ships, the frequency and consequences of ship collisions are becoming increasingly severe. Collisions can cause significant losses, including casualties, vessel damage, and environmental pollution, and can also disrupt normal shipping operations, bringing immeasurable negative impacts to global trade and the economy. Accurate assessment and effective prevention of ship collision risks have become core tasks of modern shipping safety management.

[0003] Traditional collision risk assessment methods typically focus on quantitatively analyzing the impact of a specific factor on collision risk, while neglecting the interaction between multiple factors. For example, some assessment methods may only focus on the ship's speed and course, assuming that higher speed means higher collision risk. However, in reality, the occurrence of ship collision risk often involves multiple factors, including sea traffic flow, ship behavior patterns, and sea weather conditions. These factors work together to influence the probability of a collision accident.

[0004] Many existing assessment methods fail to consider the macro-level context, focusing only on real-time vessel behavior, the dynamic changes of encountering vessels, and the environmental impacts of weather and waterways. This results in insufficient accuracy and timeliness of their assessments. Especially in complex maritime traffic environments, single-parameter models often cannot comprehensively assess changes in collision risk and potential hazards. For example, in busy waterways, even at moderate vessel speeds, collisions can still occur if the dynamic behavior of other vessels and external environmental factors are not taken into account. Therefore, existing risk assessment methods often suffer from inaccurate assessments and incomplete consideration of factors.

[0005] In the prior art, CN111951606B discloses a method and system for ship collision risk assessment and early warning. This method includes: acquiring hydrological and meteorological information about the location of a designated ship, and acquiring navigation information between the designated ship and other ships; assessing the real-time collision risk of the designated ship using a trained adaptive ship collision risk assessment model based on the hydrological, meteorological, and navigation information; determining the collision risk level of the designated ship based on the real-time collision risk; and outputting an early warning message associated with the risk level to the designated ship. However, this system only assesses the collision risk of a single ship, which may not consider the interactive effects between other ships, especially in densely populated waters where collision risk is often caused by the interaction between multiple ships. This method does not fully consider the potential impact of the collaborative effects of multiple ships on the comprehensiveness and accuracy of the assessment. Furthermore, relying solely on a trained adaptive ship collision risk assessment model for risk assessment, the comprehensiveness and balance of the model's training dataset are also important factors affecting the model's accuracy. If the training data is insufficient to cover all possible collision scenarios, the model may perform poorly in certain situations, thus potentially reducing the accuracy and effectiveness of the risk assessment.

[0006] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0007] The purpose of this invention is to provide a method and system for assessing ship collision risk based on macro-micro risk, so as to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: A method for assessing ship collision risk based on macro-micro risk, comprising the following steps: Historical characteristic data of ship collision accidents are collected, including collision location, management response time, traffic flow density, signal and communication coverage, and number of collisions in the affected area corresponding to the collision location, to construct a sample dataset; Based on the sample dataset, with collision location, traffic flow density and signal communication coverage as inputs, and management response time and number of collisions as outputs, a neural network prediction model is trained to obtain a ship collision accident risk prediction model. The location information, traffic flow density, and signal communication coverage of the waters where the vessel to be analyzed is located are input into the vessel collision accident risk prediction model, which outputs the predicted management response time and the predicted number of collisions, and calculates the macro-collision risk assessment value. Collect navigation behavior parameters of the ship to be analyzed, encounter ship behavior parameters and environmental impact parameters, generate ship behavior state index, encounter risk coefficient and environmental impact factor, and calculate micro collision risk assessment value; The ship collision risk assessment coefficient is calculated based on the macro-collision risk assessment value and the micro-collision risk assessment value, and the ship collision risk assessment coefficient is compared with the preset risk assessment threshold to output the corresponding risk assessment result.

[0009] Furthermore, the collision location is represented by geographical coordinates, and the collision impact area is determined with the collision location as the center and a preset radius range. Within the collision impact area, traffic flow density, signal communication coverage, and the number of collisions were statistically analyzed. The signal communication coverage is characterized by a combination of signal reception quality, communication network availability time, and signal reception strength within the sea area. The management response time is the time from the occurrence of the accident to the start of intervention by the management agency.

[0010] Furthermore, a neural network prediction model is constructed using a Long Short-Term Memory (LSTM) network. The model uses Tanh as the activation function, Adam as the optimization algorithm, and is configured with a three-layer network structure, 200 iterations, a learning rate of 0.001, a batch size of 32, 100 training iterations, 256 batch processing units, and 32 hidden layer neurons. The model inputs include the location of the ship collision, traffic flow density, and signal communication coverage, while the management response time and the number of collisions are used as the model outputs.

[0011] Furthermore, based on the obtained management response time prediction value and ship collision frequency prediction value, combined with the traffic flow density and signal communication coverage of the water area where the ship to be analyzed is located, a macro collision risk assessment value is generated according to the preset macro risk fusion rule. The macro risk fusion rule is used to comprehensively reflect the frequency of ship collisions, regional traffic pressure, accident handling efficiency and communication support level. The higher the predicted number of ship collisions, the higher the traffic flow density, the longer the predicted management response time, and the lower the signal communication coverage, the higher the macro collision risk assessment value. The signal communication coverage rate is determined comprehensively based on the signal reception quality, communication network availability time, and signal reception strength within the sea area. The better the signal reception quality, the longer the communication network availability time, and the higher the signal reception strength, the higher the signal communication coverage rate.

[0012] Furthermore, real-time navigation behavior parameters, encounter vessel behavior parameters, and environmental impact parameters of the vessel to be analyzed are collected. Based on the navigation behavior parameters, encounter vessel behavior parameters, and environmental impact parameters, a vessel behavior state index, an encounter risk coefficient, and an environmental impact factor are generated. The vessel behavior state index is used to reflect the degree of drastic change in the navigation state of the vessel to be analyzed. The greater the rate of change in speed and the rate of change in course, the higher the vessel behavior state index. The encounter risk coefficient is used to reflect the degree of danger of approach between the ship being analyzed and the encountering ship. The smaller the shortest distance and the greater the relative speed, the higher the encounter risk coefficient. The result corresponding to the most serious risk among the encountering ships is taken as the encounter risk coefficient. The environmental impact factor is used to reflect the degree of impact of external sea conditions on collision risk, and the higher the maximum wind speed, maximum wave height and maximum water flow velocity, the higher the environmental impact factor.

[0013] Furthermore, based on the obtained ship behavior state index, encounter risk coefficient, and environmental impact factors, a micro-collision risk assessment value is generated according to the preset micro-risk fusion logic; The micro-risk fusion logic includes: enhancing the environmental impact factors, superimposing the enhanced environmental impact factors with the smoothed result of the ship behavior state index, and nonlinearly adjusting the superimposed result according to the encounter risk coefficient to obtain the micro-collision risk assessment value. The higher the ship behavior state index, the greater the environmental impact factor, and the higher the encounter risk coefficient, the higher the micro-collision risk assessment value.

[0014] Furthermore, based on the micro-collision risk assessment value and the macro-collision risk assessment value, the ship collision risk evaluation coefficient is calculated according to the preset fusion logic; The fusion logic is used to perform weighted fusion of macro-collision risk assessment values ​​and micro-collision risk assessment values, and the weight corresponding to the micro-collision risk assessment value is not lower than the weight corresponding to the macro-collision risk assessment value. The ship collision risk assessment coefficient is compared with a preset risk assessment threshold, and the corresponding risk assessment result is generated based on the comparison result. When the collision risk assessment coefficient is greater than or equal to the risk assessment threshold, it is determined to be a high collision risk, indicating that the water area has a high collision risk and the vessel should leave as soon as possible. When the ship collision risk assessment coefficient is greater than or equal to 40% of the risk assessment threshold and less than the risk assessment threshold, it is judged as a medium collision risk, indicating that there is a collision risk in the waters and one should be more vigilant and navigate with caution. When the collision risk assessment coefficient is less than 40 percent of the risk assessment threshold, it is judged as low collision risk, indicating that the probability of a collision in the waters is low and it is safe to navigate.

[0015] This invention also provides a ship collision risk assessment system based on macro-micro risk, which is used to execute the aforementioned ship collision risk assessment method based on macro-micro risk, including: The sample data acquisition module is used to collect historical characteristic data of ship collision accidents. The historical characteristic data includes the collision location, management response time, traffic flow density, signal communication coverage and number of collisions in the area affected by the collision location, and constructs a sample dataset. The neural network training module is used to train a neural network prediction model based on the sample dataset, taking collision location, traffic flow density and signal communication coverage as inputs, and management response time and number of collisions as outputs, to obtain a ship collision accident risk prediction model. The macro-collision risk analysis module is used to input the location information, traffic flow density and signal communication coverage of the waters where the vessel to be analyzed is located into the vessel collision accident risk prediction model, output the management response time prediction value and the collision occurrence number prediction value, and calculate the macro-collision risk assessment value. The micro-collision risk analysis module is used to collect navigation behavior parameters of the ship to be analyzed, encounter ship behavior parameters and environmental impact parameters, generate ship behavior state index, encounter risk coefficient and environmental impact factor, and calculate micro-collision risk assessment value. The collision risk assessment module is used to calculate the ship collision risk assessment coefficient based on the macro collision risk assessment value and the micro collision risk assessment value, compare the ship collision risk assessment coefficient with the preset risk assessment threshold, and output the corresponding risk assessment result.

[0016] Compared with the prior art, the beneficial effects of the present invention are: First, this invention collects various historical characteristic data, including multiple macro-level factors such as collision location, traffic flow density, signal and communication coverage, and management response time, and combines this data to establish a neural network prediction model, generating a ship collision risk prediction model. By comprehensively considering multiple factors, the collision risk assessment becomes more accurate and reliable. The model trained using historical data can predict potential collision risks in advance and provide data support for actual navigation, enabling ships to better cope with different traffic environments and risk scenarios during navigation. Second, it combines macro-risk assessment with micro-risk analysis. By collecting real-time navigation behavior parameters of the ship to be analyzed, behavior parameters of encountering ships, and environmental impact parameters, it further generates a ship behavior state index, encounter risk coefficient, and environmental impact factor, conducting a detailed assessment of micro-risks. It accurately captures changes in the ship during dynamic navigation, comprehensively considering the impact of changes in ship speed, course, relative position and speed of encountering ships, and environmental conditions (such as wind speed and wave height) on collision risk. It can comprehensively reflect changes in the ship's behavior state and external environmental factors. Finally, by combining the results of macro-risk assessment and micro-risk assessment, a ship collision risk evaluation coefficient is calculated. This coefficient is then compared with a set risk evaluation threshold to quickly determine the ship's risk level. This comprehensive evaluation method provides shipping companies, maritime authorities, and ship commanders with real-time and accurate risk assessment results, enabling them to take timely countermeasures and effectively reduce the occurrence of collision accidents. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall method flow of the present invention; Figure 2 This is a schematic diagram of the overall system structure of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0019] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0020] Example: Please see Figure 1 The present invention provides a technical solution: A method for assessing ship collision risk based on macro-micro risk, comprising the following steps: Step 1: Collect historical characteristic data of ship collision accidents. The historical characteristic data includes the collision location, management response time, traffic flow density, signal communication coverage and number of collisions in the area affected by the collision location, and construct a sample dataset.

[0021] The collision location is represented by geographic coordinates, and the collision impact area is determined with the collision location as the center and a preset radius range. It involves specific sea areas, ports or waterways. Geographic information systems are used to visualize and analyze the distribution of these locations. The logic for determining the collision impact area based on ship collision location data is as follows: the scope of each collision location is a circular area with a radius of 3 kilometers centered on the geographic coordinates of that point, forming an impact area. Within the collision impact area, traffic flow density, signal communication coverage, and the number of collisions were statistically analyzed. The signal communication coverage is characterized by a combination of signal reception quality, communication network availability time, and signal reception strength within the sea area. The management response time is the time from the occurrence of the accident to the start of intervention by the management agency. The management response time specifically refers to the time taken by the management agency to respond to accident reports from all collision incidents that occurred in the collision-affected area within the time period from the current moment to 5 years ago. This includes the time from the occurrence of the accident to the start of management intervention, and is expressed in hours. Specifically, the management response time is the average of the response times of all collision incidents.

[0022] The methods for obtaining the location of a ship collision include: obtaining information on the location of a ship collision through accident reports from maritime authorities, shipping company records, and AIS (Automatic Identification System) data; Method for obtaining traffic flow density: Use AIS data to monitor the ship flow in the collision impact area from the current time to 5 years ago, and count the number of ships passing through the collision impact area in the hour before each collision, and express it as the number of ships passing through the collision impact area in one hour; Methods for obtaining the number of ship collisions: The number of ship collisions in the affected area from the current time to 5 years ago is obtained through historical accident records and maritime department statistics. The annual average number of collisions in the affected area during the period from the current time to 5 years ago is used as the number of ship collisions in the affected area.

[0023] Step 2: Based on the sample dataset, with collision location, traffic flow density and signal communication coverage as inputs, and management response time and number of collisions as outputs, train a neural network prediction model to obtain a ship collision accident risk prediction model.

[0024] A neural network prediction model is constructed using a Long Short-Term Memory (LSTM) network. The model uses Tanh as the activation function, Adam as the optimization algorithm, and is configured with a three-layer network structure, 200 iterations, a learning rate of 0.001, a batch size of 32, 100 training iterations, 256 batch processing units, and 32 hidden layer neurons. The model inputs include the location of the ship collision, traffic flow density, and signal communication coverage, while the management response time and the number of collisions are used as the model outputs.

[0025] The Tanh function was chosen as the activation function, and Adam was chosen as the optimization algorithm for the LSTM model. The formula for the Tanh function is: In the formula, Represents the Tanh function, with the independent variable... This represents the weighted sum of the neuron's inputs, that is, the result of the weighted sum of the inputs received by the neuron from the previous layer; Simultaneously, the hyperparameters of the LSTM model are set, including: number of network layers, number of iterations, learning rate, batch size, number of training iterations, number of batches, and number of hidden layer neurons. The network is set to a 3-layer network structure, the number of iterations is set to 200, the learning rate is set to 0.001, the batch size is set to 32, the number of training iterations is set to 100, the batch size is set to 256, and the number of hidden layer neurons is set to 32. The inputs to the trained ship collision accident risk prediction model are three feature parameters: ship collision location, traffic flow density, and signal and communication coverage. The outputs are management response time and the number of ship collisions.

[0026] Ship collision risk assessment involves multiple dynamic factors (such as historical characteristics of collision locations, changes in traffic flow density, and changes in signal and communication coverage), which are essentially time-series data. LSTM models have a natural advantage in processing time-series data, capable of capturing long-term and short-term dependencies over time, thus predicting risks more accurately. Ship collision risk is a complex, multi-factor, interactive nonlinear problem. As a deep learning model, LSTM has powerful nonlinear modeling capabilities, capturing the complex relationship between input features (such as collision locations, traffic flow density, and signal and communication coverage) and objectives (such as management response time and the number of collisions). Compared to traditional linear models, LSTM can more accurately reflect real risk patterns.

[0027] The risk of ship collisions can be influenced by cumulative factors over long periods of time. For example, a sustained high traffic density in a particular body of water over an extended period may significantly increase the risk of future collisions. LSTM is well-suited to capturing such long-term trends, while traditional neural network models (such as typical feedforward neural networks) struggle to handle this long-term dependency problem.

[0028] Step 3: Input the location information, traffic flow density and signal communication coverage of the waters where the vessel to be analyzed is located into the vessel collision accident risk prediction model, output the management response time prediction value and the collision occurrence number prediction value, and calculate the macro collision risk assessment value.

[0029] Based on the obtained management response time prediction and ship collision frequency prediction, combined with the traffic flow density and signal communication coverage of the waters where the ship to be analyzed is located, a macro collision risk assessment value is generated according to the preset macro risk fusion rules. The macro risk fusion rules are used to comprehensively reflect the frequency of ship collisions, regional traffic pressure, accident handling efficiency and communication support level. The higher the predicted number of ship collisions, the higher the traffic flow density, the longer the predicted management response time, and the lower the signal communication coverage, the higher the macro collision risk assessment value. The formula used to calculate the macroscopic collision risk assessment value is as follows: In the formula, This is the macro-level collision risk assessment value. This is a predicted value for the number of ship collisions. To determine the traffic flow density in the waters where the vessel to be analyzed is located, To determine the signal communication coverage of the waters where the ship to be analyzed is located, For managing response time predictions.

[0030] Among them, the macro collision risk assessment value This assessment combines characteristic parameters such as predicted collision frequency and predicted management response time to provide a macroscopic representation of collision risk. It should be noted that this macroscopic collision risk assessment value... The larger the value, the higher the risk of a collision.

[0031] Among them, the predicted number of ship collisions This reflects the predicted frequency of historical ship collisions. This is a core factor directly impacting risk assessment. A higher frequency indicates a greater historical collision risk in that waterway; therefore, the predicted number of ship collisions... Risk assessment value of collision with macroeconomics Proportional.

[0032] Traffic flow density in the waters where the vessel is located to be analyzed This indicates the traffic density of the waterway where the vessel being analyzed is located. Higher traffic density increases the likelihood of collisions between vessels. Therefore, It is also an important positive factor in risk assessment, therefore the traffic flow density in the waters where the vessel is located needs to be analyzed. Risk assessment value of collision with macroeconomics Proportional to the predicted number of ship collisions. The product form together represents a proportional relationship.

[0033] Signal communication coverage of the waters where the ship to be analyzed is located This represents the coverage of communication facilities in the waterway. Higher signal communication coverage helps improve the efficiency of communication between ships (such as Automatic Identification System [AIS] and communication and navigation signals), thereby reducing the risk of collision. Therefore, the signal communication coverage of the waterway where the ship to be analyzed is located is important. Risk assessment value of collision with macroeconomics Inversely proportional, through the denominator The form represents an inverse proportional relationship.

[0034] Management response time prediction This refers to the estimated timeframe for relevant management departments (such as maritime authorities) to take response measures when a collision is likely to occur. The more timely the response (i.e., the faster the response), the better. The smaller the value, the lower the risk; the slower the response, the higher the collision risk. Therefore, the management response time prediction value... Risk assessment value of collision with macroeconomics Proportional, through the logarithmic function Indicates smaller It will not increase significantly However, the larger The effect is gradually amplified through a logarithmic function. This design reasonably reflects the importance of response time without allowing it to have an excessive impact on the result.

[0035] The signal communication coverage rate is determined comprehensively based on the signal reception quality, communication network availability time, and signal reception strength within the sea area. The better the signal reception quality, the longer the communication network availability time, and the higher the signal reception strength, the higher the signal communication coverage rate.

[0036] The specific formula used to calculate signal communication coverage is as follows: In the formula, For signal reception strength, For the available time of the communication network, This refers to the quality of signal reception within the sea area.

[0037] Among them, signal reception strength Specifically, it refers to the strength of the wireless signal received by the receiving terminal of the ship to be analyzed at its current location in the waters, usually measured in decibels and milliwatts; Available time of communication network Specifically, this refers to the proportion of time within a certain time period during which the communication network can operate normally, and the signal reception quality within the sea area. The signal received strength is expressed as the reciprocal of the bit error rate. The specific method for obtaining the numerical value is as follows: detect the wireless signal strength within a certain time period, and use the average signal strength as the received signal strength. .

[0038] The network availability time can be obtained by using the management system of the communication equipment (such as base station manager or core network monitoring system). These systems record the fault duration and runtime of the network equipment. Wireless signal strength can be measured in real time using a dedicated signal tester (such as a wireless signal analyzer from Anritsu or Keysight). Signal bit error rate can be tested using dedicated equipment such as a bit error rate tester (BERT). The tester generates a standard test signal and detects signal errors at the receiving end. The specified time period is, for example, 10 minutes or 30 minutes.

[0039] Among them, the signal communication coverage of the waters where the ship to be analyzed is located The larger the value, the greater the signal communication coverage and the better the communication effect.

[0040] Among them, signal reception strength This reflects the strength of communication signals within the water area being analyzed. Higher signal strength indicates greater stability and reliability of communication. It is an important positive factor affecting communication coverage, therefore signal reception strength Signal communication coverage of the waters where the vessel to be analyzed is located Proportional, through an exponential function It indicates a direct proportional relationship.

[0041] Available time of communication network The availability time represents the proportion of a communication network that can operate normally within a given time period, reflecting the reliability of the communication infrastructure. The longer the availability time, the higher the communication coverage; therefore, the longer the availability time of the communication network... Signal communication coverage of the waters where the vessel to be analyzed is located Proportional.

[0042] Signal reception quality within the sea area This indicates the clarity and stability of signal reception, taking into account transmission losses within the water (such as multipath interference and obstruction). Higher signal quality results in higher communication coverage; therefore, higher signal reception quality within the sea area is crucial. Signal communication coverage of the waters where the vessel to be analyzed is located Proportional to the availability of the communication network. Through square root The form of the equations together represents a proportional relationship, and the output of the square root function gradually decreases as the input increases. This indicates that the availability of the communication network and the quality of signal reception have limited impact on improving coverage.

[0043] Step 4: Collect navigation behavior parameters of the ship to be analyzed, encounter ship behavior parameters and environmental impact parameters, generate ship behavior state index, encounter risk coefficient and environmental impact factor, and calculate micro-collision risk assessment value.

[0044] The environmental impact parameters mentioned above specifically refer to the data of the waters where the vessel being analyzed is currently located.

[0045] Real-time navigation behavior parameters, encounter vessel behavior parameters, and environmental impact parameters of the vessel to be analyzed are collected. Based on these parameters, a vessel behavior state index, an encounter risk coefficient, and an environmental impact factor are generated. The vessel behavior state index reflects the severity of changes in the vessel's navigation state; the higher the rate of change in speed and course, the higher the index. The formula used to calculate the vessel behavior state index is as follows: In the formula, This is a ship behavior status index. For the rate of change of velocity, The rate of change of heading; wherein, the rate of change of velocity The specific calculation method is as follows: calculate the absolute difference between the current speed and the speed 3 minutes ago, and use the ratio of this absolute difference to the time of 3 minutes as the rate of change of speed. Rate of change of heading The specific calculation method is as follows: calculate the absolute difference between the heading angle at the current time and 3 minutes ago, and use the ratio of this absolute angle difference to the time of 3 minutes as the heading change rate. .

[0046] Among them, the ship behavior status index The larger the value, the greater the change in the ship's behavior, the more turbulent the navigation process, and the greater the possibility of a collision.

[0047] Rate of change of velocity This represents the degree of change in a ship's speed. A larger rate of change in speed generally indicates more unstable ship behavior, potentially posing a higher risk or requiring greater attention. Therefore, the rate of change in speed... Ship behavior status index Proportional. Through It can handle the nonlinear effects of the rate of change of velocity. The introduction of the logarithmic function makes the effect of small velocity changes on the exponent more moderate, while large changes significantly increase the exponent. This helps to distinguish different degrees of velocity change.

[0048] Rate of change of heading This represents the degree of change in a ship's course. A greater change in course indicates a more drastic turn, which may also indicate abnormal or unstable behavior; therefore, the rate of change in course is important. Ship behavior status index Proportional. Through the exponential function. Expressing the impact of course changes can capture the significant effects of abrupt course changes.

[0049] The encounter risk coefficient reflects the degree of danger of approach between the analyzed vessel and other vessels in the encounter. The smaller the shortest distance and the greater the relative speed, the higher the encounter risk coefficient. The encounter risk coefficient is based on the result corresponding to the vessel with the most severe risk among all vessels in the encounter. The formula used to calculate the encounter risk coefficient is as follows: In the formula, To the risk factor, Let be the current shortest distance between the i-th encountering vessel and the vessel to be analyzed. Let be the relative speed between the i-th encountering vessel and the vessel to be analyzed, where i is the index of the encountering vessel, and , The total number of vessels encountered; Among them, the risk factor will be encountered This is the result of a comprehensive risk assessment between multiple encountering vessels and the vessel being analyzed, reflecting the most dangerous encounter scenarios. A higher value indicates a higher risk of encounter, meaning a greater probability of collision.

[0050] The shortest distance between the i-th encountered vessel and the vessel to be analyzed This represents the shortest distance (usually the predicted dynamic distance, not the current static distance) between the i-th encountering vessel and the vessel to be analyzed. A smaller shortest distance indicates a higher risk of collision between the two vessels. Therefore, the shortest distance between the i-th encountering vessel and the vessel to be analyzed... Risk factor of attending the meeting Inversely proportional.

[0051] The relative speed between the i-th encountering vessel and the vessel to be analyzed Let represent the relative speed between the i-th encountering vessel and the vessel to be analyzed. A higher relative speed indicates a faster encounter, and consequently, a greater risk. Therefore, the relative speed between the i-th encountering vessel and the vessel to be analyzed... Risk factor of attending the meeting Proportional.

[0052] The environmental impact factor reflects the degree of influence of external sea conditions on collision risk. The higher the maximum wind speed, maximum wave height, and maximum current velocity, the higher the environmental impact factor. The formula used to calculate the environmental impact factor is as follows: In the formula, As environmental impact factors, At maximum wind speed, Maximum wave height, This represents the maximum water flow velocity in the water area.

[0053] Among them, environmental impact factors This value is used to measure the degree of impact of environmental factors on ship navigation. The higher the value, the greater the adverse impact of the environment on navigation.

[0054] Maximum wind speed This indicates that the area is a circle with the current location of the ship to be analyzed as the center and a radius of... The maximum wind speed that may be encountered within a tested waterway of kilometers. Higher wind speeds generally have a greater impact on the stability and maneuverability of a vessel; therefore, the maximum wind speed... Environmental Influencing Factors Proportional.

[0055] Maximum wave height This indicates the maximum possible wave height within the monitored waters. Higher wave heights can significantly impact a vessel's comfort, safety, and stability; therefore, the maximum wave height is crucial. Environmental Influencing Factors Proportional.

[0056] Maximum water flow velocity in the water area This indicates the maximum current speed within the detected water area. Current speed affects a ship's navigation path and speed, and can potentially lead to loss of control; therefore, the maximum current speed within the water area is crucial. Environmental Influencing Factors Proportional.

[0057] Maximum wind speed and maximum wave height Adding the values ​​and squaring them reflects the combined effects of wind speed and wave height. The squaring of their linear superposition highlights the adverse effects of simultaneous strong winds and waves on ship navigation. The squaring operation makes the effects in extreme cases more significant; wind and waves often occur together, and their impact on ships has a cumulative effect. Therefore, considering them together is reasonable, and the squaring operation better represents their superposition effect.

[0058] Using a logarithmic function to process the square of the water flow velocity represents the nonlinear effect of water flow on ship navigation. A characteristic of the logarithmic function is that the output changes rapidly when the input is small, but tends to level off when the input is large. This reflects that at lower water flow velocities, its impact on ship handling is significant, while at higher speeds, the impact is relatively stable. The effect of water flow is often proportional to the square of the flow velocity. The logarithmic function avoids this overly extreme situation, making the representation of the influence of environmental factors more robust.

[0059] Step 5: Calculate the ship collision risk assessment coefficient based on the macro collision risk assessment value and the micro collision risk assessment value, compare the ship collision risk assessment coefficient with the preset risk assessment threshold, and output the corresponding risk assessment result.

[0060] Based on the obtained ship behavior state index, encounter risk coefficient and environmental impact factor, a micro collision risk assessment value is generated according to the preset micro risk fusion logic; The micro-risk fusion logic includes: enhancing environmental impact factors, superimposing the enhanced environmental impact factors with the smoothed result of the ship behavior state index, and nonlinearly adjusting the superimposed result according to the encounter risk coefficient to obtain a micro-collision risk assessment value. The higher the ship behavior state index, the larger the environmental impact factor, and the higher the encounter risk coefficient, the higher the micro-collision risk assessment value. The formula used to calculate the micro-collision risk assessment value is as follows: In the formula, This is the value for assessing the risk of microscopic collisions.

[0061] Among them, the micro-collision risk assessment value The micro-collision risk is characterized by a comprehensive analysis of ship behavior state index, encounter risk coefficient, and environmental impact factors. The micro-collision risk assessment value is among these factors. The larger the size, the higher the risk of a collision.

[0062] Among them, environmental impact factors This is used to measure the degree of impact of environmental factors on ship navigation. A higher value indicates a greater adverse impact of the environment on navigation. Therefore, the environmental impact factor... Microscopic collision risk assessment value Proportional, in the form of squares This indicates that the impact of environmental conditions (such as wind, waves, and currents) on ship collision risk is often non-linear. For example, when wind speed and wave height reach a certain level, ship maneuverability may decrease sharply. By using square magnification, the formula can more sensitively reflect the accelerating effect of adverse environments on risk.

[0063] Ship behavior status index The index reflects the stability of a ship's navigational state. A higher index indicates greater instability and risk in the ship's operation; therefore, the ship behavior state index... Microscopic collision risk assessment value Proportional, through the logarithmic function It reflects the nonlinear impact of the ship's own operational status on collision risk, characterized by when When the size is small, the contribution is large, and when... When the magnitude is large, the growth gradually slows down. This limits the excessive amplification of overall risk by the instability of ship behavior.

[0064] This indicates the collision risk between the vessel to be analyzed and the vessel that will be encountered. The larger the value, the higher the risk, therefore, it is related to the microscopic collision risk assessment value. Proportional, introduced through the denominator The contribution ratios of environmental factors and ship behavior to overall risk were dynamically adjusted. When the encounter risk is high, the impact of environmental and operational conditions becomes more significant; when the encounter risk is low, the denominator is larger, and the overall risk value is appropriately reduced.

[0065] Since the above has already explained the relationship between the ship behavior state index, encounter risk coefficient, environmental impact factor and collision risk, it will not be repeated here.

[0066] Based on the micro-collision risk assessment value and the macro-collision risk assessment value, the ship collision risk evaluation coefficient is calculated according to the preset fusion logic. The fusion logic is used to weight and fuse the macroscopic and microscopic collision risk assessment values, with the weight corresponding to the microscopic collision risk assessment value being no less than the weight corresponding to the macroscopic collision risk assessment value; the formula used to calculate the ship collision risk evaluation coefficient is as follows: In the formula, For ship collision risk assessment coefficient, and These are the weighting coefficients for the macro-collision risk assessment value and the micro-collision risk assessment value, respectively. and and All are greater than 0.

[0067] Ship collision risk assessment coefficient This integrates micro and macro collision risk assessment values, including the ship collision risk evaluation coefficient. The higher the value, the higher the collision risk and the more dangerous the waters where the ship being analyzed is located.

[0068] Since both macroscopic and microscopic collision risk assessment values ​​indicate a higher risk of collision, these values ​​are correlated with the ship collision risk evaluation coefficient. Proportional, through logarithm This indicates that as the macroscopic collision risk assessment value increases, the ship collision risk evaluation coefficient... The direct proportional effect gradually decreases, preventing over-reliance on historical data. This is achieved through the index. The form of micro-collision risk assessment value is used to evaluate the ship collision risk coefficient. The greater the micro-collision risk assessment value, the more significant the impact on the ship collision risk evaluation coefficient. The impact has increased significantly.

[0069] The description of the direct proportional relationship between the micro-collision risk assessment value and the macro-collision risk assessment value and the collision risk is already provided and will not be repeated here.

[0070] Since the micro-collision risk assessment value takes into account the collision risk during real-time navigation, it is set as follows: and and All are greater than 0.

[0071] The ship collision risk assessment coefficient is compared with a preset risk assessment threshold, and the corresponding risk assessment result is generated based on the comparison result. When the collision risk assessment coefficient is greater than or equal to the risk assessment threshold, it is determined to be a high collision risk, indicating that the water area has a high collision risk and the vessel should leave as soon as possible. When the ship collision risk assessment coefficient is greater than or equal to 40% of the risk assessment threshold and less than the risk assessment threshold, it is judged as a medium collision risk, indicating that there is a collision risk in the waters and one should be more vigilant and navigate with caution. When the ship collision risk assessment coefficient is less than 40% of the risk assessment threshold, it is judged as having low collision risk, indicating that the probability of a collision in this waterway is low and navigation is safe. Specifically, this means: when If the area is deemed to have a high collision risk, it means that the waterway poses a high risk of collision and the vessel should leave as soon as possible. when When the risk level is assessed as moderate, it indicates that there is a risk of collision in the waterway, and caution should be exercised when driving. when When the risk of collision is judged as low, it means that the probability of a collision in this waterway is low, providing safe driving conditions. in A pre-set risk assessment threshold.

[0072] Please see Figure 2 The present invention also provides a ship collision risk assessment system based on macro-micro risk, which is used to execute the above-described ship collision risk assessment method based on macro-micro risk, including: The sample data acquisition module is used to collect historical feature data of ship collision accidents. The historical feature data includes the collision location, management response time, traffic flow density, signal communication coverage and number of collisions in the area affected by the collision location, and constructs a sample dataset. The neural network training module is used to train a neural network prediction model based on the sample dataset, taking collision location, traffic flow density and signal communication coverage as inputs, and management response time and number of collisions as outputs, to obtain a ship collision accident risk prediction model. The macro-collision risk analysis module is used to input the location information, traffic flow density and signal communication coverage of the waters where the vessel to be analyzed is located into the vessel collision accident risk prediction model, output the management response time prediction value and the collision occurrence number prediction value, and calculate the macro-collision risk assessment value. The micro-collision risk analysis module is used to collect navigation behavior parameters of the ship to be analyzed, encounter ship behavior parameters and environmental impact parameters, generate ship behavior state index, encounter risk coefficient and environmental impact factor, and calculate micro-collision risk assessment value. The collision risk assessment module is used to calculate the ship collision risk assessment coefficient based on the macro collision risk assessment value and the micro collision risk assessment value, compare the ship collision risk assessment coefficient with the preset risk assessment threshold, and output the corresponding risk assessment result.

[0073] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0074] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0075] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0076] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for assessing ship collision risk based on macro-micro risk, characterized in that, The specific steps include: Historical characteristic data of ship collision accidents are collected, including collision location, management response time, traffic flow density, signal and communication coverage, and number of collisions in the affected area corresponding to the collision location, to construct a sample dataset; Based on the sample dataset, with collision location, traffic flow density and signal communication coverage as inputs, and management response time and number of collisions as outputs, a neural network prediction model is trained to obtain a ship collision accident risk prediction model. The location information, traffic flow density, and signal communication coverage of the waters where the vessel to be analyzed is located are input into the vessel collision accident risk prediction model, which outputs the predicted management response time and the predicted number of collisions, and calculates the macro-collision risk assessment value. Collect navigation behavior parameters of the ship to be analyzed, encounter ship behavior parameters and environmental impact parameters, generate ship behavior state index, encounter risk coefficient and environmental impact factor, and calculate micro collision risk assessment value; The ship collision risk assessment coefficient is calculated based on the macro-collision risk assessment value and the micro-collision risk assessment value, and the ship collision risk assessment coefficient is compared with the preset risk assessment threshold to output the corresponding risk assessment result.

2. The ship collision risk assessment method based on macro-micro risk according to claim 1, characterized in that: The collision location is represented by geographic coordinates, and the collision impact area is determined with the collision location as the center and a preset radius range. Within the collision impact area, traffic flow density, signal communication coverage, and the number of collisions were statistically analyzed. The signal communication coverage is characterized by a combination of signal reception quality, communication network availability time, and signal reception strength within the sea area. The management response time is the time from the occurrence of the accident to the start of intervention by the management agency.

3. The ship collision risk assessment method based on macro-micro risk according to claim 2, characterized in that: A neural network prediction model is constructed using a Long Short-Term Memory (LSTM) network. The model uses Tanh as the activation function, Adam as the optimization algorithm, and is configured with a three-layer network structure, 200 iterations, a learning rate of 0.001, a batch size of 32, 100 training iterations, 256 batch processing units, and 32 hidden layer neurons. The model inputs include the location of the ship collision, traffic flow density, and signal communication coverage, while the management response time and the number of collisions are used as the model outputs.

4. The ship collision risk assessment method based on macro-micro risk according to claim 2, characterized in that: Based on the obtained management response time prediction and ship collision frequency prediction, combined with the traffic flow density and signal communication coverage of the waters where the ship to be analyzed is located, a macro collision risk assessment value is generated according to the preset macro risk fusion rules. The macro risk fusion rules are used to comprehensively reflect the frequency of ship collisions, regional traffic pressure, accident handling efficiency and communication support level. The higher the predicted number of ship collisions, the higher the traffic flow density, the longer the predicted management response time, and the lower the signal communication coverage, the higher the macro collision risk assessment value. The signal communication coverage rate is determined comprehensively based on the signal reception quality, communication network availability time, and signal reception strength within the sea area. The better the signal reception quality, the longer the communication network availability time, and the higher the signal reception strength, the higher the signal communication coverage rate.

5. The ship collision risk assessment method based on macro-micro risk according to claim 4, characterized in that: Real-time navigation behavior parameters, encounter vessel behavior parameters, and environmental impact parameters of the vessel to be analyzed are collected. Based on the navigation behavior parameters, encounter vessel behavior parameters, and environmental impact parameters, a vessel behavior state index, an encounter risk coefficient, and an environmental impact factor are generated. The vessel behavior state index is used to reflect the degree of drastic change in the navigation state of the vessel to be analyzed. The greater the rate of change in speed and the rate of change in course, the higher the vessel behavior state index. The encounter risk coefficient is used to reflect the degree of danger of approach between the ship being analyzed and the encountering ship. The smaller the shortest distance and the greater the relative speed, the higher the encounter risk coefficient. The result corresponding to the most serious risk among the encountering ships is taken as the encounter risk coefficient. The environmental impact factor is used to reflect the degree of impact of external sea conditions on collision risk, and the higher the maximum wind speed, maximum wave height and maximum water flow velocity, the higher the environmental impact factor.

6. The ship collision risk assessment method based on macro-micro risk according to claim 5, characterized in that: Based on the obtained ship behavior state index, encounter risk coefficient and environmental impact factor, a micro collision risk assessment value is generated according to the preset micro risk fusion logic; The micro-risk fusion logic includes: enhancing the environmental impact factors, superimposing the enhanced environmental impact factors with the smoothed result of the ship behavior state index, and nonlinearly adjusting the superimposed result according to the encounter risk coefficient to obtain the micro-collision risk assessment value. The higher the ship behavior state index, the greater the environmental impact factor, and the higher the encounter risk coefficient, the higher the micro-collision risk assessment value.

7. The ship collision risk assessment method based on macro-micro risk according to claim 6, characterized in that: Based on the micro-collision risk assessment value and the macro-collision risk assessment value, the ship collision risk evaluation coefficient is calculated according to the preset fusion logic. The fusion logic is used to perform weighted fusion of macro-collision risk assessment values ​​and micro-collision risk assessment values, and the weight corresponding to the micro-collision risk assessment value is not lower than the weight corresponding to the macro-collision risk assessment value. The ship collision risk assessment coefficient is compared with a preset risk assessment threshold, and the corresponding risk assessment result is generated based on the comparison result. When the collision risk assessment coefficient is greater than or equal to the risk assessment threshold, it is determined to be a high collision risk, indicating that the water area has a high collision risk and the vessel should leave as soon as possible. When the ship collision risk assessment coefficient is greater than or equal to 40% of the risk assessment threshold and less than the risk assessment threshold, it is judged as a medium collision risk, indicating that there is a collision risk in the waters and one should be more vigilant and navigate with caution. When the collision risk assessment coefficient is less than 40 percent of the risk assessment threshold, it is judged as low collision risk, indicating that the probability of a collision in the waters is low and it is safe to navigate.

8. A ship collision risk assessment system based on macro-micro risk, characterized in that: The aforementioned ship collision risk assessment system based on macro-micro risk is used to execute the ship collision risk assessment method based on macro-micro risk as described in any one of claims 1-7, including: The sample data acquisition module is used to collect historical feature data of ship collision accidents. The historical feature data includes the collision location, management response time, traffic flow density, signal communication coverage and number of collisions in the area affected by the collision location, and constructs a sample dataset. The neural network training module is used to train a neural network prediction model based on the sample dataset, taking collision location, traffic flow density and signal communication coverage as inputs, and management response time and number of collisions as outputs, to obtain a ship collision accident risk prediction model. The macro-collision risk analysis module is used to input the location information, traffic flow density and signal communication coverage of the waters where the vessel to be analyzed is located into the vessel collision accident risk prediction model, output the management response time prediction value and the collision occurrence number prediction value, and calculate the macro-collision risk assessment value. The micro-collision risk analysis module is used to collect navigation behavior parameters of the ship to be analyzed, encounter ship behavior parameters and environmental impact parameters, generate ship behavior state index, encounter risk coefficient and environmental impact factor, and calculate micro-collision risk assessment value. The collision risk assessment module is used to calculate the ship collision risk assessment coefficient based on the macro collision risk assessment value and the micro collision risk assessment value, compare the ship collision risk assessment coefficient with the preset risk assessment threshold, and output the corresponding risk assessment result.

Images