Navigation mark early warning method and system based on multi-factor fuzzy comprehensive judgment

Abstract

The invention provides a navigation mark early warning method and system based on multi-factor fuzzy comprehensive judgment. The method comprises the following steps: firstly, determining key factors influencing navigation safety, and distributing weights according to the importance of each factor; secondly, environment data are collected in real time through a sensor, data of nearby ships are collected through an automatic identification system (AIS), and cleaning and formatting processing are carried out; thirdly, on the basis of the judgment factors and the weights, constructing a fuzzy relation matrix, inputting data collected in real time, processing the data through a fuzzy logic algorithm, and outputting a fuzzy vector; and finally, according to the fuzzy vector, the current risk level is evaluated, and an early warning signal is generated and broadcasted through communication equipment. The system comprises a data collection module, a data preprocessing module, a fuzzification module, a fuzzy reasoning module, an optimization module, a defuzzification module and an early warning generation module. The fuzzy relation matrix is optimized by combining the genetic algorithm and the particle swarm optimization algorithm, and the accuracy and robustness of the early warning system are remarkably improved.

Description

Technical Field

[0001] The invention belongs to the field of navigation mark warning, and specifically relates to a navigation mark warning method and system based on multi-factor fuzzy comprehensive judgment. Background Art

[0002] In the management of maritime navigation safety, beacons are an important navigation aid, and the accuracy of their warning function is directly related to the safe operation of ships and the efficiency of maritime traffic. With the development of science and technology, the beacon system has gradually evolved from a traditional static sign to a dynamic management system integrating high technology. Especially with the rapid development of information technology and sensor technology, the collection and processing of real-time data has become the key to improving the performance of the beacon warning system. In the prior art, the multi-factor fuzzy comprehensive evaluation method has been gradually applied to the beacon warning system, which comprehensively considers various ocean and meteorological conditions, such as waves, wind speed WS, visibility V and other factors, to achieve the purpose of improving the accuracy and real-time performance of the warning.

[0003] Although the existing navigation mark warning system is already able to process a variety of data, it still has shortcomings in terms of real-time performance and accuracy. For example, when processing large-scale real-time data, the traditional fuzzy comprehensive evaluation method often delays the response due to the limited algorithm processing capacity, and it is difficult to adapt to the rapidly changing marine environment. In addition, the flexibility and scalability of the existing system in data integration and processing are also relatively limited, which limits its application effect under complex marine conditions. In response to these problems, the present invention proposes an optimization method, which not only improves the speed and accuracy of data processing by integrating more efficient real-time environmental sensors and adopting advanced data processing algorithms, but also enhances the system's ability to adapt to complex environmental changes. For example, by introducing machine learning technology, the system can automatically adjust the evaluation criteria based on historical data to more accurately reflect changes in the real-time environment, thereby significantly improving the accuracy and timeliness of the warning. Summary of the invention

[0004] Although the existing navigation mark warning system is already able to process a variety of data, it still has deficiencies in terms of real-time performance and accuracy. For example, when processing large-scale real-time data, the traditional fuzzy comprehensive evaluation method often delays response due to limited algorithm processing capabilities, and is difficult to adapt to the rapidly changing marine environment. In addition, the flexibility and scalability of the existing system in data integration and processing are also relatively limited, which limits its application effect under complex marine conditions. In response to these problems, the present invention proposes an optimization method that not only improves the speed and accuracy of data processing, but also enhances the system's ability to adapt to complex environmental changes by integrating more efficient real-time environmental sensors and adopting advanced data processing algorithms.

[0005] To achieve the above object, the present invention provides the following technical solution, a navigation mark warning method based on multi-factor fuzzy comprehensive judgment, including the following specific technical steps:

[0006] The first step is to identify the key factors that affect navigation safety and assign corresponding weights to each factor according to its importance in safe navigation.

[0007] The second step is to use sensors to collect environmental data in real time. The position P, speed S, heading D, and ship density SD of nearby ships are collected through the automatic identification system (AIS). The collected data are cleaned and formatted to meet the needs of subsequent processing.

[0008] The third step is to construct a fuzzy relationship matrix based on the evaluation factors and weights. The matrix describes the contribution of each factor to the warning level. The real-time collected data is input into the fuzzy relationship matrix, processed by the fuzzy logic algorithm, and the fuzzy output vector B is output;

[0009] In the fourth step, based on the fuzzy output vector B, the system evaluates the current risk level and generates corresponding warning signals, which are broadcast to all relevant ships and maritime management agencies through the navigation system and other communication equipment.

[0010] As a preferred technical solution for a navigation mark warning method based on multi-factor fuzzy comprehensive judgment, the first step includes the following specific steps:

[0011] S11. Collecting navigation safety related data from weather stations and ship automatic identification systems (AIS), wherein the navigation safety related data includes environmental condition data and navigation behavior data; wherein the environmental condition data includes wind speed WS, wave height WH and visibility V, and the navigation behavior data includes position P, speed S, heading D, and ship density SD;

[0012] S12. Discuss and analyze the collected environmental data and navigation data, preliminarily identify potential key factors affecting navigation safety, conduct qualitative analysis on the preliminarily identified key factors, evaluate the potential impact of each factor on navigation safety, and use risk assessment methods to further screen and confirm key factors directly related to navigation safety;

[0013] S13. Use the Delphi method to score the relative importance of each key factor identified, with a scoring standard of 1 to 10. Collect and summarize the scoring results and calculate the average importance score of each factor; normalize the average score of each factor to ensure that the sum of the weights of all factors is 1;

[0014] S14. Apply the preliminarily determined weights in a simulation environment, simulate different navigation scenarios, evaluate the prediction accuracy and response speed of the early warning system, adjust the weights according to the simulation results, and ensure that the weight distribution reflects the safety risks in actual navigation to the greatest extent. The adjusted and verified weights are finally determined and applied to the fuzzy comprehensive evaluation module of the navigation warning system.

[0015] As a preferred technical solution of a navigation mark warning method based on multi-factor fuzzy comprehensive judgment, in the second step, the cleaning and formatting of the collected data includes the following specific steps:

[0016] S21. Check the timestamp of the data to ensure the continuity and consistency of the timestamp;

[0017] S22. Use the isolation forest algorithm to detect outliers in environmental data and ship data. The anomaly detection formula is I(x) = 2 -E(h(x)) / e(n) , where I(x) is the anomaly score of sample x, E(h(x)) is the expected tree height of sample x, and c(n) is the adjustment coefficient;

[0018] S23. For missing points in time series data, linear interpolation is performed using adjacent data. The interpolation formula is: Among them, t1 and t2 are the timestamps of adjacent data points, f(t 1 ) and f(t 2 ) is the corresponding data value; for missing data that cannot be supplemented by interpolation, the average value of historical data in the past 30 days is used to supplement to ensure data integrity;

[0019] S24. Convert data from different sources and types into a unified format and integrate the cleaned environmental data and AIS data to form a comprehensive data record including timestamps, environmental parameters and vessel parameters.

[0020] 4. The navigation mark warning method based on multi-factor fuzzy comprehensive judgment according to claim 1 is characterized in that the third step includes the following specific steps:

[0021] S31. Set the warning level, L is the low risk warning level, M is the medium risk warning level, H is the high risk warning level, define the membership function for each evaluation factor, and map the specific value of the factor to the fuzzy set;

[0022] S32. Organize the contribution values ​​of each evaluation factor into a fuzzy relationship matrix R,

[0023]

[0024] Among them, the element r in the matrix ij It represents the contribution of the i-th factor to the j-th warning level;

[0025] S33. Input the specific numerical values ​​of the data collected by the sensor and the AIS system into the membership function, convert them into fuzzy values, and organize the fuzzified real-time data into an input vector A;

[0026] S34. The fuzzy input vector A is operated with the fuzzy relation matrix R by a fuzzy logic algorithm to obtain an output fuzzy output vector B:

[0027] B=A·R

[0028] Among them, B represents the fuzzy membership of different risk levels.

[0029] As a preferred technical solution for the navigation mark warning method based on multi-factor fuzzy comprehensive judgment, the fuzzy relationship matrix R is optimized by combining the genetic algorithm (GA) and the particle swarm optimization algorithm (PSO). The specific steps include the following:

[0030] S321. Define the particle swarm size and initialization parameters, randomly generate the initial particle position and velocity, and when initializing the population, use the optimal solution and some of the better solutions obtained by PSO as the initial population of GA;

[0031] S32. Calculate the fitness value of each particle, update the particle's speed and position according to the particle's current speed, individual best position and global best position, compare the current particle's fitness value, update the individual and global best positions, and determine whether the maximum number of iterations of PSO has been reached or the fitness value has reached a predetermined threshold. If the termination condition is met, end the PSO stage; otherwise, continue iteration;

[0032] S33. Use the optimal solution and some of the better solutions in the PSO stage as the initial population of GA, select individuals for crossover and mutation operations according to the fitness value, generate new offspring individuals, add the offspring individuals to the population, retain the best individuals, and determine whether the maximum number of iterations of GA is reached or the fitness value reaches the predetermined threshold; if the termination condition is met, end the GA stage; otherwise, continue iterating.

[0033] 6. The navigation mark warning method based on multi-factor fuzzy comprehensive judgment according to claim 1 is characterized in that, in the fourth step, according to the fuzzy output vector B, the system evaluates the current risk level, and the specific steps are:

[0034] S41. Use a multivariable fuzzy comprehensive evaluation model to integrate multiple fuzzy output vectors to obtain a comprehensive risk assessment vector C; where: Among them, α i is the weight of each evaluation factor, B i is the i-th fuzzy output vector,

[0035] S42. Dynamically adjust the weights of different risk levels according to real-time data and environmental changes, and calculate the comprehensive risk level y after dynamic weight adjustment. Among them, β i is the dynamic weight of different risk levels, C i is the i-th component in the comprehensive risk assessment vector;

[0036] S43. Use the defuzzification method to calculate the specific risk level value z, where γ i is the weight parameter in the defuzzification process, y i is the component in the comprehensive risk assessment vector;

[0037] S44. Divide the specific risk level value z into several intervals. When the value range of z is [0,2), the risk level is low risk; when the value range of z is [2,4), the risk level is medium risk; when the value range of z is [4,5], the risk level is high risk.

[0038] A navigation mark warning system based on multi-factor fuzzy comprehensive judgment includes: a data collection module, which is responsible for collecting environmental data and ship data in real time to ensure that the system obtains accurate input data; a data preprocessing module, which cleans and formats the collected data to ensure the accuracy and consistency of the data; a fuzzification module, which fuzzifies the processed data and converts specific values ​​into fuzzy values; a fuzzy reasoning module, which performs reasoning based on fuzzy logic to obtain a comprehensive safety rating; an optimization module, which uses an optimization algorithm to optimize the parameters in the fuzzy relationship matrix RRR to improve the sensitivity and robustness of the system; a defuzzification module, which converts the fuzzy output vector B obtained by the fuzzy reasoning module into a specific safety rating value; and an early warning generation module, which generates corresponding early warning signals and early warning information according to the comprehensive safety rating results.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] Improve warning accuracy: By combining genetic algorithm (GA) and particle swarm optimization algorithm (PSO), the fuzzy relationship matrix is ​​dynamically optimized, so that the early warning system can maintain efficient and accurate risk assessment under different environmental conditions, thereby improving the accuracy of early warning.

[0041] Enhanced system robustness: The method of the present invention can still provide stable and reliable early warning results under extreme weather and complex navigation conditions, indicating that it has superior robustness and can effectively cope with rapidly changing environments and abnormal situations.

[0042] Real-time data processing: Use sensors and AIS systems to collect environmental data and ship data in real time, and clean and format them through data preprocessing modules to ensure data continuity and consistency, so that the system can adapt to environmental changes in real time and improve the timeliness and effectiveness of early warnings.

[0043] Multi-factor comprehensive evaluation: The method of the present invention determines multiple key factors that affect navigation safety and constructs a fuzzy relationship matrix based on these factors to conduct a multi-factor comprehensive evaluation. Compared with a single factor or simple weighted method, it can evaluate navigation risks more comprehensively and scientifically.

[0044] Strong system optimization capability: Combining the advantages of GA and PSO algorithms, using the global search capability of GA and the local search capability of PSO, the fuzzy relationship matrix is ​​optimized, which improves the sensitivity and robustness of the system and avoids the problem that a single algorithm is prone to fall into the local optimal solution.

[0045] Strong model interpretability: Compared with traditional machine learning prediction methods, the method of the present invention has higher stability and accuracy when dealing with data and environmental changes. At the same time, through the fuzzy logic algorithm, the prediction process of the model is more interpretable and can clearly explain the reasons for the prediction.

[0046] Reduce the misjudgment rate: Under complex and extreme conditions, the method of the present invention significantly reduces the misjudgment rate, avoids the warning lag or misjudgment phenomenon that is prone to occur in the prior art under conditions of high wind speed, high wave height, etc., and improves navigation safety.

[0047] Wide scope of application: The method of the present invention is applicable to various environmental conditions and navigation scenarios, has strong versatility and adaptability, can play an important role in practical applications, and improves the safety and reliability of navigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 is a flow chart of this application;

[0049] Figure 2 It is the principle diagram of the navigation aid warning system; DETAILED DESCRIPTION

[0050] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent replacements for some of the technical features therein.

[0051] Example 1

[0052] Although the existing navigation warning system is able to process a variety of data, it still has shortcomings in terms of real-time and accuracy. Figure 1 As shown, the present invention provides a navigation mark warning method based on multi-factor fuzzy comprehensive judgment, including the following specific technical steps:

[0053] The first step is to identify the key factors that affect navigation safety and assign corresponding weights to each factor according to its importance in safe navigation.

[0054] The second step is to use sensors to collect environmental data in real time. The position P, speed S, heading D, and ship density SD of nearby ships are collected through the automatic identification system (AIS). The collected data are cleaned and formatted to meet the needs of subsequent processing.

[0055] The third step is to construct a fuzzy relationship matrix based on the evaluation factors and weights. The matrix describes the contribution of each factor to the warning level. The real-time collected data is input into the fuzzy relationship matrix, processed by the fuzzy logic algorithm, and the fuzzy output vector B is output;

[0056] In the fourth step, based on the fuzzy output vector B, the system evaluates the current risk level and generates corresponding warning signals, which are broadcast to all relevant ships and maritime management agencies through the navigation system and other communication equipment.

[0057] In this embodiment, it should be specifically explained that determining the key factors affecting navigation safety and assigning corresponding weights according to the importance of each factor in safe navigation includes the following specific steps:

[0058] S11. Collect navigation safety-related data from meteorological stations and ship automatic identification systems (AIS), wherein the navigation safety-related data include environmental condition data and navigation behavior data; wherein the environmental condition data include wind speed WS, wave height WH and visibility V, wind speed (WS) refers to the speed at which air flows per unit time, and is an important indicator for judging wind strength; wave height (WH) refers to the vertical distance from the trough to the crest, and is an important parameter for measuring sea conditions; visibility (V) refers to the maximum distance at which a target can be clearly seen during the day, and is an important factor affecting navigation safety; navigation behavior data include position (P), speed (S), heading (D) and ship density (SD); position (P) is the specific coordinate of a ship at sea, which is used to determine the real-time position of the ship; speed (S) refers to the distance traveled by a ship per unit time, and is an important parameter reflecting the ship's driving status; heading (D) refers to the direction angle of the ship's advance, and is an important reference for navigation; ship density (SD) refers to the number of ships in a unit sea area, and is an important indicator for assessing the degree of congestion in the sea area. By collecting these data in real time through sensors and AIS systems, we can fully understand the marine environment and the operating status of the ship, ensure the timeliness and accuracy of the data, and provide reliable data support for subsequent navigation safety assessments;

[0059] S12. Discuss and analyze the collected environmental data and navigation data, preliminarily identify potential key factors affecting navigation safety, conduct qualitative analysis on the preliminarily identified key factors, evaluate the potential impact of each factor on navigation safety, and use risk assessment methods to further screen and confirm key factors directly related to navigation safety;

[0060] S13. Use the Delphi method to score the relative importance of each key factor identified. First, an expert review team in the field of navigation safety is formed. The experts score the key factors identified in the early stage based on their professional knowledge and experience. The scoring standard is 1 to 10, with 1 representing the lowest importance and 10 representing the highest importance. The scoring results are collected and summarized, and the average importance score of each factor is calculated. The results are fed back to the expert team. After multiple rounds of scoring and feedback, the experts adjust the scores until the results are stable and a consensus is reached. Then, the average score of each factor is normalized to ensure that the sum of the weights of all factors is 1. The specific method is to divide the average score of each factor by the sum of the average scores of all factors to obtain the normalized weight value. Through the Delphi method, the weights of key factors are scientifically evaluated and allocated, which improves the accuracy and credibility of risk assessment.

[0061] S14. Apply the preliminarily determined weights in a simulation environment, simulate different navigation scenarios, evaluate the prediction accuracy and response speed of the early warning system, adjust the weights according to the simulation results, ensure that the weight distribution reflects the safety risks in actual navigation to the greatest extent, and finally determine the adjusted and verified weights and apply them to the fuzzy comprehensive evaluation module of the navigation mark early warning system; for example, if the response speed of the early warning system is slow under high wind speed conditions, it may be necessary to increase the weight of the wind speed factor, and repeat the simulation and adjustment until the early warning system shows high prediction accuracy and rapid response capabilities in all test scenarios. After multiple rounds of adjustment and verification, finally determine the weight of each key factor, and apply these weights to the fuzzy comprehensive evaluation module of the navigation mark early warning system. In this way, the optimized early warning system can more accurately and timely evaluate and warn of navigation risks and ensure navigation safety.

[0062] It should be noted that in this embodiment, the cleaning and formatting of the collected data includes the following specific steps:

[0063] S21. Check the timestamps of the data to ensure the continuity and consistency of the timestamps, which significantly improves the overall quality of the data;

[0064] S22. Use the isolation forest algorithm to detect outliers in environmental data and ship data. The anomaly detection formula is I(x) = 2 -E(h(x)) / e(n), where I(x) is the anomaly score of sample x, E(h(x)) is the expected tree height of sample x, and c(n) is the adjustment coefficient. The isolation forest algorithm is used to detect and remove outliers to ensure data accuracy and reduce the impact of noise on the analysis results;

[0065] S23. For missing points in time series data, linear interpolation is performed using adjacent data. The interpolation formula is: Among them, t 1 and t 2 is the timestamp of the adjacent data points, f(t 1 ) and f(t 2 ) is the corresponding data value; for missing data that cannot be supplemented by interpolation, the average value of historical data in the past 30 days is used to supplement it to ensure data integrity and improve data utilization;

[0066] S24. Convert data from different sources and types into a unified format to facilitate subsequent processing and analysis, and integrate the cleaned environmental data and AIS data to form a comprehensive data record including timestamps, environmental parameters and vessel parameters.

[0067] The following is a Python code example to implement the above data cleaning and formatting steps:

[0068]

[0069]

[0070]

[0071]

[0072] In this embodiment, it should be specifically explained that based on the evaluation factors and weights, a fuzzy relationship matrix is ​​constructed, which describes the contribution of each factor to the warning level, and the real-time collected data is input into the fuzzy relationship matrix, and the fuzzy output vector B is output through fuzzy logic algorithm processing, including the following specific steps:

[0073] S31. Set the warning level, L is the low risk warning level, M is the medium risk warning level, H is the high risk warning level, define the membership function for each evaluation factor, and map the specific value of the factor to the fuzzy set;

[0074] S32. Organize the contribution values ​​of each evaluation factor into a fuzzy relationship matrix R,

[0075]

[0076] Among them, the element r in the matrix ij It represents the contribution of the i-th factor to the j-th warning level;

[0077] S33. Input the specific numerical values ​​of the data collected by the sensor and the AIS system into the membership function, convert them into fuzzy values, and organize the fuzzified real-time data into an input vector A;

[0078] S34. The fuzzy input vector A is operated with the fuzzy relation matrix R by a fuzzy logic algorithm to obtain an output fuzzy output vector B:

[0079] B=A·R

[0080] Among them, B represents the fuzzy membership of different risk levels.

[0081] It is necessary to explain that by constructing a fuzzy relationship matrix and using fuzzy logic algorithms to process real-time collected data, a comprehensive, real-time and accurate navigation risk assessment and warning can be achieved. This method not only improves the accuracy and robustness of the warning system, but also has a high degree of flexibility and scalability, and can adapt to the needs of different environments and tasks.

[0082] It should be specifically explained in this embodiment that the fuzzy relationship matrix R is optimized by combining the genetic algorithm (GA) and the particle swarm optimization algorithm (PSO), and the specific steps include the following:

[0083] S321. Define the particle swarm size and initialization parameters, randomly generate the initial particle position and velocity, and when initializing the population, use the optimal solution and some of the better solutions obtained by PSO as the initial population of GA;

[0084] S32. Calculate the fitness value of each particle, update the particle's speed and position according to the particle's current speed, individual best position and global best position, compare the current particle's fitness value, update the individual and global best positions, and determine whether the maximum number of iterations of PSO has been reached or the fitness value has reached a predetermined threshold. If the termination condition is met, end the PSO stage; otherwise, continue iteration;

[0085] S33. Use the optimal solution and some of the better solutions in the PSO stage as the initial population of GA, select individuals for crossover and mutation operations according to the fitness value, generate new offspring individuals, add the offspring individuals to the population, retain the best individuals, and determine whether the maximum number of iterations of GA is reached or the fitness value reaches the predetermined threshold; if the termination condition is met, end the GA stage; otherwise, continue iterating.

[0086] It needs to be explained that the optimization of the fuzzy relationship matrix R by combining the genetic algorithm (GA) and the particle swarm optimization algorithm (PSO) significantly improves the performance and efficiency of the early warning system. GA provides a powerful global search capability through selection, crossover and mutation operations to prevent falling into the local optimal solution, while PSO achieves rapid convergence and fine optimization through information sharing and dynamic adjustment between individuals. Combining the two algorithms not only speeds up the convergence speed, but also increases the probability of finding the global optimal solution. Through the diversity of GA and the collaborative search of PSO, the system effectively avoids the premature convergence phenomenon and enhances the adaptability and robustness to complex multidimensional optimization problems. In addition, the optimized fuzzy relationship matrix can more accurately reflect the actual situation, improving the accuracy and real-time performance of the early warning system. In short, the combination of GA and PSO provides an efficient, stable and intelligent optimization solution for the early warning system, significantly improving the early warning capability of the system in complex environments.

[0087] Please refer to Figure 2 As shown, a navigation mark warning system based on multi-factor fuzzy comprehensive judgment includes: a data collection module, which is responsible for collecting environmental data and ship data in real time to ensure that the system obtains accurate input data; a data preprocessing module, which cleans and formats the collected data to ensure the accuracy and consistency of the data; a fuzzification module, which fuzzifies the processed data and converts specific values ​​into fuzzy values; a fuzzy reasoning module, which performs reasoning based on fuzzy logic to obtain a comprehensive safety rating; an optimization module, which uses an optimization algorithm to optimize the parameters in the fuzzy relationship matrix R to improve the sensitivity and robustness of the system; a defuzzification module, which converts the fuzzy output vector B obtained by the fuzzy reasoning module into a specific safety rating value; and an early warning generation module, which generates corresponding early warning signals and early warning information according to the comprehensive safety rating results.

[0088] Example 2

[0089] In order to verify the accuracy of the navigation mark warning method based on multi-factor fuzzy comprehensive judgment, a detailed experiment was designed and implemented. The experiment includes the following steps:

[0090] Determine key factors and assign weights: Determine the key factors that affect navigation safety, including wind speed (WS), wave height (WH), visibility (V), position (P), speed (S), course (D) and ship density (SD). The Delphi method is used to assess the potential impact of each factor on navigation safety and determine its weight.

[0091] Data collection and preprocessing: Install high-precision meteorological sensors and automatic ship identification systems (AIS) to collect real-time data such as wind speed, wave height, visibility, ship position, speed, heading and density. Data preprocessing steps include: checking data timestamps to ensure continuity and consistency; using the isolation forest algorithm to detect outliers and perform linear interpolation to complete missing points; converting data from different sources into a unified format to form a comprehensive data record.

[0092] Constructing fuzzy relationship matrix: Based on the evaluation factors and weights, construct the fuzzy relationship matrix RRR to describe the contribution of each factor to the warning level. The data collected in real time is input into the fuzzy relationship matrix, processed by the fuzzy logic algorithm, and the fuzzy output vector BBB is output.

[0093] Assess risk level and generate warning signals: Based on the fuzzy output vector BBB, the system assesses the current risk level and generates corresponding warning signals, which are broadcast to all relevant ships and maritime management agencies through the navigation system and communication equipment.

[0094] Optimize the fuzzy relationship matrix: Combining the genetic algorithm (GA) and the particle swarm optimization algorithm (PSO), the fuzzy relationship matrix RRR is dynamically optimized to ensure the efficiency and accuracy of the system under different environmental conditions.

[0095] Table 1 is the experimental data of Example 2

[0096]

[0097] It can be clearly seen from Table 1 that the method of the invention is better than the prior art in dealing with different environments and navigation conditions. The genetic algorithm searches for the global optimal solution by simulating the natural selection process. Although GA performs well in global search, its local search ability is weak and it is easy to fall into the local optimum in complex environments. For example, under the conditions of wind speed of 6.3m / s and wave height of 0.9m, the risk level predicted by GA is high, while the actual risk level is medium. This shows that GA tends to overestimate the risk in some complex situations. PSO searches for the optimal solution through information sharing and collaborative search between individuals. PSO has the characteristics of fast convergence, but is not as good as GA in global search ability. For example, under the conditions of wind speed of 6.2m / s and wave height of 0.9m, the risk level predicted by PSO is high, while the actual risk level is medium. When PSO processes high volatility data, overfitting problems may occur. The machine learning model is trained using historical data for risk prediction. The machine learning algorithm performs well when processing large-scale data, but its prediction ability for unseen data or extreme situations is poor. For example, under the conditions of wind speed of 6.1m / s and wave height of 0.8m, the risk level predicted by machine learning is medium, and the actual risk level is also medium. Although the prediction results of machine learning are relatively accurate, the robustness and interpretability of its model are poor. In contrast, the inventive method dynamically optimizes the fuzzy relationship matrix by combining genetic algorithm and particle swarm optimization algorithm, so that the system can maintain high warning accuracy and stability. For example, under the conditions of wind speed of 6.1m / s and wave height of 0.7m, visibility is maintained at 9.0km, speed is 14.5kn, heading is 280°, and ship density is 3.0ships / km2, the predicted risk level is medium, and the actual risk level is also medium. This shows that the method of the present invention exhibits higher stability and accuracy when processing data and environmental changes.

[0098] Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A navigation mark warning method based on multi-factor fuzzy comprehensive judgment, characterized in that: The specific technical steps include the following: The first step is to identify the key factors that affect navigation safety and assign corresponding weights to each factor according to its importance in safe navigation. The second step is to use sensors to collect environmental data in real time. The position P, speed S, heading D, and ship density SD of nearby ships are collected through the automatic identification system (AIS). The collected data are cleaned and formatted to meet the needs of subsequent processing. The third step is to construct a fuzzy relationship matrix based on the evaluation factors and weights. The matrix describes the contribution of each factor to the warning level. The real-time collected data is input into the fuzzy relationship matrix, processed by the fuzzy logic algorithm, and the fuzzy output vector B is output; In the fourth step, based on the fuzzy output vector B, the system evaluates the current risk level and generates corresponding warning signals, which are broadcast to all relevant ships and maritime management agencies through the navigation system and other communication equipment.

2. The navigation mark warning method based on multi-factor fuzzy comprehensive judgment according to claim 1 is characterized in that: The first step includes the following specific steps: S11. Collecting navigation safety related data from weather stations and ship automatic identification systems (AIS), wherein the navigation safety related data includes environmental condition data and navigation behavior data; wherein the environmental condition data includes wind speed WS, wave height WH and visibility V, and the navigation behavior data includes position P, speed S, heading D, and ship density SD; S12. Discuss and analyze the collected environmental data and navigation data, preliminarily identify potential key factors affecting navigation safety, conduct qualitative analysis on the preliminarily identified key factors, evaluate the potential impact of each factor on navigation safety, and use risk assessment methods to further screen and confirm key factors directly related to navigation safety; S13. Use the Delphi method to score the relative importance of each key factor identified, with a scoring standard of 1 to 10. Collect and summarize the scoring results and calculate the average importance score of each factor; normalize the average score of each factor to ensure that the sum of the weights of all factors is 1; S14. Apply the preliminarily determined weights in a simulation environment, simulate different navigation scenarios, evaluate the prediction accuracy and response speed of the early warning system, adjust the weights according to the simulation results, and ensure that the weight distribution reflects the safety risks in actual navigation to the greatest extent. The adjusted and verified weights are finally determined and applied to the fuzzy comprehensive evaluation module of the navigation warning system.

3. The navigation mark warning method based on multi-factor fuzzy comprehensive judgment according to claim 1 is characterized in that: In the second step, the cleaning and formatting of the collected data includes the following specific steps: S21. Check the timestamp of the data to ensure the continuity and consistency of the timestamp; S22. Use the isolation forest algorithm to detect outliers in environmental data and ship data. The anomaly detection formula is I(x) = 2 -E(h(x)) / e(n) , where I(x) is the anomaly score of sample x, E(h(x)) is the expected tree height of sample x, and C(n) is the adjustment coefficient; S23. For missing points in time series data, linear interpolation is performed using adjacent data. The interpolation formula is: Where t1 and t2 are the timestamps of adjacent data points, and f(t1) and f(t2) are the corresponding data values. For missing data that cannot be filled in by interpolation, the mean of historical data in the past 30 days is used to fill in the missing data to ensure data integrity. S24. Convert data from different sources and types into a unified format and integrate the cleaned environmental data and AIS data to form a comprehensive data record including timestamps, environmental parameters and vessel parameters.

4. The navigation mark warning method based on multi-factor fuzzy comprehensive judgment according to claim 1 is characterized in that: The third step includes the following specific steps: S31. Set the warning level, L is the low risk warning level, M is the medium risk warning level, H is the high risk warning level, define the membership function for each evaluation factor, and map the specific value of the factor to the fuzzy set; S32. Organizing the contribution values ​​of each evaluation factor into a fuzzy relationship matrix R; S33. Input the specific numerical values ​​of the data collected by the sensor and the AIS system into the membership function, convert them into fuzzy values, and organize the fuzzified real-time data into an input vector A; S34. The fuzzy input vector A is operated with the fuzzy relation matrix R by a fuzzy logic algorithm to obtain an output fuzzy output vector B: B=A·R Among them, B represents the fuzzy membership of different risk levels.

5. The navigation mark warning method based on multi-factor fuzzy comprehensive judgment according to claim 4 is characterized in that: The fuzzy relationship matrix R is optimized by combining the genetic algorithm (GA) and the particle swarm optimization algorithm (PSO). The specific steps include the following: S321. Define the particle swarm size and initialization parameters, randomly generate the initial particle position and velocity, and when initializing the population, use the optimal solution and some of the better solutions obtained by PSO as the initial population of GA; S32. Calculate the fitness value of each particle, update the particle's speed and position according to the particle's current speed, individual best position and global best position, compare the current particle's fitness value, update the individual and global best positions, and determine whether the maximum number of iterations of PSO has been reached or the fitness value has reached a predetermined threshold. If the termination condition is met, end the PSO stage; otherwise, continue iteration; S33. Use the optimal solution and some of the better solutions in the PSO stage as the initial population of GA, select individuals for crossover and mutation operations according to the fitness value, generate new offspring individuals, add the offspring individuals to the population, retain the best individuals, and determine whether the maximum number of iterations of GA is reached or the fitness value reaches the predetermined threshold; if the termination condition is met, end the GA stage; otherwise, continue iterating.

6. The navigation mark warning method based on multi-factor fuzzy comprehensive judgment according to claim 1 is characterized in that: In the fourth step, according to the fuzzy output vector B, the system evaluates the current risk level. The specific steps are: S41. Use a multivariable fuzzy comprehensive evaluation model to integrate multiple fuzzy output vectors to obtain a comprehensive risk assessment vector C; where: Among them, α i is the weight of each evaluation factor, B i is the i-th fuzzy output vector, S42. Dynamically adjust the weights of different risk levels according to real-time data and environmental changes, and calculate the comprehensive risk level y after dynamic weight adjustment. Among them, β i is the dynamic weight of different risk levels, C i is the i-th component in the comprehensive risk assessment vector; S43. Use the defuzzification method to calculate the specific risk level value z, where γ i is the weight parameter in the defuzzification process, y i is the component in the comprehensive risk assessment vector; S44. Divide the specific risk level value z into several intervals. When the value range of z is [0,2), the risk level is low risk; when the value range of z is [2,4), the risk level is medium risk; when the value range of z is [4,5], the risk level is high risk.

7. A navigation mark warning system based on multi-factor fuzzy comprehensive judgment, based on the navigation mark warning method based on multi-factor fuzzy comprehensive judgment according to any one of claims 1 to 5, characterized in that: include: The data collection module is responsible for collecting environmental data and ship data in real time to ensure that the system obtains accurate input data; Data preprocessing module cleans and formats the collected data to ensure data accuracy and consistency; The fuzzification module fuzzifies the processed data and converts the specific numerical values ​​into fuzzy values; The fuzzy reasoning module performs reasoning based on fuzzy logic to obtain a comprehensive safety rating; The optimization module uses the optimization algorithm to optimize the parameters in the fuzzy relationship matrix R to improve the sensitivity and robustness of the system; A defuzzification module converts the fuzzy output vector B obtained by the fuzzy reasoning module into a specific safety rating value; and The early warning generation module generates corresponding early warning signals and early warning information according to the comprehensive safety rating results.

Images