Path risk assessment method and system based on dynamic weight correction of task context

Abstract

The application discloses a kind of path risk assessment method and system based on task situation dynamic weight correction, the method includes: obtaining candidate path and segmenting;For each section, obtain multiple risk indicators, and carry out credibility mark;Extract original physical characteristics, and calculate risk aggregation weight factor;Hierarchical analysis model is constructed, and relative importance scale is defined;Construct judgment matrix, and calculate corresponding weight, obtain global reference weight matrix;Define positive and negative correlation risk normalization operator, and construct risk mapping model, convert original physical characteristic value into risk value, and carry out risk value correction;According to task level matching situation factor vector, and calculate weight correction vector;Global reference weight matrix is dynamically adjusted, and the comprehensive risk index of each candidate path is obtained by risk aggregation and weighted summation.The method realizes the accurate quantitative evaluation of path comprehensive risk, improves the credibility and robustness of risk assessment result.

Description

Technical Field

[0001] This invention belongs to the field of path planning technology, specifically relating to a path risk assessment method and system based on dynamic weight correction according to task context. Background Technology

[0002] With the evolution of intelligent logistics and military command and control technologies, path planning has shifted from simply finding the shortest distance to minimizing risk. In the context of military mission execution, scientific risk assessment of candidate paths is crucial for ensuring mission success. While existing methods have achieved basic assessment functions, they still have limitations when dealing with multi-dimensional risk indicators, making it difficult to meet the actual needs of diverse missions in complex battlefield environments.

[0003] First, traditional risk assessment models employ a pre-defined static weighting scheme. This fixed weighting ignores the differences in risk sensitivity among different tasks. For example, in time-sensitive mobile missions, the weight of time window risk is adjusted higher than in ordinary missions; while in equipment and material transportation where safety is paramount, road smoothness and safety risks should take precedence. The aforementioned scheme cannot dynamically adjust weights according to mission type, resulting in assessment results lacking mission-specific relevance.

[0004] Secondly, in practice, due to the high complexity of task environments, the impact of the same risk indicator varies at different task levels. Therefore, it is necessary to reasonably adapt the weights of different indicators according to the actual situation. However, existing assessment models lack an effective mechanism to correlate and map task levels, contextual factors, and the sensitivity of specific risk indicators, resulting in a disconnect between assessment results and actual task requirements.

[0005] Finally, existing path risk assessment methods often simply superimpose single-dimensional information such as road geometry or meteorological data, lacking quantitative processing of the data's reliability and failing to construct an evolutionary model from "road segment characteristics" to "comprehensive path risk under task context," thus reducing the scientific rigor and accuracy of risk assessment.

[0006] In summary, existing path risk assessment methods cannot be optimized and adjusted to meet the changing needs of different missions in terms of timeliness, security, concealment, and reliability. This reduces the credibility and robustness of the risk assessment results, resulting in assessments that lack mission relevance, are out of touch with actual needs, and lack scientific rigor, making it difficult to provide reliable support for military mission path planning. Summary of the Invention

[0007] The purpose of this invention is to provide a path risk assessment method based on dynamic weight correction in mission context. It aims to achieve accurate quantitative assessment of the comprehensive risk of candidate paths and optimal path decision-making in military mission execution environment by integrating data quality coefficients and dynamic weight correction mechanisms based on mission level.

[0008] The technical problem to be solved by this invention is achieved through the following technical solution: Firstly, this invention proposes a path risk assessment method based on dynamic weight correction according to task context, including: S1: Obtain a set of candidate paths and divide each path into several continuous road segments; for each road segment, obtain multiple risk indicators under different risk categories, and use data quality coefficients to mark credibility. S2: For each road segment, extract the original physical features corresponding to multiple risk indicators and construct a physical feature vector; at the same time, calculate the risk aggregation weight factor based on the length ratio of each road segment. S3: Construct a hierarchical analysis model for comprehensive path risk assessment based on different risk categories and risk indicators, and define the relative importance scale between elements within each level; S4: For each level in the hierarchical analysis model, construct a judgment matrix using the relative importance scale, calculate the weight of each judgment matrix, and then obtain the global baseline weight matrix. S5: Define positive and negative correlation risk normalization operators based on the original physical characteristics, and further construct risk mapping models for various risk indicators to transform the original physical characteristic values ​​into risk values; correct the risk values ​​according to the data quality coefficient to obtain the corrected risk values; S6: Match the context factor vector in the pre-built context factor database according to the current task level, and calculate the weight correction vector in combination with the indicator sensitivity matrix; use the context factor vector and the weight correction vector to adjust the global benchmark weight matrix to obtain the dynamic weight of each road segment under the current task. S7: Based on the risk aggregation weight factor and the corrected risk value of each road segment, risk aggregation is performed, and a weighted sum is performed based on dynamic weights to obtain the comprehensive risk index of each candidate path, so as to select the optimal execution path.

[0009] Secondly, this invention proposes a path risk assessment system based on dynamic weight correction according to task context, used to implement the method proposed in the first aspect of this invention. The system includes: The data acquisition and credibility labeling module is used to acquire a set of candidate paths and divide each path into several continuous road segments; for each road segment, it acquires multiple risk indicators under different risk categories and uses data quality coefficients to label credibility. The feature extraction and road segment weight calculation module is used to extract the original physical features corresponding to multiple risk indicators for each road segment and construct a physical feature vector; at the same time, it calculates the risk aggregation weight factor based on the length ratio of each road segment. The evaluation system modeling and scaling module is used to construct a hierarchical analysis model for comprehensive path risk assessment based on different risk categories and risk indicators, and to define the relative importance scale between elements within each level; The benchmark weight calculation module constructs a judgment matrix for each level in the hierarchical analysis model using a relative importance scale, calculates the weight of each judgment matrix, and then obtains the global benchmark weight matrix. The risk mapping and correction module is used to define positive and negative correlation risk normalization operators based on the original physical characteristics, and further construct risk mapping models for various risk indicators to transform the original physical characteristic values ​​into risk values; the risk values ​​are corrected according to the data quality coefficient to obtain the corrected risk values; The dynamic weight correction module is used to match the context factor vector in the pre-built context factor database according to the current task level, and calculate the weight correction vector in combination with the indicator sensitivity matrix; the global benchmark weight matrix is ​​adjusted using the context factor vector and the weight correction vector to obtain the dynamic weight of each road segment under the current task. The risk comprehensive assessment and route decision module is used to aggregate risks based on the risk aggregation weight factors and the corrected risk values ​​of each road segment, and to perform weighted summation based on dynamic weights to obtain the comprehensive risk index of each candidate route in order to select the optimal execution route.

[0010] The beneficial effects of this invention are: The path risk assessment method based on dynamic weight correction according to task context provided by this invention, on the one hand, constructs a dynamic weight correction mechanism based on task level by pre-setting a context factor database and an indicator sensitivity matrix. This allows the risk assessment criteria to be optimized and adjusted according to changes in the task's requirements for timeliness, safety, concealment, and reliability, greatly enhancing the adaptability of path planning to complex task contexts and overcoming the limitations of static weights in traditional assessment models. On the other hand, to address the uncertainty of external system perception data, a data quality coefficient is introduced and the initial risk value is conservatively corrected, effectively reducing decision-making biases caused by sensor noise or missing information, and significantly improving the credibility and robustness of the risk assessment results. Furthermore, this invention also achieves a precise characterization from local road segment characteristics to a global comprehensive path risk index by fusing multiple risk indicators of different risk categories and combining the analytic hierarchy process (AHP) with a physical feature risk mapping model, providing comprehensive quantitative support for optimal path decision-making for vehicle maneuvers under task execution environments.

[0011] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0012] Figure 1 A flowchart illustrating the path risk assessment method based on dynamic weight correction in task context provided in this embodiment of the invention; Figure 2 This is a schematic diagram of the hierarchical analysis model for comprehensive path risk assessment provided in an embodiment of the present invention; Figure 3 This is a structural block diagram of a path risk assessment system based on dynamic weight correction in task context, provided in an embodiment of the present invention. Detailed Implementation

[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] The first aspect of this invention provides a path risk assessment method based on dynamic weight adjustment according to task context. See also... Figure 1 , Figure 1 This is a flowchart illustrating the path risk assessment method based on dynamic weight correction according to task context provided in an embodiment of the present invention. The method mainly includes the following steps: S1: Obtain a set of candidate paths and divide each path into several continuous road segments; for each road segment, obtain multiple risk indicators under different risk categories, and use data quality coefficients to mark credibility.

[0015] S11. Obtain the set of candidate paths to be evaluated obtained from global path planning, denoted as: ; in, The set of candidate paths to be evaluated. The number of candidate paths, For the first 10 candidate paths.

[0016] Each candidate path is represented as an ordered sequence of several consecutive road segments, as follows: ; For path The first in Each section of road, For path The number of road segments included.

[0017] S12, for each road segment Based on different risk categories, multiple risk indicators related to path risk assessment are obtained.

[0018] Optionally, as one implementation method, this embodiment can obtain risk indicator data related to route risk assessment from three risk categories: meteorological risk, timeliness risk, and road risk.

[0019] Specifically, regarding meteorological risks, information can be obtained from external environmental sensing systems and road sections. Meteorological data corresponding to the spatial location and time of passage are used as meteorological risk indicators. Examples include road visibility, road surface adhesion coefficient, and crosswind-induced deviation factors.

[0020] Regarding timeliness risks, information about road segments can be obtained from external task planning systems. Information on time constraints related to passage serves as an indicator of timeliness risk. Examples include maneuver time window deviation, time window width, and time redundancy.

[0021] Regarding road risk, road geometry parameters for each road segment can be obtained from external vehicle-road cooperative systems as road risk indicators. These include, for example, the overall road gradient, road curvature, and minimum road width.

[0022] Preferably, this embodiment mainly uses the above-mentioned three risk types and a total of nine risk indicators as examples to describe the method of the present invention in detail. In this embodiment, the three risk types can be referred to as meteorological risk. Timeliness risk and road risks The nine indicators can be denoted as to ,in, to These represent three meteorological risk indicators: road visibility, road surface adhesion coefficient, and crosswind-induced deviation factor. to These represent three timeliness risk indicators: maneuver time window deviation, time window width, and time redundancy. to These represent three road risk indicators: comprehensive road slope, road curvature, and minimum road width.

[0023] It is understandable that, in practice, the risk indicators for the above three risk types are not limited to the nine listed. Other indicators can be reduced, added, or replaced according to specific circumstances and needs. Furthermore, other risk types and their corresponding risk indicators can be selected; this embodiment does not impose specific limitations on this.

[0024] S13, for each road segment Introducing a data quality coefficient This is used to indicate the completeness and reliability of the data for that road segment, thereby assigning a reliability label to that road segment.

[0025] Specifically, the path The first in Data quality coefficient of each road segment satisfy The smaller the value, the higher the uncertainty of the data.

[0026] S2: For each road segment, extract the original physical features corresponding to multiple risk indicators and construct a physical feature vector; at the same time, calculate the risk aggregation weight factor based on the length ratio of each road segment.

[0027] S21: For each road segment Extract the original physical features that correspond one-to-one with the nine risk indicators in S12 to form a physical feature vector, as follows:

[0028] in, For physical feature vectors, For road section In terms of indicators The original physical characteristic values ​​below.

[0029] The physical characteristics and their meanings corresponding to each indicator are as follows: Road visibility index : in, For road section Visibility at the location.

[0030] Road surface adhesion coefficient index : in, For road section The road adhesion coefficient.

[0031] Crosswind-induced deviation factor index : in, For road section crosswind intensity, For the section of road The duration of exposure.

[0032] Maneuver time window deviation index : in, For the expected destination Time, For road section The permitted time window.

[0033] Time window width index : in, The width is set for the time window.

[0034] Time redundancy index : in, For path The planned total travel time, The latest allowed completion time for the task.

[0035] Comprehensive index of road slope : in, For road section Maximum longitudinal slope, For road section The maximum lateral slope.

[0036] Road curvature index : in, For the section of road The maximum road curvature.

[0037] Minimum road width index : in, This is the minimum width of the road.

[0038] S22: Calculate the risk aggregation weight factor based on the length proportion of each road segment to aggregate road segment risks into path risks, as shown below:

[0039] In the formula, For path Total length, For road section Length, For path The number of road segments included. For road section Risk aggregation weighting factor.

[0040] S3: Construct a hierarchical analysis model for comprehensive path risk assessment based on different risk categories and risk indicators, and define the relative importance scale between elements within each level.

[0041] S31: Construct a hierarchical analysis model based on different risk categories and risk indicators, including an objective layer, a criterion layer, and an indicator layer; wherein, the objective layer is used to describe the overall objective of minimizing the comprehensive risk of the path; the criterion layer covers risk categories of different dimensions; and the indicator layer covers multiple risk indicators under different risk categories of different dimensions.

[0042] For example, using the three risk categories and nine sub-indicators mentioned above, a three-tiered hierarchical analysis model is constructed, comprising an objective layer, a criterion layer, and an indicator layer, for comprehensive path risk assessment. Please refer to [link / reference]. Figure 2 , Figure 2 This is a schematic diagram of the hierarchical analysis model for comprehensive path risk assessment provided in an embodiment of the present invention. It can be seen that the objective layer is the minimization of comprehensive risk, which describes the overall objective of minimizing comprehensive path risk. The criterion layer includes meteorological risk. Timeliness risk and road risks This layer is used to describe different risk categories, and includes nine risk indicators. to Used to describe specific quantifiable risk indicators. to Belonging to , to Belonging to , to Belonging to .

[0043] S32: Define a scale of relative importance between any two elements.

[0044] In this embodiment, the relative importance scale between any two elements is used to compare elements pairwise within the same level to construct a judgment matrix, as specifically represented below:

[0045] In the formula, For any two elements at the same level, as elements Relative to elements Importance scale as elements Relative to elements The importance scale has the following semantics: 1 indicates equal importance; 3 indicates slightly important; 5 indicates significant importance; 7 indicates strong importance; 9 indicates extreme importance; and 2 / 4 / 6 / 8 are the median values ​​of adjacent levels.

[0046] S4: For each level in the hierarchical analysis model, construct a judgment matrix using the relative importance scale, calculate the weight of each judgment matrix, and then obtain the global baseline weight matrix.

[0047] S41: For any level other than the target level, construct the corresponding judgment matrix using the relative importance scale between any two elements in the current level as elements.

[0048] Specifically, this embodiment uses Figure 2 Taking the hierarchical analysis model shown as an example, judgment matrices can be constructed for the criterion layer and the indicator layer.

[0049] First, regarding the three criterion elements in the criterion layer... A criterion-level judgment matrix is ​​constructed to reflect the relative importance of the three criterion elements, as follows:

[0050] in, The criterion layer judgment matrix, As a standard Relative to criteria The importance comparison value.

[0051] The matrix elements satisfy the following relationship:

[0052] Then, for the indicator layer, in each criterion layer The following constructs index judgment matrices to reflect the criteria layer. The relative importance of the following three indicators is determined by constructing each judgment matrix in the same way as the criterion layer, as follows:

[0053] in, This is the indicator layer judgment matrix. As an indicator relative to indicators Importance comparison value, meteorological risk indicator judgment matrix Corresponding indicators Timeliness Risk Indicator Judgment Matrix Corresponding indicators Road risk indicator judgment matrix Corresponding indicators .

[0054] It should be noted that, in order to verify the rationality of the judgment matrix, after constructing the judgment matrix, the following consistency check is performed on each judgment matrix. The check process is as follows: a) Calculate the consistency index using the following formula:

[0055] In the formula, As a consistency indicator, To determine the largest eigenvalue of a matrix, To determine the order of a matrix; b) Calculate the consistency ratio based on the consistency index. The calculation formula is as follows:

[0056] In the formula, The consistency ratio, It is a random consistency indicator; c) If judging the consistency ratio If the value is less than the preset threshold, the consistency of the judgment matrix is ​​verified; otherwise, return to S3 to redefine the relative importance scale. Specifically, return to S32 to readjust the scale until the constructed judgment matrix passes the consistency check.

[0057] S42: For each judgment matrix, the weight of each element is calculated using the geometric mean method, thereby obtaining the weight vector of the corresponding level.

[0058] First, calculate the... The row geometric mean is as follows:

[0059] in, For the first The geometric mean of the row The first in the matrix Line number The elements of the column.

[0060] Then, the geometric mean is normalized to obtain the weights, as follows:

[0061] in, For the first The weight of each element.

[0062] Therefore, the criterion layer weight vector is obtained. Indicator layer weight vector .

[0063] S44: Combine the weight vectors of all levels to obtain the global baseline weight matrix.

[0064] Specifically, by combining the global baseline weights of the nine indicators, a global baseline weight matrix is ​​obtained, as follows: =

[0065] in, This is the global baseline weight matrix. For the first The global benchmark weight of each indicator.

[0066] S5: Define positive and negative correlation risk normalization operators based on the original physical characteristics, and further construct risk mapping models for various risk indicators to transform the original physical characteristic values ​​into risk values; correct the risk values ​​according to the data quality coefficient to obtain the corrected risk values.

[0067] S51: Based on the original physical characteristics, define the positive correlation risk normalization operator and the negative correlation risk normalization operator, as follows:

[0068] In the formula, , These are the positive correlation risk normalization operator and the negative correlation risk normalization operator, respectively. These are the original physical characteristic values. These are the minimum and maximum values ​​of the feature, respectively. This is a truncation function used to restrict the result to a range. Inside; S52: For each risk indicator, construct a corresponding risk mapping model to transform the original physical characteristic values ​​into risk values.

[0069] Specifically, taking the above nine risk indicators as an example, their corresponding risk mapping models are as follows: A. The road visibility risk mapping model is as follows:

[0070] in, For path The Middle Each road section is in the indicator The risk value below, These are the minimum and maximum reference values ​​for visibility, respectively.

[0071] B. The adhesion coefficient risk mapping model is as follows:

[0072] in, For path The Middle Each road section is in the indicator The risk value below, These are the minimum and maximum reference values ​​for the adhesion coefficient, respectively.

[0073] C. The crosswind-induced skew risk mapping model is as follows:

[0074] in, For road section The combined amount of crosswind-induced deviation, Crosswind intensity, For exposure duration, These are the minimum and maximum reference values ​​for the combined crosswind-induced deflection, respectively.

[0075] D. The risk mapping model for the maneuver time window deviation is as follows:

[0076] in, For the estimated arrival time, These are the lower and upper bounds of the planned passage time window, respectively. The center moment of the time window, Due to time deviation, This represents the maximum tolerable time deviation.

[0077] E. The time window urgency risk mapping model is as follows:

[0078] in, The width of the time window, the value is equal to - , These are the minimum and maximum reference values ​​for the time window width.

[0079] F. The task completion time redundancy risk mapping model is as follows:

[0080] in, For path Time redundancy, For path The planned travel time For path The maximum allowed passage time.

[0081] G. The road slope risk mapping model is as follows:

[0082] in, This is a comprehensive measure of slope. For road section Maximum longitudinal slope, For road section Maximum lateral slope, To balance the weights, These are the minimum and maximum reference values ​​for the comprehensive slope, respectively.

[0083] H. The road curvature risk mapping model is as follows:

[0084] in, For road section The maximum curvature, These are the minimum and maximum reference values ​​for road curvature, respectively.

[0085] I. The road width risk mapping model is as follows:

[0086] in, For road section Minimum width, These are the minimum and maximum reference values ​​for road width, respectively.

[0087] S53: Based on the data quality coefficient, a conservative quality correction is made to the risk value of the current road segment, expressed by the formula:

[0088] In the formula, For path The Middle The first section of the road The uncorrected risk value of the indicator. For path The Middle The first section of the road The adjusted risk value of the indicator. path The first in Data quality coefficient for each road segment.

[0089] S6: Match the context factor vector in the pre-built context factor database according to the current task level, and calculate the weight correction vector in combination with the indicator sensitivity matrix; use the context factor vector and the weight correction vector to adjust the global baseline weight matrix to obtain the dynamic weight of each road segment under the current task.

[0090] S61: Construct a context factor database to describe the context characteristics of tasks at different task levels; and pre-configure a corresponding context factor vector for each task level in the context factor database.

[0091] In this embodiment, the task levels are divided using a finite number of discrete levels; specifically, as follows:

[0092] in, These are task levels, and different levels can be used to differentiate the overall intensity of requirements for a task in terms of timeliness, security, secrecy, or reliability.

[0093] In the context factor database, for each task level Pre-configure the corresponding context factor vectors, as follows:

[0094] in, For the task level The context vector corresponding to the time, For the first The values ​​of each context factor.

[0095] S62: Constructing the Indicator Sensitivity Matrix This is used to represent the sensitivity of each indicator under different task contexts; where, the indicator sensitivity matrix... elements in Indicates the first Individual indicators For the first Task Context Dimension The sensitivity, with a numerical range of .

[0096] S63: A correction vector is calculated using the context factor vector and the indicator sensitivity matrix. This vector represents the weight adjustment factor for each indicator in the current task context and is used to dynamically adjust the weight of each indicator. The calculation formula is as follows:

[0097] In the formula, For the correction vector, For the first One correction factor.

[0098] S64: Using the global baseline weight matrix of the correction vector, obtain the dynamic weights of each road segment under the current task, expressed by the formula:

[0099] In the formula, The weights are dynamic, and Q represents the total number of risk indicators.

[0100] S7: Based on the risk aggregation weight factor and the corrected risk value of each road segment, risk aggregation is performed, and a weighted sum is performed based on dynamic weights to obtain the comprehensive risk index of each candidate path, so as to select the optimal execution path.

[0101] S71: For paths The weighted risk value of the route under each indicator is calculated based on the risk aggregation weight factor and the corrected risk value for each road segment. The calculation formula is as follows:

[0102] In the formula, For path The The weighted risk value of each indicator. For path The number of road segments included. For path Section of the road Risk aggregation weighting factor, For path The Middle The first section of the road The adjusted risk value of the indicator; S72: Calculate the path by weighting the risk value and dynamic weights. The comprehensive risk index is expressed by the formula:

[0103] In the formula, For path The comprehensive risk index, where Q represents the total number of risk indicators. For dynamic weights.

[0104] S73: Based on the comprehensive risk index, candidate paths are ranked, and the path with the lowest risk is selected as the optimal path. The formula is as follows:

[0105] in, This is the optimal path.

[0106] Thus, multi-path risk assessment and optimal path selection have been achieved.

[0107] The path risk assessment method based on dynamic weight correction according to task context provided by this invention, on the one hand, constructs a dynamic weight correction mechanism based on task level by pre-setting a context factor database and an indicator sensitivity matrix. This allows the risk assessment criteria to be optimized and adjusted according to changes in the task's requirements for timeliness, safety, concealment, and reliability, greatly enhancing the adaptability of path planning to complex task contexts and overcoming the limitations of static weights in traditional assessment models. On the other hand, to address the uncertainty of external system perception data, a data quality coefficient is introduced and the initial risk value is conservatively corrected, effectively reducing decision-making biases caused by sensor noise or missing information, and significantly improving the credibility and robustness of the risk assessment results. Furthermore, this invention also achieves a precise characterization from local road segment characteristics to a global comprehensive path risk index by fusing multiple risk indicators of different risk categories and combining the analytic hierarchy process (AHP) with a physical feature risk mapping model, providing comprehensive quantitative support for optimal path decision-making for vehicle maneuvers under task execution environments.

[0108] Based on the same inventive concept, a second aspect of this invention also provides a path risk assessment system based on dynamic weight correction according to task context. Please see [link to relevant documentation]. Figure 3 , Figure 3 This is a structural block diagram of a path risk assessment system based on task context dynamic weight correction, provided in an embodiment of the present invention. The system includes: The data acquisition and credibility labeling module is used to acquire a set of candidate paths and divide each path into several continuous road segments; for each road segment, it acquires multiple risk indicators under different risk categories and uses data quality coefficients to label credibility. The feature extraction and road segment weight calculation module is used to extract the original physical features corresponding to multiple risk indicators for each road segment and construct a physical feature vector; at the same time, it calculates the risk aggregation weight factor based on the length ratio of each road segment. The evaluation system modeling and scaling module is used to construct a hierarchical analysis model for comprehensive path risk assessment based on different risk categories and risk indicators, and to define the relative importance scale between elements within each level; The benchmark weight calculation module constructs a judgment matrix for each level in the hierarchical analysis model using a relative importance scale, calculates the weight of each judgment matrix, and then obtains the global benchmark weight matrix. The risk mapping and correction module is used to define positive and negative correlation risk normalization operators based on the original physical characteristics, and further construct risk mapping models for various risk indicators to transform the original physical characteristic values ​​into risk values; the risk values ​​are corrected according to the data quality coefficient to obtain the corrected risk values; The dynamic weight correction module is used to match the context factor vector in the pre-built context factor database according to the current task level, and calculate the weight correction vector in combination with the indicator sensitivity matrix; the global benchmark weight matrix is ​​adjusted using the context factor vector and the weight correction vector to obtain the dynamic weight of each road segment under the current task. The risk comprehensive assessment and route decision module is used to aggregate risks based on the risk aggregation weight factors and the corrected risk values ​​of each road segment, and to perform weighted summation based on dynamic weights to obtain the comprehensive risk index of each candidate route in order to select the optimal execution route.

[0109] It should be noted that, for the system implementation, since it is basically similar to the method implementation and can achieve the same or similar beneficial effects as the method implementation, the description is relatively simple, and relevant parts can be referred to in the description of the method implementation.

[0110] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A path risk assessment method based on dynamic weight adjustment according to task context, characterized in that, include: S1: Obtain a set of candidate paths and divide each path into several consecutive road segments; For each road segment, multiple risk indicators under different risk categories are obtained, and the credibility is marked using the data quality coefficient; S2: For each road segment, extract the original physical features corresponding to the multiple risk indicators and construct a physical feature vector; at the same time, calculate the risk aggregation weight factor based on the length ratio of each road segment. S3: Construct a hierarchical analysis model for comprehensive path risk assessment based on different risk categories and risk indicators, and define the relative importance scale between elements within each level; S4: For each level in the hierarchical analysis model, construct a judgment matrix using the relative importance scale, calculate the weight of each judgment matrix, and then obtain the global benchmark weight matrix. S5: Define positive and negative correlation risk normalization operators based on the original physical characteristics, and further construct risk mapping models for various risk indicators to transform the original physical characteristic values ​​into risk values; correct the risk values ​​according to the data quality coefficient to obtain the corrected risk values; S6: Match the context factor vector in the pre-built context factor database according to the current task level, and calculate the weight correction vector in combination with the indicator sensitivity matrix; adjust the global benchmark weight matrix using the context factor vector and the weight correction vector to obtain the dynamic weight of each road segment under the current task. S7: Based on the risk aggregation weight factor of each road segment and the corrected risk value, perform risk aggregation, and perform weighted summation based on the dynamic weight to obtain the comprehensive risk index of each candidate path, so as to select the optimal execution path.

2. The path risk assessment method based on dynamic weight correction according to task context as described in claim 1, characterized in that, S1 includes: S11. Obtain the set of candidate paths to be evaluated obtained from global path planning. Each candidate path is represented as an ordered sequence of several consecutive road segments, expressed as... ;in, The set of candidate paths to be evaluated. The number of candidate paths, For the first Candidate paths, For path The first in Each section of road, For path The number of road segments included; S12, for each road segment Based on different risk categories, multiple risk indicators related to path risk assessment are obtained; S13, for each road segment Introducing a data quality coefficient This is used to indicate the completeness and reliability of the data for that road segment, thereby assigning a reliability label to that road segment.

3. The path risk assessment method based on dynamic weight correction according to task context as described in claim 1, characterized in that, In S2, the formula for calculating the risk aggregation weighting factor is as follows: In the formula, For path Total length, For road section Length, For path The number of road segments included. For road section Risk aggregation weighting factor.

4. The path risk assessment method based on dynamic weight correction according to task context as described in claim 1, characterized in that, S3 include: S31: Construct a hierarchical analysis model based on different risk categories and risk indicators, including an objective layer, a criterion layer, and an indicator layer; wherein, the objective layer is used to describe the overall objective of minimizing the comprehensive risk of the path; the criterion layer covers risk categories of different dimensions; and the indicator layer covers multiple risk indicators under different risk categories of different dimensions. S32: Define a scale of relative importance between any two elements, expressed as follows: In the formula, For any two elements at the same level, as elements Relative to elements Importance scale as elements Relative to elements The importance scale has the following semantics: 1 indicates equal importance; 3 indicates slightly important; 5 indicates significant importance; 7 indicates strong importance; 9 indicates extreme importance; and 2 / 4 / 6 / 8 are the median values ​​of adjacent levels.

5. The path risk assessment method based on dynamic weight correction according to task context as described in claim 1, characterized in that, S4 includes: S41: For any level other than the target level, construct the corresponding judgment matrix using the relative importance scale between any two elements in the current level as elements; S42: For each judgment matrix, the weight of each element is calculated using the geometric mean method, thereby obtaining the weight vector of the corresponding level; S44: Combine the weight vectors of all levels to obtain the global baseline weight matrix.

6. The path risk assessment method based on dynamic weight correction according to task context as described in claim 5, characterized in that, After S41 and before S42, the process also includes: performing a consistency check on each judgment matrix, the check process is as follows: a) Calculate the consistency index using the following formula: In the formula, As a consistency indicator, To determine the largest eigenvalue of a matrix, To determine the order of a matrix; b) Calculate the consistency ratio based on the aforementioned consistency index. The calculation formula is as follows: In the formula, The consistency ratio, It is a random consistency indicator; c) If judging the consistency ratio If the value is less than the preset threshold, the matrix consistency is deemed acceptable; otherwise, return to S3 to redefine the relative importance scale.

7. The path risk assessment method based on dynamic weight correction according to task context as described in claim 1, characterized in that, S5 include: S51: Based on the original physical characteristics, define positive correlation risk normalization operators and negative correlation risk normalization operators, as follows: In the formula, , These are the positive correlation risk normalization operator and the negative correlation risk normalization operator, respectively. These are the original physical characteristic values. These are the minimum and maximum values ​​of the feature, respectively. This is a truncation function used to restrict the result to a range. Inside; S52: For each risk indicator, construct a corresponding risk mapping model to transform the original physical characteristic values ​​into risk values; S53: Based on the aforementioned data quality coefficient, a conservative quality correction is made to the risk value of the current road segment, expressed by the following formula: In the formula, For path The Middle The first section of the road The uncorrected risk value of the indicator. For path The Middle The first section of the road The adjusted risk value of the indicator. path The first in Data quality coefficient for each road segment.

8. The path risk assessment method based on task context dynamic weight correction according to claim 1, characterized in that, S6 include: S61: Construct a context factor database to describe the context characteristics of tasks at different task levels; in the context factor database, a corresponding context factor vector is pre-configured for each task level, wherein the task levels are divided using a finite number of discrete levels; the context factor vector is represented as: In the formula, For mission level, For the task level The context vector corresponding to the time, For the first The values ​​of each context factor; S62: Constructing the Indicator Sensitivity Matrix The sensitivity matrix is ​​used to represent the sensitivity of each indicator under different task contexts; wherein, the indicator sensitivity matrix... elements in Indicates the first Individual indicators For the first Task Context Dimension The sensitivity, with a numerical range of ; S63: The correction vector is calculated using the context factor vector and the index sensitivity matrix. The calculation formula is as follows: In the formula, For the correction vector, For the first One correction factor; S64: Using the global baseline weight matrix of the correction vector, obtain the dynamic weights of each road segment under the current task, expressed by the formula: In the formula, The weights are dynamic, and Q represents the total number of risk indicators.

9. The path risk assessment method based on dynamic weight correction according to task context as described in claim 1, characterized in that, S7 includes: S71: For paths The weighted risk value of the path under each indicator is calculated based on the risk aggregation weight factor of each road segment and the corrected risk value. The calculation formula is as follows: In the formula, For path The The weighted risk value of each indicator. For path The number of road segments included. For path Section of the road Risk aggregation weighting factor, For path The Middle The first section of the road The adjusted risk value of the indicator; S72: Based on the weighted risk value and the dynamic weight, perform a weighted calculation to obtain the path. The comprehensive risk index is expressed by the formula: In the formula, For path The comprehensive risk index, where Q represents the total number of risk indicators. Dynamic weights; S73: Based on the comprehensive risk index, the candidate paths are sorted, and the path with the lowest risk is selected as the optimal path.

10. A path risk assessment system based on dynamic weight correction according to task context, used to implement the method described in any one of claims 1-9, characterized in that, The system includes: The data acquisition and credibility labeling module is used to acquire a set of candidate paths and divide each path into several continuous road segments; for each road segment, it acquires multiple risk indicators under different risk categories and uses data quality coefficients to label credibility. The feature extraction and road segment weight calculation module is used to extract the original physical features corresponding to the multiple risk indicators for each road segment and construct a physical feature vector; at the same time, it calculates the risk aggregation weight factor based on the length ratio of each road segment. The evaluation system modeling and scaling module is used to construct a hierarchical analysis model for comprehensive path risk assessment based on different risk categories and risk indicators, and to define the relative importance scale between elements within each level; The benchmark weight calculation module constructs a judgment matrix for each level in the hierarchical analysis model using the relative importance scale, calculates the weight of each judgment matrix, and then obtains the global benchmark weight matrix. The risk mapping and correction module is used to define positive and negative correlation risk normalization operators based on the original physical characteristics, and further construct risk mapping models for various risk indicators to transform the original physical characteristic values ​​into risk values; and correct the risk values ​​according to the data quality coefficient to obtain the corrected risk values. The dynamic weight correction module is used to match the context factor vector in the pre-built context factor database according to the current task level, and calculate the weight correction vector in combination with the indicator sensitivity matrix; and adjust the global benchmark weight matrix using the context factor vector and the weight correction vector to obtain the dynamic weight of each road segment under the current task. The risk comprehensive assessment and route decision module is used to aggregate risks based on the risk aggregation weight factors of each road segment and the corrected risk values, and to perform weighted summation based on the dynamic weights to obtain the comprehensive risk index of each candidate path, so as to select the optimal execution path.

Images