Route planning evaluation method adaptive to network connection unit cluster

Abstract

The embodiment of the invention provides a route planning evaluation method adaptive to a network connection unit cluster, and the method comprises the following steps: constructing a cluster collaborative digital twin environment, carrying out dynamic priority weight distribution, carrying out multi-target route planning and conflict, carrying out multi-dimensional performance evaluation, and carrying out dynamic parameter optimization. Through the evaluation method, the balance between task urgency degree guarantee and conflict resolution efficiency is realized, and meanwhile, efficient balance is realized among real-time conflict detection, dynamic weight adjustment and distributed negotiation, so that the collaborative security and task guarantee capability of a network-connected unit cluster are remarkably improved, and the network-connected unit cluster safety is improved. And the method has the characteristics of high dynamic adaptability, quantification of priority guarantee and conflict resolution efficiency through evaluation dimension multivariate, wide scene coverage and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation and unmanned systems, and particularly relates to a route planning intelligent evaluation method adapted to a networked unit cluster, and is particularly applicable to the collaborative path planning and performance evaluation of multi-agent clusters such as unmanned aerial vehicles and unmanned vehicles in complex environments. Background Art

[0002] Intelligent equipment such as unmanned aerial vehicles and unmanned vehicles can usually independently establish route planning. Literature such as CN115235491A, CN113312562A, and CN105203120A has proposed methods for evaluating the route planning of single intelligent equipment. However, there has never been a good solution to the route planning evaluation method for networked unit clusters.

[0003] That is, the path planning of conventional networked unit clusters is usually based on fixed priority rules (such as task type or submission time) or centralized control strategies, and has the following problems: insufficient dynamic adaptability, unable to dynamically adjust path priorities according to real-time task urgency or resource competition; single evaluation dimension, only evaluating basic indicators such as path length and time, lacking quantification of priority guarantee and conflict resolution efficiency; limited scenario coverage: testing depends on preset scenarios, and it is difficult to generate extreme confrontation scenarios that conform to physical laws to verify robustness.

[0004] Therefore, a new technical solution is needed to solve the above technical problems. Summary of the Invention

[0005] To this end, the present invention provides a route planning evaluation method adapted to a networked unit cluster to solve the above technical problems.

[0006] A route planning evaluation method adapted to a networked unit cluster, characterized by comprising the following steps: S100: Construct a cluster collaborative digital twin environment: load the kinematic model of the unit, task attribute tags, and a high-precision environmental map, where the task attribute tags include urgency grading; S200: Dynamic priority weight assignment: Calculate the initial priority weight through a fuzzy logic model based on the task urgency grading and the remaining endurance of the unit; Real-time detect path conflicts between units, and update the priority weight based on a distributed negotiation protocol. The priority weight update formula satisfies: the priority weight value is negatively correlated with the conflict distance overlimit amount, and the priority weight normalization ensures that the total priority weight sum of the cluster is 1; S300: Multi-objective path planning and conflict resolution: Embed the priority weight into the path cost function to generate an initial path; When an initial path crossing is detected, the low-priority aircraft adopts a heading angle adjustment strategy to avoid conflicts and ensure the minimum safe distance; S400: Multi-dimensional performance evaluation: Calculate the actual completion time deviation rate of high-emergency tasks as an index of the priority guarantee rate; Statistically calculate the proportion of the average detour distance caused by priority adjustment as an index of conflict resolution efficiency; S500: Dynamic parameter optimization: Reverse-adjust the parameters of the fuzzy logic model according to the evaluation results; If the priority guarantee rate fails to meet the standard continuously, trigger an alarm and lock the current parameter set.

[0007] Among them, in step 200, the fuzzy logic model is preset with a fuzzy rule table, where the fuzzy rule table is preset with conditional combinations and rule weights, and each conditional combination corresponds to a rule weight, and based on the task urgency classification and the remaining endurance of the aircraft, calculating the initial priority weight through the fuzzy logic model includes the following steps: S201: Based on the task urgency classification and the remaining endurance of the aircraft, match the conditional combinations of the fuzzy rule table and assign rule weights; S202: Obtain the initial priority weight according to the membership degree values of the task urgency and the remaining endurance and the rule weights. Among them, the membership degree function of the task urgency adopts a combination of triangular and trapezoidal functions to cover the domain, and the membership degree function of the remaining endurance adopts a trapezoidal function to cover the domain.

[0008] Among them, in step S200, conflict detection is performed by means of spatio-temporal overlap determination.

[0009] Among them, the update of the priority weight based on the distributed negotiation protocol is carried out according to the following steps: When it is detected that aircraft i i and j j have a path conflict, each aircraft broadcasts its initial priority weight and ; Each aircraft updates the weight according to the following formula: ; In the formula: : The updated priority weight of aircraft ; : The initial priority weight of aircraft (calculated by the fuzzy logic model); : Aircraft The initial priority weight (calculated by the fuzzy logic model); : The dynamic attenuation coefficient; Is the conflict distance overlimit; : Represents the actual spatial distance between two units at a certain moment; : With the unit The set of units in conflict; : The set of all conflicting units; And, the updated priority weight should satisfy , so that the total priority weight of the cluster is conserved.

[0010] Among them, embedding the priority weight in the path cost function in step S300 is carried out through the following steps: S301: Obtain the priority weight of the current unit ; S302: Embed the priority weight Into the multi-objective optimization framework, comprehensively calculate the time cost T, energy consumption cost E, and safety cost S, and weight them according to the dynamic weight.

[0011] Among them, after step S300, there is a step of increasing the priority weight of the avoidance unit.

[0012] Among them, after generating the initial path in step S300, there is also a step of optimizing the local path using the gradient descent method.

[0013] Among them, if there is still a path conflict in the initial path or the initial path optimized by the local path, then return to execute step S200: Real-time detection of path conflicts between units, updating the priority weight based on the distributed negotiation protocol, and the priority weight update formula satisfies: The priority weight value is negatively correlated with the conflict distance overlimit, and the normalization of the priority weight ensures that the sum of the total priority weights of the cluster is 1.

[0014] Among them, adjusting the parameters of the fuzzy logic model in step S500 includes adjusting the parameters of the membership function and the rule weights in the rule table.

[0015] Among them, the average detour distance ratio ≤ 15%, and the priority guarantee rate ≥ 90%.

[0016] Beneficial effects: The embodiment of the present invention provides a route planning evaluation method adapted to a networked unit cluster, including the following steps: constructing a cluster collaborative digital twin environment, dynamically allocating priority weights, multi-objective path planning and conflict handling, multi-dimensional performance evaluation, and dynamic parameter optimization. Through the above evaluation method, the balance between task urgency guarantee and conflict resolution efficiency is achieved. At the same time, an efficient balance is realized among real-time conflict detection, dynamic weight adjustment, and distributed negotiation, significantly improving the collaborative security and task guarantee ability of the networked unit cluster, and having characteristics such as high dynamic adaptability, multiple evaluation dimensions capable of quantifying priority guarantee and conflict resolution efficiency, and wide scenario coverage. Brief Description of the Drawings

[0017] Figure 1 is a flowchart of the route planning evaluation method of the present invention adapted to a networked unit cluster. Detailed Embodiments

[0018] Refer to the attached Figure 1 , the embodiment of the present invention provides a route planning evaluation method adapted to a networked unit cluster, including the following steps: S100: Construct a cluster collaborative digital twin environment: Load the kinematic model of the unit, task attribute tags, and a high-precision environmental map, where the task attribute tags include urgency grading.

[0019] Specifically, when the unit is a drone unit, the drone can adopt a six-degree-of-freedom model. When the unit is an unmanned vehicle unit, the unmanned vehicle can adopt a bicycle model. Preferably, the unit is a drone unit.

[0020] Among them, the task attribute tags can set urgency level tags and maximum allowable delay tags according to the task type.

[0021] Among them, the high-precision environmental map can use lidar and visual SLAM to generate a grid map with a resolution ≤ 0.1 m. At the same time, the initial distribution of dynamic obstacles is randomly generated through a Poisson process.

[0022] S200: Dynamically allocate priority weights: a. Calculate the initial priority weight through a fuzzy logic model based on the task urgency grading and the remaining endurance of the unit.

[0023] Among them, the task urgency refers to the time sensitivity level of the task. In this embodiment, the task urgency is represented by E, and the larger the value, the more urgent it is. In other embodiments, it can also be set such that the smaller the value, the more urgent it is.

[0024] In this embodiment, the value range of the task urgency E is E ∈ [0, 5], and its specific meaning represents a finite level of 5 levels. In other embodiments, the task urgency can also be other value ranges, such as E ∈ [0, 10]. It can be understood that the value range of the task urgency E should be finite and excessive dispersion should be avoided.

[0025] Among them, the remaining endurance ability refers to the percentage of the current remaining energy of the unit in the total capacity. In this embodiment, the remaining endurance ability is represented by R, and its value range is R ∈ [0%, 100%]. It can be understood that the lower the endurance, the more limited the ability to execute high-priority tasks.

[0026] The fuzzy logic model is preset with a fuzzy rule table. Among them, the fuzzy rule table is preset with conditional combinations and rule weights, and each conditional combination corresponds to a rule weight. At the same time, the conditional combination is a fuzzy set cross-combination of the task urgency E and the remaining endurance ability R.

[0027] In a specific embodiment, the fuzzy rule base is the rule shown in the following table.

[0028]

[0029] Calculating the initial priority weight through the fuzzy logic model based on the task urgency classification and the remaining endurance ability of the unit includes the following steps: S201: Based on the task urgency classification and the remaining endurance ability of the unit, match the conditional combination of the fuzzy rule table and assign the rule weight; S202: Obtain the initial priority weight according to the membership degree values of the task urgency, the remaining endurance ability, and the rule weight.

[0030] Among them, obtaining the initial priority weight according to the membership degree values of the task urgency, the remaining endurance ability, and the rule weight uses the centroid method to calculate and obtain the initial priority weight.

[0031] Specifically, it is calculated through the following formula: ; In the formula: : The initial priority weight; : The rule weight of each rule; : The activation intensity of each rule, Among them and are the membership degree values of the task urgency and the remaining endurance ability respectively.

[0032] It can be understood that the initial priority weight is the weighted average of the activation intensity of each rule and the rule weight.

[0033] Among them, the membership degree value of the task urgency belongs to the membership function. In this embodiment, the membership function of the task urgency E adopts a combination of triangular and trapezoidal functions to cover the domain: Low urgency (Low): ; Medium urgency (Medium): ; High urgency (High): ; Among them, the membership degree value of the remaining battery life belongs to the membership function. In this embodiment, the membership function of the remaining battery life R adopts a combination of trapezoidal functions to cover the domain: Low battery life (Low): ; Medium battery life (Medium) ; High battery life (High): ; It can be understood that in other embodiments, when the task urgency E is in other value ranges, such as E ∈ [0, 10], the membership function can be re-determined according to the value range of the task urgency E. However, the membership functions of each fuzzy set (low, medium, high) should cover the entire domain without gaps.

[0034] In a specific embodiment, when the task urgency of a certain UAV is E = 4 (medium-high urgency) and the remaining battery life R = 70% (medium battery life), its: Membership degree of task urgency: High = 0.5, Medium = 0.5; Membership degree of remaining battery life: Medium = 1.0; Match rules 1, 2, and 4 in the fuzzy rule table: Rule 1 (High(E≥3), Any R): Rule weight w1 = 0.9, activation intensity μ1 = min(0.5, 1) = 0.5; Rule 2 (High(E≥3), Medium (80%>R>30%)): Rule weight w2 = 0.7, activation intensity μ2 = min(0.5, 1) = 0.5; Rule 4 (Medium (1 < E < 5), Medium (80% > R > 30%)): The rule weight w3 = 0.5, and the activation intensity μ5 = min(0.5, 1) = 0.5; Then, the initial priority weight is: ; b. Real-time detection of path conflicts between units, and update the priority weight based on the distributed negotiation protocol. The weight update formula satisfies: The weight value is negatively correlated with the conflict distance excess limit, and the weight normalization ensures that the total weight of the cluster is 1.

[0035] Among them, conflict detection is carried out by means of spatio-temporal overlap determination.

[0036] Specifically, it includes: Path prediction: Each unit periodically broadcasts its predicted path for the next T seconds.

[0037] The above-mentioned periodicity can be per second, and the format of the predicted path can be time-position sequence: ; In the formula: The predicted path of unit ; t: Timestamp; : The horizontal coordinate of unit at time ; : The vertical coordinate of unit at time ;

[0038] Conflict determination: Preset a safety distance threshold , and when the paths of two units and enter the spatial overlap area within the same time period, and the distance between them is less than or equal to , there is a conflict.

[0039] Specifically, it can be determined by the following formula: ; In the formula: : Safety distance threshold; : Time t is within the interval ;

[0040] Among them, It can be set through experiments or industry standards or adjusted based on real-time status. When the is preset as a fixed value, is 8 - 12m. When is configured to be adjusted based on real-time status, The value of can be dynamically calculated according to the current speed of the unit, the environmental risk level, etc. Specifically, it can be dynamically adjusted by the following formula: ; In the formula: The influence weight of speed on the safety distance; : The influence weight of acceleration on the safety distance; : The influence weight of the environmental risk factor on the safety distance; ; : The instantaneous acceleration of the unit at time t; : The environmental risk factor, such as increasing the margin in low visibility; : The dynamically adjusted safety distance (value at time t); In the formula, , , : Usually, it can be calibrated through historical data.

[0041] The update of the priority weight based on the distributed negotiation protocol is carried out according to the following steps: S210: When it is detected that there is a path conflict between the unit and , each unit broadcasts its initial fuzzy logic weight and .

[0042] S211: Each unit updates the weight according to the following formula: ; In the formula: : The updated priority weight of unit ; : The initial priority weight of unit (calculated by the fuzzy logic model); : The initial priority weight of unit (calculated by the fuzzy logic model); is the dynamic attenuation coefficient: ; λ is the conflict sensitivity coefficient, with a value range of [0.05, 0.2]. The smaller the value, the more sensitive it is to the conflict distance; is the conflict distance overlimit: ; : represents the actual spatial distance between two units at a certain moment; : the set of units in conflict with unit ; : the set of all conflict units.

[0043] It can be understood that the weight value is negatively correlated with the conflict distance overlimit, that is, the overlimit is larger, is smaller, and the greater the reduction in weight. That is, is closer to 1, then the optimization time and energy consumption are prioritized (high-priority units pursue efficiency), while is closer to 0, then risk avoidance is prioritized (low-priority units actively avoid).

[0044] At the same time, the updated weight should satisfy , ensuring the conservation of the total cluster weight.

[0045] In a specific embodiment: Initial state: Unit A (medical task): = 0.6; Unit B (logistics task): = 0.3; Unit C (inspection task): = 0.1; Safety distance = 10 m → γ = 0.1.

[0046] Conflict event: 1. A conflicts with B: The predicted distance between A and B: d AB = 8 m → Δd AB = 2 m; 2. B conflicts with C: The predicted distance between B and C: d BC = 9 m → Δd BC = 1 m.

[0047] After weight update = 0.4912 / 0.8039 ≈ 0.611; = 0.2222 / 0.8039 ≈ 0.276; = 0.0905 / 0.8039 ≈ 0.113; ; That is: The weight of the high - priority unit A increases (0.6 → 0.611) because its conflict is less serious (Δd = 2); the weight of unit B decreases (0.3 → 0.276) and it undertakes more avoidance responsibilities; the total weight remains 1, meeting the normalization constraint.

[0048] Meanwhile, through the above - mentioned step - by - step mechanism, the system achieves an efficient balance among real - time conflict detection, dynamic weight adjustment, and distributed negotiation, significantly improving the collaborative security and task guarantee capabilities of the connected unit cluster.

[0049] S300: Multi - objective path planning and conflict resolution.

[0050] a. Embed the priority weight into the path cost function to generate the initial path.

[0051] The path cost function is a heuristic function based on the A* algorithm: ; where is the actual path cost from the starting point to node n; is the heuristic estimated cost from node n to the end point.

[0052] The initial path is the path of the node with the minimum

[0053] Among them, the embedding of the priority weight is carried out in the following way: S301: Obtain the priority weight of the current unit ; S302: Embed the priority weight into the multi - objective optimization framework, comprehensively calculate the time cost T, energy consumption cost E, and safety cost S, and weight them according to the dynamic weight: ; In the formula: : The actual path cost from the starting point to node n; : Time cost = path length / maximum speed of the unit; : Energy consumption cost = integral of the square of acceleration (reflecting control energy consumption); : Safety risk cost, ; : Weight coefficient of time cost; : Weight coefficient of energy consumption cost; : Weight coefficient of safety cost; Each weight coefficient should also satisfy the normalization constraint, that is: .

[0054] Furthermore, the gradient descent method is used to optimize the local path.

[0055] Among them, the optimized variable is the sequence of path point heading angles { , ,..., }; The objective function is: ; In the formula: ; ; : Conflict penalty coefficient, usually = 10; : Represents the actual spatial distance between two units at a certain moment; : After determining the initial path , that is: ; ϵ: Extremely small positive number; Iteratively update the objective function to: : ; In the formula: : Learning rate, usually, = 0.01.

[0056] : .

[0057] Furthermore, it also includes the step of verifying the feasibility of the initial path.

[0058] Specifically, it includes: Conduct kinematic constraint check: Among them, the maximum steering angle limit: ; : Maximum allowable steering angle; It can be understood that It can be set according to the kinematic model of the unit. Preferably, the angle of the drone is 30° and the angle of the unmanned vehicle is 25°.

[0059] Acceleration limit: ; : The maximum allowable acceleration.

[0060] Furthermore, the initial path or the path optimized through local path still < , then return to step 200b to perform dynamic weight update.

[0061] It can be understood that through the multi-objective path planning embedded with priority weights and the gradient descent local optimization, the efficient cooperation and autonomous conflict resolution of the connected unit cluster are realized.

[0062] b. When a path intersection is detected, the low-priority unit adopts a heading angle adjustment strategy to avoid conflicts and ensure the minimum safety distance.

[0063] Furthermore, before executing this step, it includes a priority comparison step: That is, compare the priority weights of the two units and . The unit with the lower weight is marked as the avoidance responsible party. At the same time, if the weights are the same, the unit with the smaller ID is the avoidance responsible party.

[0064] The described heading angle adjustment strategy includes the following steps: S310: Obtain the relative position and velocity vector of the two units at the conflict point: ; In the formula: : The position vector difference of the two units at the predicted conflict time ; : The predicted conflict time; : The unit at time position coordinate; : The unit at time position coordinate; : The unit instantaneous velocity vector, the direction is consistent with the tangent direction of the motion trajectory; : The unit instantaneous velocity vector, the direction is consistent with the tangent direction of the motion trajectory.

[0065] Obtain the minimum safety distance constraint: ; In the formula: : The conflict prediction time window, which is the time window from the current moment to the predicted conflict time and is used to calculate the offset of the future position.

[0066] Obtain the heading angle adjustment direction: Calculate the relative heading angle between the two aircraft: ; If the current heading of the aircraft is , preferentially adjust in the direction perpendicular to the relative heading (turn left or right).

[0067] Among them, , ; : The relative heading angle, which represents the azimuth angle from aircraft to aircraft .

[0068] Obtain the heading angle increment based on geometric avoidance: ; Among them, the constraint condition: , is the maximum allowable steering angle.

[0069] S311: Execute the heading angle adjustment.

[0070] The heading angle adjustment can be executed by a single turn or a progressive adjustment.

[0071] Among them, when using a single turn, directly apply to execute the heading angle adjustment.

[0072] When executing the heading angle adjustment in a progressive adjustment manner, increase by each step, and a total of steps are used for adjustment to avoid sharp turns.

[0073] It can be understood that it is also necessary to verify whether the acceleration of the adjusted path meets , if not, reduce the speed or adjust in segments.

[0074] In addition, generate the adjusted path according to the new heading angle and re-broadcast it to the cluster.

[0075] In addition, after this step, it is also necessary to verify the avoidance effect: re-predict the distance between the two aircraft at time tc , if ≥ , then the avoidance is successful, if, < , then step 300b is executed again.

[0076] Furthermore, due to the additional energy consumption or time loss caused by the avoidance aircraft's detour, its weight can be temporarily increased , such as , that is, increased by 1.1 times, to avoid continuous resource preemption.

[0077] Through the above method, the minimum adjustment amount can be quickly generated, the detour distance can be reduced, the temporary weight increase mechanism can balance the avoidance cost, and the computational complexity is low to ensure real-time performance. That is, through the priority-driven heading angle adjustment strategy, low-priority aircraft can avoid conflicts at the minimum path cost on the premise of ensuring a safe distance, and ensure the efficient execution of high-priority tasks.

[0078] S400: Multi-dimensional performance evaluation.

[0079] a. Calculate the actual completion time deviation rate of high-urgency tasks as the priority guarantee rate index.

[0080] The high-urgency tasks mentioned above are tasks defined as High in the task urgency classification, or tasks with E≥4 in the 5-level urgency classification, or tasks defined as high-urgency tasks in the urgency level identification.

[0081] The time deviation rate is obtained through the following formula: ; In the formula: ; ; , for task , its theoretically optimal time, ; where, : The length of the shortest conflict-free path; : The maximum speed defined in the aircraft kinematic model.

[0082] It can be understood that When it is a negative deviation, it means early completion, then =0%, When it is a positive deviation, it means delayed completion.

[0083] Priority guarantee rate statistics: ; In the formula: ; ; N: The total number of high-urgency tasks; It is understandable that if the task is not completed (such as cancelled or failed), then = 100%.

[0084] Furthermore, the priority guarantee rate can be calculated with weights, such as assigning weights according to the task urgency , then the priority guarantee rate is calculated according to the following formula: ; In the formula: : weight; : .

[0085] Furthermore, if the task is not completed due to system failure, it is marked as = 100%, and is included in the guarantee rate calculation for statistics.

[0086] Furthermore, is positively correlated with the task urgency E. Specifically, when E = 5, = 5, when E = 4 = 3.

[0087] It is understandable that by calculating the time deviation rate of high-urgency tasks and statistically analyzing the priority guarantee rate, the system can accurately evaluate the effectiveness of dynamic path planning and conflict resolution strategies, providing core data support for parameter optimization and operation and maintenance decision-making.

[0088] b. Statistically analyze the proportion of the average detour distance caused by priority adjustment as an index of conflict resolution efficiency.

[0089] Among them, the proportion of the detour distance is obtained according to the following formula: ; In the formula: : The theoretically shortest path of each unit without conflict can be obtained by calculating through the path planning algorithm in step S300a.

[0090] : The actual travel path length of the unit due to priority adjustment is measured in real time through a trajectory tracking algorithm (such as SLAM).

[0091] It is understandable that the proportion of the detour distance needs to be calculated for each conflict event ; Among them, : represents the mth conflict event, ; : The total number of effective conflict events that occurred between units during the evaluation period; = 1: No detour; < 1: Path increase; > 1: Path optimization.

[0092] Global average efficiency ≥ 85%, that is, the proportion of the average detour distance ≤ 15%. Among them, the global average efficiency is obtained through the following formula: ; S500: Dynamic parameter optimization: a. Reverse-adjust the parameters of the fuzzy logic model according to the evaluation results.

[0093] The adjusted parameters of the fuzzy logic model can be the parameters of the membership function, and the rule weight w in the rule table k .

[0094] Among them, the parameters of the membership function can be the vertex position of the triangular function of the task urgency E, the turning points of the trapezoidal function of the remaining endurance R, etc.

[0095] The reverse adjustment of the parameters of the fuzzy logic model according to the evaluation results also includes the priority guarantee rate and the conflict resolution efficiency Merge the evaluation results (priority guarantee rate and conflict resolution efficiency) into a single optimization goal: ; In the formula: : The weight coefficient of the conflict resolution efficiency. In this embodiment = 0.5, and in other embodiments, it can be adjusted according to the scenario; J: Comprehensive optimization goal value (to be maximized); It can be understood that should be maximized to improve both the priority guarantee and the conflict resolution efficiency simultaneously.

[0096] b. If the priority guarantee rate fails to meet the standard continuously, trigger an alarm and lock the current parameter set.

[0097] The continuous failure of the priority guarantee rate to meet the standard means that it has not met the standard in the most recent K consecutive periods. The so-called meeting the standard means that the priority guarantee rate ≥ 90%. The so-called K is usually 3 - 5, and in this embodiment, K = 3.

[0098] It can be understood that in this embodiment, by inversely adjusting the parameters of the fuzzy logic model, the system can autonomously optimize the dynamic priority allocation strategy, significantly improve the collaborative performance of the networked unit cluster in complex scenarios, and at the same time ensure the continuous improvement of the evaluation indicators. Meanwhile, through continuous monitoring, intelligent warning and parameter locking mechanisms, the system can quickly respond when the priority guarantee rate is continuously insufficient, ensure the reliability of key task execution, and provide stable operation guarantee for the networked unit cluster in complex dynamic environments.

[0099] The above are only the embodiments of the present invention, and thus do not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A route planning and evaluation method adapted for a networked cluster, characterized in that: The following steps are involved: S100: Build a cluster collaborative digital twin environment: load the unit kinematic model, task attribute labels and high-precision environment map, where the task attribute labels include urgency classification; S200: Dynamic priority weight allocation: Based on the mission urgency classification and the crew's remaining endurance, the initial priority weight is calculated through a fuzzy logic model; Real-time detection of path conflicts between units, update priority weights based on distributed negotiation protocol, and the priority weight update formula satisfies: the priority weight value is negatively correlated with the conflict distance limit, and the priority weight is normalized to ensure that the total priority weight of the cluster is 1; S300: Multi-objective path planning and conflict resolution: Embed priority weights in the path cost function to generate the initial path; When the initial path intersection is detected, the low-priority crew adopts a heading angle adjustment strategy to avoid conflict and ensure the minimum safety distance; S400: Multi-dimensional performance evaluation: Calculate the actual completion time deviation rate of high-urgent tasks as a priority guarantee rate indicator; The average detour distance caused by priority adjustment is calculated as the conflict resolution efficiency indicator; S500: Dynamic parameter optimization: Reversely adjust the fuzzy logic model parameters according to the evaluation results; If the priority guarantee rate fails to meet the standard continuously, an alarm is triggered and the current parameter set is locked.

2. The evaluation method according to claim 1, characterized in that: In step 200, the fuzzy logic model is preset with a fuzzy rule table, wherein the fuzzy rule table is preset with condition combinations and rule weights, and each condition combination corresponds to a rule weight, and based on the mission urgency classification and the remaining endurance of the crew, calculating the initial priority weight through the fuzzy logic model includes the following steps: S201: Based on the mission urgency classification and the crew's remaining endurance, match the condition combination of the fuzzy rule table and assign rule weights; S202: Obtaining an initial priority weight according to the task urgency, the membership value of the remaining endurance and the rule weight, wherein the membership function of the task urgency adopts a combination of triangular and trapezoidal functions to cover the domain, and the membership function of the remaining endurance adopts a combination of trapezoidal functions to cover the domain.

3. The evaluation method according to claim 2, characterized in that: In step S200, conflict detection is performed by means of spatiotemporal overlap determination.

4. The evaluation method according to claim 2, characterized in that: The updating of priority weights based on the distributed negotiation protocol is performed in the following steps: When the unit is detected i With j j When there is a conflict in the paths, each unit broadcasts its initial priority weight and ; The weight of each unit is updated as follows: ； Where: :unit Updated priority weights; :unit The initial priority weights of (calculated by the fuzzy logic model); :unit The initial priority weights of (calculated by the fuzzy logic model); : dynamic attenuation coefficient; The conflict distance exceeds the limit; : Indicates the actual spatial distance between the two units at a certain moment; :With the crew There are conflicting crew sets; : The collection of all conflicting crews; And, the updated priority weight should satisfy , so that the total priority weight of the cluster is conserved.

5. The evaluation method according to claim 4, characterized in that: In step S300, the priority weight is embedded in the path cost function by the following steps: S301: Get the priority weight of the current unit ; S302: Priority weight Embedded in a multi-objective optimization framework, the time cost T, energy cost E, and safety cost S are comprehensively calculated and weighted according to dynamic weights.

6. The evaluation method according to claim 5, characterized in that: After step S300, the step of increasing the priority weight of the avoiding aircraft group is included.

7. The evaluation method according to claim 5, characterized in that: After the initial path is generated in step S300, a step of optimizing the local path by using a gradient descent method is also included.

8. The evaluation method according to claim 7, characterized in that: If the initial path or the initial path after local path optimization still has a path conflict, then return to execute step S200: real-time detection of path conflicts between units, update priority weights based on a distributed negotiation protocol, and the priority weight update formula satisfies: the priority weight value is negatively correlated with the conflict distance excess, and the priority weight is normalized to ensure that the total priority weight of the cluster is 1.

9. The evaluation method according to claim 5, characterized in that: In step S500, the fuzzy logic model parameters are adjusted, including adjusting the membership function parameters and the rule weights in the rule table.

10. The evaluation method according to claim 1, characterized in that: The average detour distance ratio is ≤ 15%, and the priority guarantee rate is ≥ 90%.

Images