Vehicle cluster data fusion and resource allocation method based on adaptive game algorithm

Abstract

The invention relates to a vehicle cluster data fusion and resource allocation method based on a self-adaptive game algorithm, belongs to the technical field of vehicle resources, and solves the problem that an existing vehicle cluster cannot achieve accurate task and resource allocation when facing a dynamic and changeable operation environment and vehicle performance fluctuation in operation. Comprising the following steps: acquiring multi-dimensional task data of a plurality of task periods generated in a task execution process of each vehicle in a vehicle cluster, and obtaining a task performance data set of each vehicle in the vehicle cluster; performing data fusion on the task performance data set of each vehicle in the current task period to obtain a task performance score of each vehicle; based on the task performance score of each vehicle, the task execution capability of each vehicle in the current task period is evaluated, and tasks and resources of each vehicle are dynamically allocated; and taking a preset task cycle number as an observation window, and gradually adjusting the resources allocated to each vehicle based on the fluctuation value of the task performance score of each vehicle in each task cycle in the observation window period.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle resources, and in particular to a vehicle cluster data fusion and resource allocation method based on an adaptive game algorithm. Background Art

[0002] The collaborative operation of vehicles in a cluster plays an important role in existing intelligent production and transportation. Through collaborative operation, vehicles can divide the work and cooperate with each other, each taking on the task they are best at, thereby improving overall operational efficiency. In addition, through collaborative operation between vehicles, resources can be allocated more flexibly and effectively to ensure that resources are used where they are most needed. When faced with complex and changing environments, cluster vehicles can quickly adjust strategies and actions to adapt to different task requirements and environmental changes.

[0003] The collaborative operation and resource allocation of vehicles in the existing cluster mainly rely on a centralized command system for unified scheduling and management. During the operation process, preset static indicators are usually used for task allocation and resource scheduling. However, with the dynamic changes in the operating environment and the complexity of the operating tasks, it is impossible to accurately allocate tasks to the vehicles in the cluster, which can easily lead to insufficient resources for efficient vehicles and waste of resources for inefficient vehicles, thereby reducing the operating efficiency of the entire vehicle cluster. At the same time, when facing environmental changes, the task performance of each vehicle in the cluster often fluctuates significantly, and the traditional centralized system is difficult to respond to these changes in a timely and effective manner, resulting in uneven resource allocation and delayed decision-making. Summary of the invention

[0004] In view of the above analysis, an embodiment of the present invention aims to provide a vehicle cluster data fusion and resource allocation method based on an adaptive game algorithm, so as to solve the problem that existing vehicle clusters are often unable to achieve accurate task and resource allocation when faced with a dynamic and changeable operating environment and vehicle performance fluctuations during operation.

[0005] The purpose of the present invention is mainly achieved through the following technical solutions:

[0006] The present invention provides a vehicle cluster data fusion and resource allocation method based on an adaptive game algorithm, comprising the following steps:

[0007] Acquire and pre-process the multi-dimensional task data of several task cycles generated by each vehicle in the vehicle cluster during the task execution process to obtain a task performance data set of each vehicle in the vehicle cluster in each task cycle; wherein the task performance data set includes: task completion time, task success rate, historical task performance data and vehicle status data;

[0008] In a task cycle:

[0009] Performing data fusion on the task performance data set of each vehicle in the current task cycle to obtain a task performance score of each vehicle;

[0010] Based on the task performance scores of each vehicle, the task execution capability of each vehicle in the current task cycle is evaluated, and tasks and resources of each vehicle are dynamically allocated;

[0011] The preset number of task cycles is used as an observation window, and the resources allocated to each vehicle are progressively adjusted based on the fluctuation value of the task performance score of each vehicle during each task cycle within the observation window.

[0012] Furthermore, the task performance data set of each vehicle in the current task cycle is fused to obtain a task performance score of each vehicle, including:

[0013] Standardizing each data in the task performance data set of the current task cycle to obtain a standardized value of each task performance index;

[0014] The standardized task performance index values ​​are weighted and summed to obtain the task performance score of the vehicle in the current task cycle; wherein the weight value of each task performance is calculated based on the following formula:

[0015]

[0016] X={T,S,H,R}

[0017] in, represents the weight coefficient of the i-th vehicle on the task performance index X; T represents the task completion time index; S represents the task success rate index; H represents the historical performance data index; R represents the vehicle status data index; α X represents the adaptive adjustment parameter of the task performance index X; X represents the smoothing coefficient of the task performance indicator X; X i represents the value of the task performance index X of the i-th vehicle; J represents the total number of vehicles in the vehicle cluster; X j represents the value of the task performance index X of the jth vehicle; Represents the mean of the task performance index X of all vehicles in the vehicle cluster.

[0018] Furthermore, the data obtained by each vehicle during the execution of the task also includes real-time task performance data;

[0019] Based on the task performance score of each vehicle, the task execution capability of each vehicle in the current task cycle is evaluated, and tasks of each vehicle are dynamically assigned, including:

[0020] Constructing a task performance score piecewise function based on the task performance score of each vehicle, a preset minimum threshold and a medium threshold of the task performance score;

[0021] Based on the task performance score piecewise function and the task performance score of each vehicle, calculating an initial task allocation amount for each vehicle;

[0022] Based on the real-time task performance data of each vehicle during the task execution process, feedback adjustment is performed on the initial task allocation amount of each vehicle, and the adjusted task allocation amount of each vehicle is calculated.

[0023] Furthermore, the task performance scoring piecewise function is expressed as:

[0024]

[0025] Among them, P i represents the task performance score of the i-th vehicle; P min represents the minimum threshold of task performance score; P mid Indicates the medium threshold for task performance rating; A k represents the initial coefficient of segment k; B k represents the adjustment coefficient of the sensitivity of segment k to the fluctuation of vehicle task performance during the task allocation process; C k Represents the smoothing coefficient that affects the segment turning point within segment k.

[0026] Furthermore, the initial task allocation of each vehicle is calculated using the following formula:

[0027]

[0028] Among them, T assign (i) represents the initial task allocation of the i-th vehicle; ω k represents the weight coefficient of the kth segment function; β k represents the adjustment coefficient of the k-th segment function; γ k Represents the smoothing factor of the k-th segment function.

[0029] Furthermore, the real-time task performance data obtained by each vehicle during the task execution process is standardized to obtain a standardized real-time task performance index value;

[0030] Based on the standardized real-time task performance index value of each vehicle, the adjusted task allocation of each vehicle is calculated using the following formula:

[0031]

[0032] Where M represents the number of data types in the real-time task performance data; δ mrepresents the real-time feedback weight coefficient of the mth type of data in the real-time task performance data; ξ m represents the sensitivity coefficient of the mth type of data in the real-time task performance data; κ m represents the smoothing coefficient of the mth data in the real-time task performance data; V i ′ represents the real-time task performance index value of the i-th vehicle in the vehicle cluster; V avg Represents the mean value of the real-time task performance index of all vehicles in the vehicle cluster.

[0033] Furthermore, based on the task performance score of each vehicle, resources of each vehicle are dynamically allocated within the current task cycle, including:

[0034] Based on the task performance scores of each vehicle, the task performance differences of each vehicle are evaluated to obtain the initial resources allocated to each vehicle;

[0035] Based on the initial resources of each vehicle, the task performance score of each vehicle and the highest and lowest task performance scores among the cluster vehicles, the initial resource allocation amount of each vehicle is dynamically adjusted to calculate the adjusted resource allocation amount of each vehicle.

[0036] Furthermore, the following formula is used to obtain the initial resources allocated to each vehicle:

[0037]

[0038] Among them, R i represents the initial resources allocated to the i-th vehicle; J represents the total number of vehicles in the vehicle cluster; ε j represents the fluctuation impact value of controlling the task performance score of the jth vehicle; P i and P j denote the task performance scores of the i-th vehicle and the j-th vehicle respectively; σ j represents the influence coefficient of smoothing the resource allocation of the jth vehicle; represents the impact of adjusting the task score on the resource allocation of the i-th vehicle; P avg Represents the mean score of task performance in the vehicle cluster.

[0039] Furthermore, the following formula is used to dynamically adjust the initial resource allocation of each vehicle, and the adjusted resource allocation of each vehicle is calculated:

[0040]

[0041] Among them, R′ i ′ represents the resources allocated to the i-th vehicle; τ represents the resource allocation adjustment coefficient; P i represents the task performance score of the i-th vehicle; P avg represents the mean score of task performance in the vehicle cluster; Pmax represents the highest task performance score in the vehicle cluster; P min represents the lowest task performance score in the vehicle cluster; P thresh represents the preset task performance score threshold; η represents the adjustment range of resource allocation.

[0042] Furthermore, the progressive adjustment of resources allocated to each vehicle includes:

[0043] Based on the task performance scores of each vehicle in the current task cycle and the previous task cycle, the task performance fluctuation amplitude value of each vehicle in the current task cycle is calculated;

[0044] When the task performance fluctuation amplitude value of each vehicle in the observation window period is continuously negative, based on the task performance fluctuation amplitude mean, the penalty value of each vehicle in the current task cycle is calculated using the following formula to reduce the resource allocation value of each vehicle in the next observation window period:

[0045]

[0046] Among them, △Q i represents the mean fluctuation of the task performance of the i-th vehicle during the observation window; μ represents the penalty coefficient.

[0047] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0048] 1. The present invention uses a multi-dimensional task performance data set to comprehensively consider task completion time, task success rate, historical performance data and real-time change indicators to conduct a more comprehensive and dynamic evaluation of the real-time task performance of each vehicle in the cluster, greatly improving the accuracy of task allocation and ensuring more efficient and reasonable resource allocation in complex environments.

[0049] 2. The present invention introduces an adaptive game algorithm to drive the resource allocation of vehicles in the cluster. In the resource allocation process, a dynamic game algorithm is used to achieve fair resource competition and cooperation among vehicles, solving the problem of uneven resource allocation in traditional centralized command systems. The system can not only allocate more resources to vehicles with outstanding performance, but also dynamically adjust resource allocation according to the real-time performance of the vehicle, ensuring maximum utilization of vehicle resources.

[0050] 3. The present invention uses a progressive penalty mechanism to flexibly handle resource allocation when task performance fluctuates, and sets an observation window to avoid excessive resource reduction due to short-term performance fluctuations. Compared with the existing methods that often take immediate penalties or resource reduction measures for vehicles with poor performance, which easily lead to resource allocation imbalance or task failure, the present invention effectively improves the robustness and flexibility of the system.

[0051] In the present invention, the above-mentioned technical solutions can also be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent description, and some advantages can become obvious from the description, or can be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The drawings are only for the purpose of illustrating particular embodiments and are not to be considered limiting of the present invention. Like reference symbols denote like components throughout the drawings.

[0053] Figure 1 The figure is a flow chart of a method for vehicle cluster data fusion and resource allocation based on an adaptive game algorithm in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0055] A specific embodiment of the present invention discloses a vehicle cluster data fusion and resource allocation method based on an adaptive game algorithm, such as Figure 1 As shown, the following steps are included:

[0056] Step S1, obtaining multi-dimensional task data of several task cycles generated by each vehicle in the vehicle cluster during the task execution process and preprocessing, to obtain a task performance data set of each vehicle in the vehicle cluster in each task cycle; wherein the task performance data set includes: task completion time, task success rate, historical task performance data and vehicle status data.

[0057] Specifically, the multi-dimensional mission data represents a comprehensive record and analysis of the vehicle's performance during the mission;

[0058] The task completion time is obtained by calculating the difference between the start time and the end time of the task recorded by the vehicle's task management system or time recording module. The vehicle's task management system can automatically record the start and end events of each task, thereby accurately measuring the duration of the task.

[0059] The mission success rate is obtained by counting the number of successes and failures of multiple missions. The system can automatically determine the success or failure of each mission based on the vehicle's mission goal achievement, and summarize the mission success rate over a period of time to reflect the reliability of the vehicle during mission execution.

[0060] The historical mission performance data is extracted from the vehicle's mission database and covers the vehicle's performance records in similar missions in the past, including mission completion time, success rate, and real-time performance, etc. This type of data can be obtained through the mission record module stored in the vehicle or the back-end mission database query, providing reference data for the comprehensive evaluation of the vehicle.

[0061] The vehicle status data is obtained through the vehicle's sensor system and reflects the vehicle's real-time status during the execution of the task, such as changes in vehicle speed, fuel consumption, temperature, etc. This type of data is collected in real time through on-board sensors and data acquisition modules, and integrated and transmitted through the vehicle's central control system.

[0062] The multi-dimensional task data obtained above is preprocessed to remove abnormal data and noise information, retain valid data related to vehicle task execution, and obtain a task performance data set of each vehicle in the vehicle cluster in each task cycle; exemplarily, the preprocessing method can adopt methods such as sliding average or median filtering, which are not specified in this embodiment.

[0063] Step S2: In one task cycle, the following steps S21 to S22 are executed:

[0064] Step S21: performing data fusion on the task performance data set of each vehicle in the current task cycle to obtain a task performance score of each vehicle.

[0065] Specifically, the task cycle is the time for a vehicle to perform a task within a time period. A weighted fusion method is used to obtain the task performance score of each vehicle for evaluating the actual performance of each vehicle in a task cycle.

[0066] Furthermore, each data in the task performance data set of the current task cycle is standardized to obtain each standardized task performance index value.

[0067] Specifically, by standardizing the data, the data of different dimensions and ranges are converted to the same standard. The standardization method, for example, may use a standard deviation standardization method or a Min-Max standardization method, which is not specified in this embodiment.

[0068] Furthermore, the standardized task performance index values ​​are weighted and summed using the following formula to obtain the task performance score of the vehicle in the current task cycle.

[0069]

[0070] Where Pi represents the task performance score of the i-th vehicle; T irepresents the task completion time index value of the i-th vehicle; represents the task completion time indicator weight of the i-th vehicle; S i represents the mission success rate indicator of the i-th vehicle; represents the mission success rate indicator weight of the i-th vehicle; H i represents the historical performance data indicator of the i-th vehicle; represents the weight of the historical performance data indicator of the i-th vehicle; R i Represents the vehicle status data indicator of the i-th vehicle; Represents the vehicle status data indicator weight of the i-th vehicle.

[0071] Furthermore, the weight value of each task performance is calculated based on the following formula:

[0072]

[0073] X={T,S,H,R}

[0074] in, represents the weight coefficient of the i-th vehicle on the task performance index X; T represents the task completion time index; S represents the task success rate index; H represents the historical performance data index; R represents the vehicle status data index; α X represents the adaptive adjustment parameter of the task performance index X; X represents the smoothing coefficient of the task performance indicator X; X i represents the value of the task performance index X of the i-th vehicle; J represents the total number of vehicles in the vehicle cluster; X j represents the value of the task performance index X of the jth vehicle; Represents the mean of the task performance index X of all vehicles in the vehicle cluster.

[0075] Specifically, the weight value of each task performance is adjusted according to the deviation of each indicator from the mean value, and normalized to ensure that the sum of the weights is 1.

[0076] It should be noted that the adaptive adjustment parameter α of the task performance index X is X , whose value is between 0 and 1. A larger value will make the weight more inclined to the current task performance data. For example, it can be 0.8-1.0; a smaller value will make the weight more reflective of historical data. For example, it can be 0.2-0.5.

[0077] The smoothing coefficient λ of the task performance indicator X X, its value range is between 0.1 and 10. Smaller values ​​are suitable for data with smaller fluctuations, so that the weight changes more slowly. For example, 0.1-1.0 can be used. Larger values ​​are suitable for situations where data fluctuates greatly, which can speed up the adjustment of weights and respond more quickly. For example, 5-10 can be used.

[0078] Step S22: Based on the task performance score of each vehicle, the task execution capability of each vehicle in the current task cycle is evaluated, and tasks and resources of each vehicle are dynamically allocated.

[0079] Specifically, the tasks of each vehicle refer to the specific work that each vehicle undertakes during the operation according to its functions, capabilities and operation requirements. The resources of each vehicle refer to the various supports and supplies provided to ensure the smooth execution of the assigned tasks. In the collaborative operation of vehicle clusters, the reasonable allocation of tasks and resources is the key to improving overall efficiency and effectiveness.

[0080] Dynamic task and resource allocation based on each vehicle's task performance score can ensure that resources are reasonably allocated according to actual performance and capabilities, improve operational efficiency and adaptability, incentivize vehicles to improve performance, reduce risks, optimize decision-making processes, and adapt to changing operational needs, thereby maximizing the effectiveness and competitiveness of the entire vehicle cluster.

[0081] Furthermore, based on the task performance score of each vehicle, the task execution capability of each vehicle in the current task cycle is evaluated, and the tasks of each vehicle are dynamically allocated, including steps S2201 to S2203:

[0082] Specifically, this embodiment uses piecewise functions and a dynamic feedback mechanism to evaluate the task execution capability of armored vehicles in real time, and dynamically adjusts task allocation according to task performance scores.

[0083] Step S2201: construct a task performance score piecewise function based on the task performance score of each vehicle, the preset minimum threshold and the medium threshold of the task performance score, and the formula is:

[0084]

[0085] Among them, P i represents the task performance score of the i-th vehicle; P min represents the minimum threshold of task performance score; P mid Indicates the medium threshold for task performance rating; A k represents the initial coefficient of segment k; B k represents the adjustment coefficient of the sensitivity of segment k to the fluctuation of vehicle task performance during the task allocation process; C kRepresents the smoothing coefficient that affects the segment turning point within segment k.

[0086] Specifically, a multi-dimensional piecewise function is constructed based on the task performance score of each vehicle, and the task execution capability is divided into multiple stages, each stage corresponding to a different task allocation strategy. By dividing the task execution capability into different stages, the piecewise function allows differentiated task allocation strategies to be adopted for vehicles with different performance levels. By allocating more tasks to vehicles with better performance, the vehicles are encouraged to improve their task performance. For vehicles with poor performance, a conservative task allocation strategy helps to reduce the overall risk.

[0087] In the first section of the piecewise function, the task performance score is relatively low, so the task allocation strategy is more conservative; in the second section of the piecewise function, the task performance score is at a medium level, and the task allocation strategy is more balanced; in the third section of the piecewise function, the task performance score is high, and the task allocation strategy will be more aggressive.

[0088] It should be noted that the minimum threshold P of the task performance score is min , represents the minimum requirement standard for task allocation, which is usually set based on historical task data. If the vehicle's task performance score is lower than this value, its task allocation will be significantly reduced. Exemplarily, the minimum threshold can be set to 0.4 or 0.5.

[0089] The medium threshold P for the task performance score mid , represents an intermediate level of task performance, which can be the average of the task performance of all vehicles, or a reasonable estimate of the task success rate. Exemplarily, the medium threshold can be set in the range of 0.7-0.8.

[0090] The initial coefficients of the segment k, including A1, A2 and A3, are used to set the initial amplitude of the task allocation in each stage, which is determined according to the priority of the task and the relative scarcity of vehicle resources. The initial coefficients can start from a small value. For example, A1, A2 and A3 can be set to 0.2, 0.5 and 0.8 respectively to ensure that the allocation in different stages is clearly different.

[0091] The adjustment coefficients for the sensitivity of the vehicle task performance fluctuations in the task allocation process within the segment k include B1 and B2, which are used to control the sensitivity of the task allocation to the vehicle performance fluctuations, and the value range is between 0-1. A higher adjustment coefficient value indicates sensitivity to performance changes and is suitable for rapid response to fluctuations; a lower adjustment coefficient value is suitable for more stable task allocation. For tasks with large environmental changes, a higher value can be selected, for example, B1=0.8, B2=1.0; for tasks with relatively stable environments, a lower value can be selected, for example, B1=0.4, B2=0.6.

[0092] The smoothing coefficients affecting the segment turning point in the segment k include C1, C2 and C3, which are used to control the smooth transition of task allocation at different stages, and can ensure that the allocation changes gradually when the comprehensive task performance is close to the inter-segment value to avoid sudden changes. The larger the value, the smoother the transition, and the value range is between 0.5-2. For example, in order to ensure a smooth transition between segments, C1, C2 and C3 are set to 1.0, 1.5 and 2.0 respectively.

[0093] Step S2202: Based on the task performance score piecewise function and the task performance score of each vehicle, the initial task allocation of each vehicle is calculated, and the formula is:

[0094]

[0095] Among them, T assign (i) represents the initial task allocation of the i-th vehicle; ω k represents the weight coefficient of the kth segment function; β k represents the adjustment coefficient of the k-th segment function; γ k represents the smoothing factor of the k-th segment function; f k (P i ) represents the kth piecewise function of the i-th vehicle in the vehicle cluster.

[0096] Specifically, based on the task performance scoring piecewise function, a dynamic feedback mechanism is established by combining the piecewise functions in three different intervals and the adjustment parameters in the adaptive game model, and the task allocation of each vehicle in the vehicle cluster is dynamically adjusted according to the real-time task performance and historical data of the vehicle.

[0097] It should be noted that the weight coefficient ω of the kth segment function is k , satisfying ω1+ω2+ω3=1.

[0098] The smoothing factor γ of the k-th segment function k , which is used to control the smooth transition of task allocation within a segment. For example, it can be set to 0.5 to ensure that the segmentation is moderately smooth.

[0099] The adjustment coefficient β of the k-th segment function k , used to adjust the change range of task allocation within the segment. Its value needs to adapt to the task requirements. For example, it can be set to 1.0 to ensure that the allocation change has a significant regulatory effect.

[0100] Furthermore, the data obtained by each vehicle during the task execution process also includes real-time task performance data.

[0101] Specifically, the real-time task performance data is generated by the real-time sensor data of the vehicle during the task execution process, based on the data automatically collected by the vehicle's task performance system, including but not limited to real-time data change indicators such as the vehicle's current speed, current position, real-time fuel consumption, as well as the current task progress and vehicle communication status.

[0102] Step S2203: Based on the real-time task performance data of each vehicle during the task execution process, feedback adjustment is performed on the initial task allocation of each vehicle, and the adjusted task allocation of each vehicle is calculated.

[0103] Furthermore, the real-time task performance data obtained by each vehicle during the task execution process is standardized to obtain a standardized real-time task performance index value.

[0104] Specifically, the real-time task performance data is standardized and converted to the same standard as the data in the task performance data set to participate in subsequent calculations. The standardization method, for example, can use the standard deviation standardization method or the Min-Max standardization method, which is not specified in this embodiment. It should be noted that before the real-time task performance data is standardized, it is also necessary to perform denoising and smoothing processing to remove noise and outliers in the data and make the data smoother.

[0105] Furthermore, based on the standardized real-time task performance index value of each vehicle, the adjusted task allocation of each vehicle is calculated using the following formula:

[0106]

[0107] Where M represents the number of data types in the real-time task performance data; δ m represents the real-time feedback weight coefficient of the mth type of data in the real-time task performance data; ξ m represents the sensitivity coefficient of the mth type of data in the real-time task performance data; κ m represents the smoothing coefficient of the mth data in the real-time task performance data; V i ′ represents the real-time task performance index value of the i-th vehicle in the vehicle cluster; V avg Represents the mean value of the real-time task performance index of all vehicles in the vehicle cluster.

[0108] Specifically, the number of data types M in the real-time task performance data represents different task performance indicators, which are used to set independent feedback adjustments for data in different real-time task performance data so as to accurately control task allocation adjustments. For example, when the real-time task performance data includes three dimensions: speed, fuel efficiency, and response time, M is 3.

[0109] Sensitivity coefficient ξ of the mth data in the real-time task performance data m , which is used to determine the magnitude of feedback regulation on task allocation. Exemplarily, its value can be set to 0.1 to moderately adjust the feedback intensity.

[0110] The smoothing coefficient κ of the mth data in the real-time task performance data m , used to control the smoothness of the feedback response. For example, its value can be set to 0.5 to ensure that the adjustment process is smooth when the task performance changes.

[0111] Furthermore, based on the task performance score of each vehicle, resources of each vehicle are dynamically allocated in the current task cycle, including steps S2211 and S2212:

[0112] Step S2211: Based on the task performance score of each vehicle, the task performance difference of each vehicle is evaluated to obtain the initial resources allocated to each vehicle, and the formula is as follows:

[0113]

[0114] Among them, R i represents the initial resources allocated to the i-th vehicle; J represents the total number of vehicles in the vehicle cluster; ε j represents the fluctuation impact value of controlling the task performance score of the jth vehicle; P i and P j denote the task performance scores of the i-th vehicle and the j-th vehicle respectively; σ j represents the influence coefficient of smoothing the resource allocation of the j-th vehicle; represents the impact of adjusting the task score on the resource allocation of the i-th vehicle; P avg Represents the mean score of task performance in the vehicle cluster.

[0115] Specifically, based on the task performance score of each vehicle, an adaptive game model algorithm is used to dynamically allocate the initial resources of each vehicle by considering the relative task performance scores between vehicles and their relationship with the average task performance score of the entire cluster.

[0116] It should be noted that the fluctuation influence value ε of the task performance score of the jth vehicle is controlled j , which is used to control the fluctuation impact of the vehicle's task performance score, can be set to 0.3, for example, to achieve moderate fluctuation control.

[0117] The influence coefficient σ of smoothing the resource allocation of the jth vehicle j , used to control the adjustment intensity of resource allocation. For example, it can be set to 0.5 to ensure the smoothness of task allocation.

[0118] Adjust the impact of task score on resource allocation of the i-th vehicle Exemplarily, it can be set to 1.0 to balance the impact of task scoring on resource allocation.

[0119] Step S2212: dynamically adjust the initial resource allocation of each vehicle using the following formula to calculate the adjusted resource allocation of each vehicle:

[0120]

[0121] Among them, R′ i ′ represents the resources allocated to the i-th vehicle after adjustment; τ represents the resource allocation adjustment coefficient; P i represents the task performance score of the i-th vehicle; P avg represents the mean score of task performance in the vehicle cluster; P max represents the highest task performance score in the vehicle cluster; P min represents the lowest task performance score in the vehicle cluster; P thresh represents the preset task performance score threshold; η represents the adjustment range of resource allocation.

[0122] Specifically, resource allocation is dynamically adjusted by considering the vehicle's task performance score relative to the performance of the entire vehicle cluster, and resource allocation is further increased or decreased based on whether the vehicle exceeds a preset task performance score threshold.

[0123] It should be noted that the resource allocation adjustment coefficient τ has a value range of 0-1 and is used to adjust the sensitivity of resource allocation to ensure the rationality and stability of resource allocation. Exemplarily, it can be set to 0.4 to achieve a moderate resource adjustment response.

[0124] The preset task performance score threshold P thresh , which is used to judge whether the vehicle performs well in order to determine the increase or decrease trend of resource allocation. It can be set to the median or average of all vehicle task performance scores based on historical data. For example, it can be set to 0.7 to define the boundary between excellent and average performance.

[0125] The adjustment range η of resource allocation is used to control the increase or decrease of resource allocation. Its value range is between 0-0.5 to ensure the stability of allocation adjustment. Exemplarily, it can be set to 0.2 to make the increase or decrease of resource allocation have moderate flexibility.

[0126] By adopting a resource allocation mechanism driven by an adaptive game algorithm to evaluate the differences in task performance between vehicles, resource allocation is adjusted based on the task performance scores. Vehicles that perform above the preset threshold receive more resources, while vehicles that perform below the preset threshold receive less resources. This can optimize the reasonable allocation of resources in the vehicle cluster.

[0127] Step S3: taking a preset number of task cycles as an observation window, and based on the fluctuation value of the task performance score of each vehicle during each task cycle within the observation window, progressively adjusting the resources allocated to each vehicle.

[0128] Specifically, for vehicles whose performance fluctuates greatly during the execution of tasks, a progressive penalty mechanism is applied. By setting an observation window period, the performance of the vehicle during the observation window period and the recovery of the performance are observed. If the vehicle recovers within the window period, the existing resource allocation is maintained. If the performance continues to decline, its tasks and resource allocation are gradually reduced. At the same time, setting an observation window period can prevent the system from overreacting to a single or a few poor performances. The observation window period allows the system to observe performance over a period of time, thereby reducing sensitivity to incidental events and avoiding overly aggressive adjustments due to short-term abnormal performance. Exemplarily, the number of task cycles in the observation window period can be set to 5.

[0129] Furthermore, based on the task performance scores of each vehicle in the current task cycle and the previous task cycle, the task performance fluctuation amplitude value of each vehicle in the current task cycle is calculated using the following formula:

[0130]

[0131] Among them, △Q current,i represents the task performance fluctuation amplitude value of the current task cycle of the i-th vehicle; P current,i represents the task performance score of the i-th vehicle in the current task cycle; P prev,i represents the task performance score of the i-th vehicle in the previous task cycle.

[0132] Specifically, the task performance fluctuation amplitude value is used to measure the stability and consistency of the vehicle's task performance within a certain period of time. current,i When the value is positive or does not fluctuate, the calculated resource allocation is maintained unchanged at the beginning of the next observation window; when the vehicle's performance fluctuates △Q current,i If the value is continuously negative at the beginning of the observation window period, but the performance fluctuation recovers in the later period, that is, the fluctuation becomes positive or does not fluctuate, it is regarded as fluctuation recovery. Similarly, at the beginning of the next observation window period, the calculated resource allocation amount is maintained unchanged.

[0133] Furthermore, when the task performance fluctuation amplitude value of each vehicle in the observation window period is continuously negative, based on the task performance fluctuation amplitude mean, the penalty value of each vehicle in the current task cycle is calculated using the following formula to reduce the resource allocation value of each vehicle in the next observation window period:

[0134]

[0135] Among them, △Q i represents the mean fluctuation of the task performance of the i-th vehicle during the observation window; μ represents the penalty coefficient.

[0136] Specifically, when the performance fluctuation of the vehicle continues to be negative, at the beginning of the next observation window, based on the above penalty value, the resource allocation of the vehicle is adjusted using the following formula:

[0137]

[0138] Among them, R i,adj Represents the resource allocation of the i-th vehicle after adjustment based on the penalty value.

[0139] It should be noted that when the amount of resource allocation adjusted based on the penalty value reaches the preset minimum resource allocation threshold, the penalty is stopped and the allocated resource amount is maintained at the preset minimum resource allocation threshold. Exemplarily, the preset minimum resource allocation threshold can be set to 0.3 to ensure that the vehicle can maintain basic operation under the minimum resource allocation.

[0140] In this embodiment, through a progressive penalty mechanism, resource allocation is adjusted when vehicle performance fluctuates greatly, so as to make reasonable use of resources and avoid excessive impact on resource allocation due to short-term fluctuations.

[0141] In summary, the vehicle cluster data fusion and resource allocation method based on the adaptive game algorithm according to the embodiment of the present invention has the following beneficial effects:

[0142] 1. The present invention uses a multi-dimensional task performance data set to comprehensively consider task completion time, task success rate, historical performance data and real-time change indicators to conduct a more comprehensive and dynamic evaluation of the real-time task performance of each vehicle in the cluster, greatly improving the accuracy of task allocation and ensuring more efficient and reasonable resource allocation in complex environments.

[0143] 2. The present invention introduces an adaptive game algorithm to drive the resource allocation of vehicles in the cluster. In the resource allocation process, a dynamic game algorithm is used to achieve fair resource competition and cooperation among vehicles, solving the problem of uneven resource allocation in traditional centralized command systems. The system can not only allocate more resources to vehicles with outstanding performance, but also dynamically adjust resource allocation according to the real-time performance of the vehicle, ensuring maximum utilization of vehicle resources.

[0144] 3. The present invention uses a progressive penalty mechanism to flexibly handle resource allocation when task performance fluctuates, and sets an observation window to avoid excessive resource reduction due to short-term performance fluctuations. Compared with the existing methods that often take immediate penalties or resource reduction measures for vehicles with poor performance, which easily lead to resource allocation imbalance or task failure, the present invention effectively improves the robustness and flexibility of the system.

[0145] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A vehicle cluster data fusion and resource allocation method based on an adaptive game algorithm, characterized in that: The steps include: Acquire and pre-process the multi-dimensional task data of several task cycles generated by each vehicle in the vehicle cluster during the task execution process to obtain a task performance data set of each vehicle in the vehicle cluster in each task cycle; wherein the task performance data set includes: task completion time, task success rate, historical task performance data and vehicle status data; In a task cycle: Performing data fusion on the task performance data set of each vehicle in the current task cycle to obtain a task performance score of each vehicle; Based on the task performance scores of each vehicle, the task execution capability of each vehicle in the current task cycle is evaluated, and tasks and resources of each vehicle are dynamically allocated; The preset number of task cycles is used as an observation window, and the resources allocated to each vehicle are progressively adjusted based on the fluctuation value of the task performance score of each vehicle during each task cycle within the observation window.

2. The method according to claim 1, characterized in that: The step of fusing the task performance data set of each vehicle in the current task cycle to obtain a task performance score of each vehicle includes: Standardizing each data in the task performance data set of the current task cycle to obtain a standardized value of each task performance index; The standardized task performance index values ​​are weighted and summed to obtain the task performance score of the vehicle in the current task cycle; wherein the weight value of each task performance is calculated based on the following formula: X={T,S,H,R} in, represents the weight coefficient of the i-th vehicle on the task performance index X; T represents the task completion time index; S represents the task success rate index; H represents the historical performance data index; R represents the vehicle status data index; α X represents the adaptive adjustment parameter of the task performance index X; X represents the smoothing coefficient of the task performance indicator X; X i represents the value of the task performance index X of the i-th vehicle; J represents the total number of vehicles in the vehicle cluster; X j represents the value of the task performance index X of the jth vehicle; Represents the mean of the task performance index X of all vehicles in the vehicle cluster.

3. The method according to claim 2, characterized in that: The data obtained by each vehicle during the mission execution also includes real-time mission performance data; Based on the task performance score of each vehicle, the task execution capability of each vehicle in the current task cycle is evaluated, and tasks of each vehicle are dynamically assigned, including: Constructing a task performance score piecewise function based on the task performance score of each vehicle, a preset minimum threshold and a medium threshold of the task performance score; Based on the task performance score piecewise function and the task performance score of each vehicle, calculating an initial task allocation amount for each vehicle; Based on the real-time task performance data of each vehicle during the task execution process, feedback adjustment is performed on the initial task allocation amount of each vehicle, and the adjusted task allocation amount of each vehicle is calculated.

4. The method according to claim 3, characterized in that: The task performance scoring piecewise function is expressed as: Among them, P i represents the task performance score of the i-th vehicle; P min represents the minimum threshold of task performance score; P mid Indicates the medium threshold for task performance rating; A k represents the initial coefficient of segment k; B k represents the adjustment coefficient of the sensitivity of segment k to the fluctuation of vehicle task performance during the task allocation process; C k Represents the smoothing coefficient that affects the segment turning point within segment k.

5. The method according to claim 4, characterized in that: The initial task allocation for each vehicle is calculated using the following formula: Among them, T assign (i) represents the initial task allocation of the i-th vehicle; ω k represents the weight coefficient of the kth segment function; β k represents the adjustment coefficient of the k-th segment function; γ k Represents the smoothing factor of the k-th segment function.

6. The method according to claim 5, characterized in that: Standardize the real-time task performance data obtained by each vehicle during the task execution process to obtain a standardized real-time task performance index value; Based on the standardized real-time task performance index value of each vehicle, the adjusted task allocation of each vehicle is calculated using the following formula: Where M represents the number of data types in the real-time task performance data; δ m represents the real-time feedback weight coefficient of the mth type of data in the real-time task performance data; ξ m represents the sensitivity coefficient of the mth type of data in the real-time task performance data; k m represents the smoothing coefficient of the mth data in the real-time task performance data; V i ′ V represents the real-time task performance index value of the i-th vehicle in the vehicle cluster; avg Represents the mean value of the real-time task performance index of all vehicles in the vehicle cluster.

7. The method according to claim 2, characterized in that: Based on the mission performance score of each vehicle, the resources of each vehicle are dynamically allocated during the current mission cycle, including: Based on the task performance scores of each vehicle, the task performance differences of each vehicle are evaluated to obtain the initial resources allocated to each vehicle; Based on the initial resources of each vehicle, the task performance score of each vehicle and the highest and lowest task performance scores among the cluster vehicles, the initial resource allocation amount of each vehicle is dynamically adjusted to calculate the adjusted resource allocation amount of each vehicle.

8. The method according to claim 7, characterized in that: Use the following formula to get the initial resources allocated to each vehicle: Among them, R i represents the initial resources allocated to the i-th vehicle; J represents the total number of vehicles in the vehicle cluster; ε j represents the fluctuation impact value of controlling the task performance score of the jth vehicle; P i and P j denote the task performance scores of the i-th vehicle and the j-th vehicle respectively; σ j represents the influence coefficient of smoothing the resource allocation of the jth vehicle; represents the impact of adjusting the task score on the resource allocation of the i-th vehicle; P avg Represents the mean score of task performance in the vehicle cluster.

9. The method according to claim 8, characterized in that: Use the following formula to dynamically adjust the initial resource allocation of each vehicle and calculate the adjusted resource allocation of each vehicle: Among them, R ′ i ′ represents the resources allocated to the i-th vehicle; τ represents the resource allocation adjustment coefficient; P i represents the task performance score of the i-th vehicle; P avg represents the mean score of task performance in the vehicle cluster; P max represents the highest task performance score in the vehicle cluster; P min represents the lowest task performance score in the vehicle cluster; P thresh represents the preset task performance score threshold; η represents the adjustment range of resource allocation.

10. The method according to any one of claims 1 to 9, characterized in that: The stepwise adjustment of resources allocated to each vehicle includes: Based on the task performance scores of each vehicle in the current task cycle and the previous task cycle, the task performance fluctuation amplitude value of each vehicle in the current task cycle is calculated; When the task performance fluctuation amplitude value of each vehicle in the observation window period is continuously negative, based on the task performance fluctuation amplitude mean, the penalty value of each vehicle in the current task cycle is calculated using the following formula to reduce the resource allocation value of each vehicle in the next observation window period: Among them, △Q i represents the mean fluctuation of the task performance of the i-th vehicle during the observation window; μ represents the penalty coefficient.

Images