Multi-objective Intelligent Planning Method and Product Based on Spatiotemporal Cognition Mechanism

Abstract

The present invention provides a multi-objective intelligent planning method and product based on a spatio-temporal cognitive mechanism, relating to the technical field of intelligent planning. In the present invention, spatio-temporal weights are calculated by considering time weights and space weights to dynamically integrate time information (time windows of task points) and space information (spatial positions of task points), thereby enhancing the comprehensiveness of task planning. By adopting the technical solution provided by the implementation of the present invention, it is possible to plan for an unmanned aerial vehicle to efficiently visit multiple task points within a given time in complex terrains (such as mountains, cities or forests), while taking into account energy consumption, path length and task priorities, so as to achieve comprehensive coverage and dynamic adjustment of tasks. By adopting the technical solution provided by the embodiments of the present invention, it is possible to significantly improve the speed and success rate of target tracking under conditions of resource constraints and complex environments, providing an efficient and accurate technical solution for intelligent target tracking of unmanned aerial vehicles in high-speed motion scenarios.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent planning, and in particular to a multi-objective intelligent planning method and product based on a spatio-temporal cognitive mechanism. Background Art

[0002] In complex terrains (such as mountains, cities or forests), an unmanned aerial vehicle (UAV) needs to efficiently visit multiple task points within a given time, while taking into account energy consumption, path length and task priority to achieve comprehensive coverage and dynamic adjustment of tasks. However, due to the optimization complexity of time window constraints, the time windows of multiple objectives may overlap or conflict. Moreover, task points and obstacles in a complex environment may change over time, and the computing resources of the UAV are limited, making it difficult to complete task planning and execution under the constraints of limited computing power and energy consumption. Summary of the Invention

[0003] The present invention provides a multi-objective intelligent planning method and product based on a spatio-temporal cognitive mechanism to at least partially solve the above problems.

[0004] In a first aspect of the present invention, a multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism is provided, and the method includes:

[0005] Modeling the characteristics of the time-dependent functions of the multiple task points changing with time according to the respective time windows of the multiple task points, to obtain the respective time characteristic functions of the multiple task points;

[0006] Determining the respective time weights of the multiple task points according to the respective time windows of the multiple task points and the respective time characteristic functions of the multiple task points;

[0007] Modeling the attraction relationship between a task point and other task points according to the spatial positions of the multiple task points, to determine the distribution intensity function of the task point, where the attraction relationship between task points is used to characterize the degree of being on the same route between task points;

[0008] For each task point among the multiple task points, obtaining the spatial weight of the task point according to the distribution intensity function of the task point;

[0009] Obtaining the spatio-temporal weight of the task point according to the respective time weights and spatial weights of the multiple task points;

[0010] Planning the execution order of the multiple task points with the highest task completion efficiency as the goal according to the respective spatio-temporal weights of the multiple task points.

[0011] Optionally, modeling the characteristics of the time-dependent functions of the multiple task points changing with time according to the respective time windows of the multiple task points, to obtain the respective time characteristic functions of the multiple task points, includes:

[0012] Determine the time-dependent functions of the multiple task points according to the time windows of the multiple task points respectively. The time-dependent function of a task point is used to characterize whether any moment is within the time window of this task;

[0013] Use a fractional differential equation and the time-dependent functions of the multiple task points respectively to model the characteristics of the time-dependent functions of the multiple task points changing with time, and obtain the time characteristic functions of the multiple task points respectively.

[0014] Optionally, for each task point among the multiple task points, obtain the spatial weight of this task point according to the distribution intensity function of this task point, including:

[0015] For each task point among the multiple task points, perform feature decomposition of the distribution intensity function of this task point at multiple scales to obtain the weight corresponding to each scale and the feature basis function of this task point at each scale;

[0016] For each task point among the multiple task points, obtain the spatial weight of this task point according to the feature basis function of this task point at each scale and the distances between this task point and each task point at this scale.

[0017] Optionally, the method further includes:

[0018] For any task point among the multiple task points, after detecting that the spatial position of this task point is updated, determine the updated spatial weights and updated spatio-temporal weights of the multiple task points respectively according to the updated spatial position of this task point;

[0019] According to the updated spatio-temporal weights of the multiple task points respectively, with the goal of the highest task completion efficiency, re-plan the execution order of the multiple task points.

[0020] Optionally, the method further includes:

[0021] After detecting a new task point, determine the spatio-temporal weight of this new task point, and, according to the spatial position of this new task point, determine the updated spatial weights and updated spatio-temporal weights of the multiple task points respectively;

[0022] According to the updated spatio-temporal weights of the multiple task points respectively, and the spatio-temporal weight of the new task point, with the goal of the highest task completion efficiency, re-plan the execution order of the multiple task points.

[0023] Optionally, the method further includes:

[0024] After detecting that any one of the multiple task points is deleted, determine the updated spatial weights and updated spatio-temporal weights of the remaining task points respectively;

[0025] According to the updated spatio-temporal weights of the remaining task points respectively, with the goal of maximizing the task completion efficiency, re-plan the execution order of the remaining task points.

[0026] Optionally, the method further includes:

[0027] For any one of the multiple task points, after detecting that the time window of this task point is updated, according to the updated time window of this task point, determine the updated time weight and updated spatio-temporal weight of this task point;

[0028] According to the updated spatio-temporal weight of this task point and the spatio-temporal weights of other task points respectively, with the goal of maximizing the task completion efficiency, re-plan the execution order of the multiple task points.

[0029] Optionally, after planning the execution order of the multiple task points, the method further includes:

[0030] According to the execution order of the multiple task points, determine every two adjacent task points in the execution order;

[0031] According to multiple paths between every two adjacent task points in the execution order, plan the optimal path for traversing the multiple task points.

[0032] The second aspect of the present invention provides a multi-objective intelligent planning device based on a spatio-temporal cognitive mechanism. The multi-objective intelligent planning device based on a spatio-temporal cognitive mechanism includes:

[0033] A time modeling module, configured to model the characteristics of the time-dependent function of each of the multiple task points changing with time according to the time window of each of the multiple task points, and obtain the time characteristic function of each of the multiple task points;

[0034] A time weight determination module, configured to determine the time weights of the multiple task points respectively according to the time window of each of the multiple task points and the time characteristic function of each of the multiple task points;

[0035] A space modeling module, configured to model the attraction relationship between this task point and other task points according to the spatial positions of the multiple task points, and determine the distribution intensity function of this task point. The attraction relationship between task points is used to characterize the degree of being on the way between task points;

[0036] A space weight determination module, configured to obtain the space weight of each task point among the multiple task points according to the distribution intensity function of this task point;

[0037] The spatio-temporal weight determination module obtains the spatio-temporal weight of each task point according to the respective time weight and space weight of the multiple task points;

[0038] The planning module is configured to plan the execution order of the multiple task points with the highest task completion efficiency according to the respective spatio-temporal weights of the multiple task points.

[0039] The third aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes, it implements the multi-objective intelligent planning method based on the spatio-temporal cognitive mechanism as described in the first aspect of the present invention.

[0040] The fourth aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the multi-objective intelligent planning method based on the spatio-temporal cognitive mechanism as described in the first aspect of the present invention.

[0041] The fifth aspect of the present invention provides a computer program product, including a computer program / instructions. When the computer program / instructions are implemented by a processor, they implement the steps in the multi-objective intelligent planning method based on the spatio-temporal cognitive mechanism as described in the first aspect of the present invention.

[0042] The technical solution provided by the embodiments of the present invention can solve the problem of efficient task planning and execution of intelligent devices (such as drones) in complex dynamic environments, and is particularly suitable for multi-objective task allocation and path optimization with time window constraints. By adopting the technical solution provided by the embodiments of the present invention, it is possible to plan drones to efficiently visit multiple task points within a given time in complex terrains (such as mountains, cities, or forests), while taking into account energy consumption, path length, and task priorities, and achieving comprehensive coverage and dynamic adjustment of tasks.

[0043] By adopting the technical solution provided by the embodiments of the present invention, it is possible to significantly improve the speed and success rate of target tracking under resource-constrained and complex environments, providing an efficient and accurate technical solution for intelligent target tracking of drones in high-speed motion scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for the description of the present invention will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0045] Figure 1 is a flowchart of the steps of the multi-objective intelligent planning method based on the spatio-temporal cognitive mechanism provided by the present invention;

[0046] Figure 2 It is a schematic diagram for performing feature decomposition of the distribution intensity function of task points at multiple scales in the multi-objective intelligent planning method based on the spatio-temporal cognitive mechanism provided by the present invention;

[0047] Figure 3 It is a flowchart of the steps of the multi-objective intelligent planning method based on the spatio-temporal cognitive mechanism provided by the present invention;

[0048] Figure 4 It is a schematic diagram of the process of the multi-objective intelligent planning method based on the spatio-temporal cognitive mechanism provided by the present invention;

[0049] Figure 5 It is a structural block diagram of the multi-objective intelligent planning device based on the spatio-temporal cognitive mechanism provided by the present invention. Detailed implementation manners

[0050] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation manners.

[0051] In the related art, a multi-objective path planning based on dynamic programming is proposed. This method obtains the optimal path for multi-objective access through the decomposition of tasks and the recursive calculation of path states. However, as the number of task points increases, the dimension of the state space will increase exponentially, resulting in a significant increase in computational complexity. At the same time, this method cannot effectively handle the dynamically changing time window constraints and lacks flexibility and real-time performance.

[0052] The intelligent planning methods proposed in the related art often model the time and space characteristics independently and lack the joint analysis of spatio-temporal information. For example, although the dynamic programming method can optimize the path, it is difficult to handle the time window constraints, resulting in insufficient accuracy of the planning results in complex dynamic scenarios. The change of the time window will trigger frequent recalculations, further increasing the computational overhead of the system.

[0053] Reinforcement learning and dynamic programming methods have high computational requirements in high-dimensional task scenarios. Especially when the number of task points is large or the path length is long, the calculation time will increase significantly. For an embedded UAV platform with limited resources, this computational bottleneck will lead to planning delays, thereby reducing the overall real-time performance and reliability of the task.

[0054] In complex dynamic scenarios (such as changes in task point positions or dynamic adjustments of time windows), the related art lacks the ability to quickly update the spatio-temporal model, resulting in frequent recalculations for task planning. In this case, the UAV may miss the time window or select a sub-optimal path, further reducing the task completion rate and efficiency.

[0055] In related technologies, it is also proposed to use reinforcement learning for multi-objective task scheduling. Through deep reinforcement learning, drones can learn optimal task scheduling strategies in a dynamic environment, such as through Proximal Policy Optimization (PPO) or Deep Q-Network (DQN) methods. Although reinforcement learning methods have good adaptability, when facing time window constraints and complex dynamic scenarios, the convergence speed is slow and a large amount of computing resources are required, making it difficult to meet the real-time requirements of embedded devices.

[0056] Based on this, the embodiments of the present invention provide a solution that can achieve efficient task planning under complex dynamic scenarios and time constraints, and specifically propose a multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism. The embodiments of the present invention build a dynamic spatio-temporal cognitive model, and use fractional differential equations (FDE) and generalized Poisson equations to model and optimize the time characteristics and spatial characteristics of task points, improving the efficiency and accuracy of task scheduling and path planning.

[0057] As Figure 1 shown, it shows a flowchart of the steps of a multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism provided by the embodiments of the present invention. Specifically, the method includes the following steps:

[0058] S101, according to the time windows of multiple task points, model the characteristics of the time-dependent functions of the multiple task points changing with time, and obtain the time characteristic functions of the multiple task points.

[0059] In the embodiments of the present invention, the time window refers to the time interval during which each task point is allowed to be executed (for example, the task point is only valid within , where represents the start point of the time interval during which the task point is allowed to be executed, and represents the end point of the time interval during which the task point is allowed to be executed).

[0060] The time-dependent function is used to describe the function of the effectiveness of the task point changing with time. In the embodiments of the present invention, if the time t is not within the time window of the task point, the effectiveness of the task point is invalid and can be represented as 0; if the time t is within the time window of the task point, the effectiveness of the task point is valid and can be represented as 1.

[0061] The time characteristic function is a mathematical expression obtained through modeling that can quantify the time dependence of the effectiveness of task points and describe the non-linear dynamic characteristics of task effectiveness changing over time. The time characteristic function characterizes the impact of the time window constraint of task points on the priority of task points (for example: the less remaining time of a task point, the more urgent the task, and the higher the task priority).

[0062] Specifically, the step S101 includes the following sub-steps:

[0063] S1011, according to the time windows of multiple task points respectively, determine the time dependence functions of the multiple task points respectively. The time dependence function of a task point is used to characterize whether it is within the time window of the task at any moment.

[0064] Specifically, the time window of the task point can be expressed as a time dependence function :

[0065]

[0066] where:

[0067] represents the step function, which is used to define the start and end ranges of the time window;

[0068] represents whether the task point is within the effective time window at time t.

[0069] In the embodiment of the present invention, time window parsing converts the time window constraint into a mathematical expression, providing a basis for subsequent time modeling.

[0070] S1012, use the fractional differential equation and the time dependence functions of the multiple task points respectively to model the characteristics of the time dependence functions of the multiple task points changing over time, and obtain the time characteristic functions of the multiple task points respectively.

[0071] In the embodiment of the present invention, the fractional differential equation is used to capture the time-varying characteristics of the task, and the memory and historical dependence on the time characteristics are controlled through the order of the fractional derivative.

[0072] In the embodiment of the present invention, the fractional differential model introduces historical memory and can better describe the dynamic changes of the task time window. Specifically, by combining the time dependence function with the fractional differential equation, the time evolution characteristics of task effectiveness can be more finely characterized, especially the memory effect and the gradual change process during the transition inside and outside the time window.

[0073] S102. Determine the time weights of the multiple task points according to the time windows and the time characteristic functions of the multiple task points respectively.

[0074] In an embodiment of the present invention, the time characteristic functions of the multiple task points can be integrated according to the time windows of the multiple task points respectively to obtain the time weights of the task points respectively. The time weights of each task point can be used for subsequent optimization of task scheduling.

[0075] Taking the "complex terrain" as an example of an urban scene with multiple building obstacles, assuming that the intelligent device currently performing the task is a drone, the task is package delivery, and the task points include task point A (customer A) and task point B (customer B). The package of customer A needs to be delivered between 10:00 and 11:00, and the package of customer B needs to be delivered between 10:30 and 12:00. Then, when the current time is 10:30, the time weight of task point A is higher than that of task point B.

[0076] In an embodiment of the present invention, time perception modeling is completed based on steps S101 to S102. Through time perception modeling, the distribution law of the time windows of the task points is analyzed, the dependence relationship and dynamic characteristics of the time windows of each task point are captured, and task priorities are generated to ensure that the tasks are efficiently completed within the time window constraints of this task point.

[0077] S103. Model the attraction relationship between this task point and other task points according to the spatial positions of the multiple task points, and determine the distribution intensity function of this task point. The attraction relationship between task points is used to characterize the degree of being on the way between task points.

[0078] In an embodiment of the present invention, the spatial distribution positions of the task points can be represented as a Poisson field with boundary conditions, and the attraction relationship between the task points is described through the potential field of the task points. Based on the potential field of the task points, the density function of the task points can be determined to reflect the distribution intensity of the task points. In an embodiment of the present invention, the density function of the task points is used as the distribution intensity function of the task points.

[0079] In an embodiment of the present invention, the spatial positions of each task point can be represented by position vectors.

[0080] In an embodiment of the present invention, the boundary conditions can be determined according to the allowable flight boundary of the drone.

[0081] In an embodiment of the present invention, after constructing the potential field model, the attraction relationship between the task points can be modeled to provide support for path optimization.

[0082] In an embodiment of the present invention, the attraction relationship between task points is determined by the spatial position distribution of each task point, and characterizes the degree of being on the way between task points.

[0083] In the embodiments of the present invention, a potential field can be generated for each task point, and its intensity decays with distance, reflecting the range and intensity of its attraction. The superposition of the potential fields of all task points can form a total potential field. Based on the total potential field, the distribution intensity function of each task point can be obtained.

[0084] By modeling the attraction relationship between task points through a potential field and transforming the physical potential field idea into a mathematical tool, dynamic and adaptive task scheduling and path planning can be achieved.

[0085] Modeling the spatial distribution of task points as a Poisson field with boundary conditions, the spatial constraints and the attraction relationship between task points can be encoded into the density function through partial differential equations.

[0086] S104. For each of the multiple task points, according to the distribution intensity function of this task point, obtain the spatial weight of this task point.

[0087] In the embodiments of the present invention, the distribution intensity function of each task point can be first subjected to feature decomposition at multiple scales to obtain the feature basis functions of the task points considering different local features respectively. Then, according to the feature basis functions at different scales and the distances between this task point and other task points at this scale, obtain the spatial weight of this task point.

[0088] Specifically, the step S104 includes the following sub-steps:

[0089] S1041. For each of the multiple task points, perform feature decomposition on the distribution intensity function of this task point at multiple scales to obtain the weight corresponding to each scale and the feature basis function of this task point at each scale.

[0090] In the embodiments of the present invention, the distribution intensity function of each task point considers the current global distribution features, and these global distribution features consider not only the distances between task points but also influencing factors such as obstacles.

[0091] In the embodiments of the present invention, the Poisson field can be subjected to feature decomposition to extract the multi-scale spatial characteristics of task points. Specifically, the distribution intensity function of task points can be decomposed to obtain the feature intensity and feature basis function of task points at different scales, where the feature intensity represents the weight corresponding to each scale.

[0092] Each scale corresponds to a different spatial resolution. For example, a large scale may capture the overall distribution trend, while a small scale focuses on local details.

[0093] Specifically, the localization characteristics of wavelet basis functions can be utilized to decompose the distribution intensity function into components of different scales. By performing multi-scale wavelet decomposition on the distribution intensity function of the task points, the distribution characteristics at different spatial resolutions can be quantified. The scale weights indicate the importance of each level, and the eigenbasis functions reveal the local structure, providing a data-driven decision basis for subsequent intelligent planning.

[0094] Specifically, as Figure 2 shown, where the ellipses of different sizes represent different scales. It can be seen that the features at different scales can capture different spatial feature information. Taking task point A as an example, it can be seen that the spatial feature information at the smallest scale (the first scale) can focus on the local details related to the spatial information of task point A itself. For example, the obstacle information around task point A. The spatial feature information at a scale one larger than the smallest scale (the second scale) can focus on the mutual spatial distribution information between task point A and task point B. For example, the path information from task point A to task point B. The spatial feature information at an even larger scale (the third scale) can focus on the mutual spatial distribution relationship among task point A, task point B, and task point C. Assuming there are obstacles between A and B that seriously affect the traveling speed, the priority order for completing the task points can be determined as A→C→B. The spatial feature information at an even larger scale (the fourth scale) can focus on the mutual spatial distribution relationship among task point A, task point B, task point C, and task point D. Assuming the road between D and B is clear and there are no obstacles, and the unmanned aerial vehicle can maintain a high traveling speed, then the priority order for completing the task points can be determined as A→C→D→B.

[0095] Thus, in the embodiments of the present invention, by decomposing the distribution intensity function of the task points, the eigenbasis functions of the task points at different scales can be obtained, considering the mutual spatial distribution relationship between task points from different scales.

[0096] S1042. For each of the multiple task points, according to the eigenbasis function of the task point at each scale and the distance between the task point and each of the task points at that scale, obtain the spatial weight of the task point.

[0097] In the embodiments of the present invention, extracting the multi-scale features of the spatial distribution of task points can distinguish the local and global target distribution relationships, improving the robustness of the intelligent planning method.

[0098] In the embodiments of the present invention, the multi-scale features of each task point and the distances between the task point and each of the task points at the corresponding scale are summed to obtain the spatial weight of the task point.

[0099] In the embodiments of the present invention, multi-scale feature extraction can take into account local features. For example, there are many obstacles between task point A and task point B, which will affect the flight speed. The straight-line distance between the two points is 1. After local feature extraction of the distance and considering the obstacle features, the distance can be determined to be 2. Task point A and task point C are originally far away, with a straight-line distance of 2, but the middle field of vision is open and the obstacles are relatively low, and the flight speed is fast. After local feature extraction and considering the obstacle features, the distance can be determined to be 1.

[0100] In the embodiments of the present invention, based on steps S103 to S104, spatial cognitive modeling is completed. Spatial cognitive modeling can analyze the spatial distribution law of task points through the generalized Poisson equation, optimize the path planning between targets, and reduce redundant paths.

[0101] S105. Obtain the spatio-temporal weight of the task point according to the time weight and spatial weight of each of the multiple task points.

[0102] In the embodiments of the present invention, by fusing the time weight and spatial weight, a dynamic spatio-temporal cognitive model is constructed to guide task scheduling and path optimization.

[0103] In the embodiments of the present invention, when calculating the spatio-temporal weight, the adjustment coefficients of the time weight and spatial weight can also be considered to balance the influence of time factors and spatial factors on intelligent planning, so as to dynamically integrate time information (the time window of the task point) and spatial information (the spatial position of the task point), and improve the comprehensiveness of task planning.

[0104] S106. According to the spatio-temporal weights of each of the multiple task points, with the goal of the highest task completion efficiency, plan the execution order of the multiple task points.

[0105] In the embodiments of the present invention, a dynamic programming algorithm can be used to generate the optimal task execution order, determine the best task scheduling order, and maximize the task completion efficiency.

[0106] In the embodiments of the present invention, the spatio-temporal weight dynamically integrates time information and spatial information, and can achieve the global optimal solution of task scheduling and path optimization.

[0107] In the embodiments of the present invention, a multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism is also provided. Specifically, as Figure 3 shown, which shows the step flow chart of the multi-objective intelligent planning method based on the spatio-temporal cognitive mechanism. The method includes the following steps:

[0108] S201. According to the time window of each of the multiple task points, model the characteristics of the time-dependent function of each of the multiple task points changing with time, and obtain the time characteristic function of each of the multiple task points.

[0109] S202. Determine the time weights of the multiple task points according to the time windows and time characteristic functions of the multiple task points respectively.

[0110] S203. Model the attraction relationship between a task point and other task points according to the spatial positions of the multiple task points, and determine the distribution intensity function of this task point. The attraction relationship between task points is used to characterize the degree of being on the way between task points.

[0111] S204. For each task point among the multiple task points, obtain the spatial weight of this task point according to the distribution intensity function of this task point.

[0112] S205. Obtain the spatio-temporal weight of this task point according to the time weights and spatial weights of the multiple task points respectively.

[0113] S206. Plan the execution order of the multiple task points with the goal of the highest task completion efficiency according to the spatio-temporal weights of the multiple task points respectively.

[0114] Steps S201 - S206 are similar to the above steps S101 - S106, and will not be elaborated here.

[0115] S207. Determine every two adjacent task points in the execution order according to the execution order of the multiple task points.

[0116] S208. Plan the optimal path for traversing the multiple task points according to the multiple paths between every two adjacent task points in the execution order.

[0117] In the embodiment of the present invention, the ant colony algorithm can be combined with the spatio-temporal cognitive model to generate the optimal path according to the paths between every two adjacent task points in the execution order.

[0118] Specifically, in the embodiment of the present invention, after obtaining the execution order of task points considering the time information and spatial information of each task point based on the above steps S201 - S206, based on this execution order, further combine the spatial positions of each task point and all spatial information within the task execution range (including road information, obstacle information, etc.) to obtain multiple paths between every two adjacent task points in the execution order, and finally plan the optimal path for traversing the multiple task points.

[0119] In the embodiment of the present invention, in the case where the spatial position of any one task point is updated, the time weights and spatial weights of all task points in the whole can be dynamically updated to dynamically plan the execution order of multiple task points in the whole.

[0120] Specifically, the method further includes the following steps:

[0121] S11. For any one of the multiple task points, after detecting that the spatial position of this task point is updated, determine the updated spatial weights and updated spatio-temporal weights of each of the multiple task points according to the updated spatial position of this task point.

[0122] S12. According to the updated spatio-temporal weights of each of the multiple task points, with the goal of maximizing the task completion efficiency, re-plan the execution order of the multiple task points.

[0123] In the embodiments of the present invention, in the case of detecting a newly added task point, the time weights and spatial weights of all task points in the global scope can be dynamically updated based on the time information and spatial information of the newly added task point, so as to dynamically plan the execution order of multiple task points in the global scope, and further dynamically adjust the optimal path for traversing multiple task points in the global scope.

[0124] Specifically, the method further includes the following steps:

[0125] S21. After detecting a newly added task point, determine the spatio-temporal weight of this newly added task point, and according to the spatial position of this newly added task point, determine the updated spatial weights and updated spatio-temporal weights of each of the multiple task points.

[0126] S22. According to the updated spatio-temporal weights of each of the multiple task points and the spatio-temporal weight of the newly added task point, with the goal of maximizing the task completion efficiency, re-plan the execution order of the multiple task points.

[0127] In the embodiments of the present invention, in the case of detecting that any one of the multiple task points is deleted, the time weights and spatial weights of all task points in the global scope can be dynamically updated based on the time information and spatial information of deleting this task point, so as to dynamically plan the execution order of multiple task points in the global scope, and further dynamically adjust the optimal path for traversing multiple task points in the global scope.

[0128] Specifically, the method further includes the following steps:

[0129] S31. After detecting that any one of the multiple task points is deleted, determine the updated spatial weights and updated spatio-temporal weights of the remaining task points respectively.

[0130] S32. According to the updated spatio-temporal weights of the remaining task points respectively, with the goal of maximizing the task completion efficiency, re-plan the execution order of the remaining task points.

[0131] In an embodiment of the present invention, when it is detected that the time window of any task point is updated, the time weights and space weights of all task points in the global scope can be dynamically updated based on the updated time information and space information of this task point, so as to dynamically plan the execution order of multiple task points in the global scope, and further dynamically adjust the optimal path for traversing multiple task points in the global scope.

[0132] Specifically, the method further includes the following steps:

[0133] S41. For any one of the multiple task points, after it is detected that the time window of this task point is updated, according to the updated time window of this task point, determine the updated time weight and the updated space-time weight of this task point.

[0134] S42. According to the updated space-time weight of this task point and the space-time weights of other task points respectively, with the goal of the highest task completion efficiency, re-plan the execution order of the multiple task points, and further dynamically adjust the optimal path for traversing multiple task points in the global scope.

[0135] For the sake of easy understanding, the following Figure 4 further explains the multi-objective intelligent planning method based on the spatio-temporal cognitive mechanism provided by the embodiment of the present invention. As Figure 4 shown, in the embodiment of the present invention, the time window of each task point among the current multiple task points can be obtained as time information, and the spatial position of each task point among the current multiple task points can be obtained as space information, and then time window parsing and spatial distribution identification are performed for all task points. Among them, for the time information, based on time window parsing, the time window constraint of each task point is converted into a mathematical expression, and further time characteristic modeling is performed to obtain the time characteristic function of each task point, and further time-dependent weight optimization is completed to obtain the time weight of each task point. For the space information, based on the spatial distribution representation of all task points, multi-scale feature extraction is performed to obtain the feature basis function of each task point at each scale, and further based on the path optimization weight, the space weight of each task point is obtained. Then, spatio-temporal weight integration is performed based on the time weight and space weight of each task point to obtain the spatio-temporal weight of each task point, and further intelligent planning of global task points considering time information and space information is realized. During the task execution process, path planning and real-time adjustment can also be performed based on dynamic task scheduling information (for example: adding task points, deleting task points, modifying the time information or space information of task points).

[0136] Based on the same inventive concept, the present invention also provides a multi-objective intelligent planning device based on the spatio-temporal cognitive mechanism, as Figure 5As shown, it shows a structural block diagram of a multi-objective intelligent planning device based on a spatio-temporal cognitive mechanism. The multi-objective intelligent planning device 500 based on the spatio-temporal cognitive mechanism includes:

[0137] A time modeling module 501, configured to model the characteristics of the time-dependent functions of the multiple task points changing with time according to the time windows of the multiple task points, and obtain the time characteristic functions of the multiple task points;

[0138] A time weight determination module 502, configured to determine the time weights of the multiple task points according to the time windows of the multiple task points and the time characteristic functions of the multiple task points;

[0139] A space modeling module 503, configured to model the attraction relationship between the task point and other task points according to the spatial positions of the multiple task points, determine the distribution intensity function of the task point, and the attraction relationship between task points is used to characterize the degree of being on the same route between task points;

[0140] A space weight determination module 504, configured to obtain the space weight of each task point among the multiple task points according to the distribution intensity function of the task point;

[0141] A spatio-temporal weight determination module 505, configured to obtain the spatio-temporal weight of the task point according to the time weights and space weights of the multiple task points;

[0142] A planning module 506, configured to plan the execution order of the multiple task points with the goal of the highest task completion efficiency according to the spatio-temporal weights of the multiple task points.

[0143] Optionally, the time modeling module 501 is configured to:

[0144] Determine the time-dependent functions of the multiple task points according to the time windows of the multiple task points. The time-dependent function of a task point is used to characterize whether it is within the time window of the task at any moment;

[0145] Use a fractional differential equation and the time-dependent functions of the multiple task points to model the characteristics of the time-dependent functions of the multiple task points changing with time, and obtain the time characteristic functions of the multiple task points.

[0146] Optionally, the space weight determination module 504 is configured to:

[0147] For each task point among the multiple task points, perform feature decomposition of the distribution intensity function of the task point at multiple scales to obtain the weight corresponding to each scale and the feature basis function of the task point at each scale;

[0148] For each of the multiple task points, based on the eigenfunction of the feature of the task point at each scale and the distances between the task point and each of the task points at that scale, the spatial weight of the task point is obtained.

[0149] Optionally, the apparatus further includes:

[0150] A first detection module, configured to, for any one of the multiple task points, after detecting that the spatial position of the task point is updated, determine the updated spatial weights and updated spatio-temporal weights of each of the multiple task points according to the updated spatial position of the task point;

[0151] A first update module, configured to re-plan the execution order of the multiple task points with the highest task completion efficiency as the goal according to the updated spatio-temporal weights of each of the multiple task points.

[0152] Optionally, the apparatus further includes:

[0153] A second detection module, configured to, after detecting a newly added task point, determine the spatio-temporal weight of the newly added task point, and determine the updated spatial weights and updated spatio-temporal weights of each of the multiple task points according to the spatial position of the newly added task point;

[0154] A second update module, configured to re-plan the execution order of the multiple task points with the highest task completion efficiency as the goal according to the updated spatio-temporal weights of each of the multiple task points and the spatio-temporal weight of the newly added task point.

[0155] Optionally, the apparatus further includes:

[0156] A third detection module, configured to, after detecting that any one of the multiple task points is deleted, determine the updated spatial weights and updated spatio-temporal weights of the remaining task points;

[0157] A third update module, configured to re-plan the execution order of the remaining task points with the highest task completion efficiency as the goal according to the updated spatio-temporal weights of each of the remaining task points.

[0158] Optionally, the apparatus further includes:

[0159] A fourth detection module, configured to, for any one of the multiple task points, after detecting that the time window of the task point is updated, determine the updated time weight and updated spatio-temporal weight of the task point according to the updated time window of the task point;

[0160] A fourth update module, configured to re-plan the execution order of the multiple task points with the goal of maximizing the task completion efficiency according to the updated spatio-temporal weights of the task point and the spatio-temporal weights of other task points respectively.

[0161] Optionally, the device further includes:

[0162] A determination module, configured to determine every two adjacent task points in the execution order of the multiple task points;

[0163] A traversal module, configured to plan an optimal path for traversing the multiple task points according to multiple paths between every two adjacent task points in the execution order.

[0164] Based on the same inventive concept, the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes, the steps in the multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism as described in any of the above embodiments are implemented.

[0165] Based on the same inventive concept, the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps in the multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism as described in any of the above embodiments are implemented.

[0166] Based on the same inventive concept, the present invention provides a computer program product, including a computer program / instructions, and when the computer program / instructions are executed by a processor, the steps in the multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism as described in any of the above embodiments are implemented.

[0167] Each embodiment in this specification is described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other.

[0168] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a device, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0169] The present invention is described with reference to the flowcharts and / or block diagrams of methods, terminal devices (apparatus), and computer program products according to the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combinations of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable terminal devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable terminal devices produce means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0170] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable terminal device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0171] These computer program instructions can also be loaded onto a computer or other programmable terminal device, such that a series of operation steps are executed on the computer or other programmable terminal device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable terminal device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0172] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications to these embodiments once they learn the basic creative concepts. Therefore, the appended claims are intended to be construed as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0173] Finally, it should also be noted that in the present invention, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article or terminal device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or terminal device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or terminal device comprising the element.

[0174] The above has introduced in detail a multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism provided by the present invention. Specific examples are used in the present invention to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism, characterized in that, The method includes: Modeling the characteristics of the time-dependent functions of the multiple task points changing with time according to the time windows of the multiple task points respectively, to obtain the time characteristic functions of the multiple task points respectively; Determining the time weights of the multiple task points respectively according to the time windows of the multiple task points respectively and the time characteristic functions of the multiple task points respectively; Modeling the attraction relationship between this task point and other task points according to the spatial positions of the multiple task points, to determine the distribution intensity function of this task point. The attraction relationship between task points is used to characterize the degree of being on the way between task points, including: representing the spatial distribution position of task points as a Poisson field with boundary conditions, describing the attraction relationship between task points through the potential field of task points, determining the density function of task points based on the potential field of task points to reflect the distribution intensity of task points; encoding the spatial constraints and the attraction relationship between task points into the density function through partial differential equations, and taking the density function of task points as the distribution intensity function of task points; For each task point among the multiple task points, obtaining the spatial weight of this task point according to the distribution intensity function of this task point; Obtaining the spatio-temporal weight of this task point according to the time weights and spatial weights of the multiple task points respectively; Planning the execution order of the multiple task points with the goal of the highest task completion efficiency according to the spatio-temporal weights of the multiple task points respectively; Modeling the characteristics of the time-dependent functions of the multiple task points changing with time according to the time windows of the multiple task points respectively, to obtain the time characteristic functions of the multiple task points respectively, including: determining the time-dependent functions of the multiple task points respectively according to the time windows of the multiple task points respectively. The time-dependent function of a task point is used to characterize whether it is within the time window of this task at any moment; using fractional differential equations and the time-dependent functions of the multiple task points respectively, modeling the characteristics of the time-dependent functions of the multiple task points changing with time, to obtain the time characteristic functions of the multiple task points respectively; For each task point among the multiple task points, obtaining the spatial weight of this task point according to the distribution intensity function of this task point, including: Performing multi-scale feature decomposition on the distribution intensity function of each task point among the multiple task points, to obtain the weights corresponding to each scale and the feature basis functions of this task point at each scale, including: decomposing the distribution intensity function of the task point to obtain the feature intensity and feature basis functions of the task point at different scales, where the feature intensity represents the weight corresponding to each scale, and each scale corresponds to different spatial resolutions; For each task point among the multiple task points, obtaining the spatial weight of this task point according to the feature basis functions of this task point at each scale and the distances between this task point and each task point at this scale, including: summing up the multi-scale features of each task point and the distances between the task point and each task point at the corresponding scale, to obtain the spatial weight of this task point.

2. The multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism according to claim 1, characterized in that, The method further includes: For any one of the multiple task points, after detecting that the spatial position of the task point is updated, determine the updated spatial weights and updated spatio-temporal weights of each of the multiple task points according to the updated spatial position of the task point; According to the updated spatio-temporal weights of each of the multiple task points, with the goal of maximizing the task completion efficiency, re-plan the execution order of the multiple task points.

3. The multi-objective intelligent planning method based on the spatio-temporal cognitive mechanism according to claim 1, wherein The method further includes: After detecting a new task point, determine the spatio-temporal weight of the new task point, and according to the spatial position of the new task point, determine the updated spatial weights and updated spatio-temporal weights of each of the multiple task points; According to the updated spatio-temporal weights of each of the multiple task points and the spatio-temporal weight of the new task point, with the goal of maximizing the task completion efficiency, re-plan the execution order of the multiple task points.

4. The multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism according to claim 1, wherein The method further includes: After detecting that any one of the multiple task points is deleted, determine the updated spatial weights and updated spatio-temporal weights of the remaining task points; According to the updated spatio-temporal weights of the remaining task points, with the goal of maximizing the task completion efficiency, re-plan the execution order of the remaining task points.

5. The multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism according to claim 1, wherein The method further includes: For any one of the multiple task points, after detecting that the time window of the task point is updated, determine the updated time weight and updated spatio-temporal weight of the task point according to the updated time window of the task point; According to the updated spatio-temporal weight of the task point and the spatio-temporal weights of other task points, with the goal of maximizing the task completion efficiency, re-plan the execution order of the multiple task points.

6. The multi-objective intelligent planning method based on a spatio-temporal cognitive mechanism according to any one of claims 1-5, characterized in that After planning the execution order of the multiple task points, the method further includes: According to the execution order of the multiple task points, determine every two adjacent task points in the execution order; According to multiple paths between every two adjacent task points in the execution order, plan the optimal path for traversing the multiple task points.

7. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the multi-objective intelligent planning method based on spatio-temporal cognitive mechanism according to any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the multi-objective intelligent planning method based on spatio-temporal cognitive mechanism according to any one of claims 1 to 6.

Images