Map blocking method and device, electronic equipment and storage medium

Abstract

The invention discloses a map partitioning method and device, electronic equipment and a readable storage medium. The method comprises the steps of obtaining respective partitioning scene features of a plurality of map partitions; performing simulation processing on the basis of respective block scene features of the map blocks through target computing nodes allocated for the map blocks, and performing statistics on the resource utilization rate of the target computing nodes of the map blocks in the simulation processing process; determining the load rate of the target computing node of each map block according to the resource utilization rate of the target computing node of each map block; and according to the load rate of each target computing node, performing block adjustment on the plurality of map blocks to obtain a target map block corresponding to each target computing node. According to the method provided by the invention, the situation of slow simulation process caused by insufficient computing resources of the target computing node or resource waste caused by overlarge computing resources of the target computing node is reduced, and the utilization rate of the computing resources is improved.

Description

Technical Field

[0001] The present application relates to the field of electronic information technology, and more specifically, to a map segmentation method, device, electronic device and computer-readable storage medium. Background Art

[0002] Autonomous driving simulation is the application of computer simulation technology in the field of vehicle autonomous driving. It digitally restores the application scenarios of autonomous driving through mathematical modeling and establishes an autonomous driving model close to the real world. At present, when simulating autonomous driving of vehicles, there is a waste of computing resources or a slow simulation process. Summary of the invention

[0003] The present application proposes a map segmentation method, device, electronic device and computer-readable storage medium to improve the above-mentioned defects.

[0004] In a first aspect, an embodiment of the present application provides a map segmentation method, the method comprising:

[0005] Obtaining the scene features of each of the multiple map blocks; the multiple map blocks are obtained by evenly dividing the target map;

[0006] By allocating target computing nodes to each map block, simulation processing is performed based on the block scene characteristics of each map block, and during the simulation processing, the resource utilization rate of the target computing nodes of each map block is counted; the computing resource difference of each target computing node is less than the preset difference range;

[0007] Determine the load rate of the target computing node of each map block according to the resource utilization rate of the target computing node of each map block;

[0008] According to the load rate of each target computing node, multiple map blocks are adjusted to obtain the target map block corresponding to each target computing node.

[0009] In a second aspect, the embodiment of the present application further provides a map segmentation device, the device comprising:

[0010] An acquisition module is used to obtain the scene features of each of the multiple map blocks; the multiple map blocks are obtained by evenly dividing the target map;

[0011] The simulation module is used to perform simulation processing based on the block scene characteristics of each map block through the target computing nodes allocated to each map block, and in the process of simulation processing, count the resource utilization rate of the target computing nodes of each map block; the computing resource difference of each target computing node is less than the preset difference range;

[0012] A load rate determination module, used to determine the load rate of the target computing node of each map block according to the resource utilization rate of the target computing node of each map block;

[0013] The adjustment module is used to adjust the multiple map blocks according to the load rate of each target computing node to obtain the target map block corresponding to each target computing node.

[0014] In a third aspect, an embodiment of the present application further provides an electronic device, comprising: one or more processors; a memory; one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to execute the above method.

[0015] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores a program code executable by a processor, and when the program code is executed by the processor, the processor executes the above method.

[0016] In a fifth aspect, an embodiment of the present application further provides a computer program product, comprising computer-readable instructions, which perform the above method when executed by a processor.

[0017] The present application provides a map segmentation method, device, electronic device and computer-readable storage medium. In the present application, multiple map segments are obtained by evenly dividing a target map. Therefore, the computing resource difference of the target computing nodes allocated to each map segment is small (that is, the computing resource difference is smaller than a preset difference range). First, the target computing nodes allocated to each map segment are simulated based on the segmentation scene characteristics of each map segment, and the resource utilization rate of the target computing nodes of each map segment is counted. Then, according to the resource utilization rate of the target computing nodes of each map segment, the load rate of the target computing nodes of each map segment is determined; according to the load rate of each target computing node, multiple map segments are segmented and adjusted, thereby achieving the goal of indicating the target segment. The load rate of the actual resource utilization of the target computing node is used to determine the target map block of each target computing node, so that the computing resources required for the target map block allocated to each target computing node are consistent with the computing resources of the target computing node itself, thereby achieving the purpose of allocating map blocks to the target computing node on demand, and effectively avoiding the situation where the simulation process is slow due to insufficient computing resources of the target computing node or the resource waste caused by excessive computing resources of the target computing node when the target computing node re-simulates the map block in the case of evenly allocating map blocks to each target computing node with small computing resource differences, thereby improving the computing resource utilization rate, greatly reducing the waste of computing resources and the slow simulation process, and realizing the rational use of resources.

[0018] Other features and advantages of the embodiments of the present application will be described in the subsequent description, and partly become apparent from the description, or be understood by practicing the embodiments of the present application. The purposes and other advantages of the embodiments of the present application can be realized and obtained by the structures specifically pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 A flowchart of a map segmentation method proposed according to an embodiment of the present application is shown.

[0021] Figure 2 A schematic diagram of a map segmentation in an embodiment of the present application is shown.

[0022] Figure 3 A schematic diagram of another map segmentation in an embodiment of the present application is shown.

[0023] Figure 4 Shows Figure 1 The steps after step S150 of the corresponding embodiment are a flowchart in an embodiment.

[0024] Figure 5 A structural block diagram of a map segmentation device proposed in one embodiment of the present application is shown.

[0025] Figure 6 A structural block diagram of an electronic device provided according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0026] In order to make those skilled in the art better understand the present application scheme, the technical scheme in the present application embodiment will be clearly and completely described below in conjunction with the drawings in the present application embodiment. Obviously, the described embodiment is only a part of the present application embodiment, rather than all the embodiments. The components of the present application embodiment usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the present application for protection, but merely represents the selected embodiment of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0027] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0028] See also Figure 1 , Figure 1 A flowchart of a map segmentation method proposed according to an embodiment of the present application is shown, which is used in an electronic device. The method includes:

[0029] S110: Obtaining block scene features of each of the plurality of map blocks.

[0030] Among them, multiple map blocks are obtained by evenly dividing the target map.

[0031] The electronic device in this embodiment may be a physical server, a cloud server, a mobile terminal, a portable computer or a desktop computer, etc.

[0032] The map blocks can be obtained by dividing the target map, and the target map can be an electronic map of a certain area. For example, the target map is an electronic map of city A, and another example is that the target map is an electronic map of district B in city A.

[0033] In the present application, the determined target map may be an electronic map of regular shape, or an electronic map of irregular shape, for example, the target map is a rectangular map, or another example, the target map is an irregular map surrounded by curves.

[0034] In the case where the shape of the target map is a regular shape, the target map can be divided into map blocks of uniform shape, thereby obtaining a plurality of map blocks. For example, if the target map is a square map, the target map can be evenly divided into four small squares. In other words, the average division is the average in area.

[0035] In the case where the shape of the target map is irregular, the target map may be divided into map blocks of inconsistent shapes, thereby obtaining a plurality of map blocks. For example, if the target map is an irregular polygonal map, the target map may be divided into circular map blocks, square map blocks, and irregular polygonal map blocks.

[0036] The scene features of a map refer to the relevant information used to indicate the simulation process for the map, which may include the map itself, other objects on the map, simulation algorithms, simulation duration, etc. In this application, the scene features of a map include not only the map itself, but also scene complexity, sensor simulation information, algorithm complexity, simulation duration, and physical model accuracy.

[0037] The scene complexity includes the following:

[0038] Number of vehicles: The more vehicles there are, the greater the amount of computation required, and more central processing units (CPUs) and memory are needed to handle interactions and behavior predictions between vehicles.

[0039] Road environment: Complex road structures (such as multiple intersections, viaducts, roundabouts, etc.) require more computing resources to simulate vehicle driving paths and decisions.

[0040] Traffic flow: High-traffic scenarios mean more vehicles interacting simultaneously, increasing computing requirements.

[0041] Sensor simulation information includes the following:

[0042] Types and quantity of sensors: such as cameras, lidar, millimeter-wave radar, etc. The data processing of each sensor requires certain computing resources.

[0043] Sensor resolution and frame rate: Sensors with higher resolution and frame rate generate more data and require more memory and computing power to process.

[0044] Algorithm complexity can include the following:

[0045] Object detection and tracking algorithms: Complex algorithms require more computing resources to identify and track objects such as vehicles and pedestrians in real time.

[0046] Path planning and decision-making algorithms: Optimized algorithms may require more computation to generate accurate and safe driving paths.

[0047] The simulation duration refers to the duration of the vehicle's driving process when simulating the vehicle's driving process based on the scene characteristics. With respect to the simulation duration, a longer simulation duration means more calculations, and thus more resources are required to maintain stable performance.

[0048] The accuracy of the physical model includes the following:

[0049] Accuracy of vehicle dynamics model: A high-precision model requires more calculations to accurately simulate the vehicle's acceleration, braking, steering and other behaviors.

[0050] Environmental physics model: For example, the impact of weather conditions (rain, snow, wind, etc.) on vehicle driving. Complex physics models require more resources.

[0051] After the target map is divided, a plurality of map blocks are obtained. For each map block, scene features located within the range of the map block are obtained from the scene features of the target map as block scene features of the map block.

[0052] S120, performing simulation processing based on the block scene characteristics of each map block through the target computing nodes allocated to each map block, and during the simulation processing, counting the resource utilization rate of the target computing nodes of each map block.

[0053] Among them, the computing resource difference of each target computing node is less than the preset difference range; the preset difference range can be set based on demand, which is not repeated here. The computing resources involved in this application may include central processing unit (CPU) resources, memory resources, and network bandwidth resources. Among them, the central processing unit resources can be the number of CPUs, the memory resources can be the size of the memory, and the network bandwidth resources can be the specific bandwidth of the network.

[0054] For example, when the computing resources are central processing unit resources, the preset difference range is that the CPU quantity difference is 0, that is, the CPU quantity of the target computing node is the same. For another example, when the computing resources are memory resources, the preset difference range is that the memory difference is less than 100MB.

[0055] Since each map block is obtained by evenly dividing the target map, the computing resources of the target computing nodes allocated to each map block have small differences (the computing resource differences are smaller than a preset difference range) or the computing resources are consistent.

[0056] For example, Figure 2 As shown, the square target map c is evenly divided to obtain four square map blocks: map block p1, map block p2, map block p3 and map block p3. The positions of the four map blocks are shown in Table 1. Table 1 is as follows:

[0057] Table 1

[0058] Map Tiles p1 p2 p3 p4 X coordinate of the upper left corner of the block 0 0 10 10 Y coordinate of the upper left corner of the block 0 10 0 10 Block width 10 10 10 10 Block height 10 10 10 10

[0059] The block height and block width are used to indicate the size of the block, and the xy coordinates of the upper left corner of the block are used to indicate the position of the map block in the target map.

[0060] For the four map blocks in Table 1, four target computing nodes are determined. When the computing resources involved include the resources of the central processing unit and the resources of the memory, the target computing nodes allocated to the four map blocks in Table 1 are shown in Table 2. Table 2 is as follows:

[0061] Table 2

[0062] Target compute node n1 n2 n3 n4 Map Tiles p1 p2 p3 p4 Number of CPUs 1C 1C 1C 1C Memory 4G 4G 4G 4G

[0063] Among them, the number of CPUs refers to the number of cores of the required CPU, and 1C refers to 1 core of the required CPU. In Table 2, the computing resources of the target computing nodes n1, n2, n3, and n4 are all 1 core CPU and 4G memory.

[0064] For each map block, simulation processing is performed based on the block scene characteristics of the map block through the target computing nodes of each map block, and then during the simulation processing, the resource utilization rate of the target computing nodes of the map block is counted.

[0065] Among them, the resource utilization of the target computing node refers to the resource utilization of the resource objects in the target computing node used for the simulation process. Exemplarily, the resource objects may include the aforementioned central processing unit, memory and network bandwidth. At this time, the resource utilization of the target computing node may include the utilization of the central processing unit, memory and network bandwidth of the target computing node.

[0066] In some implementations, any time point in the simulation process may be selected to collect the instantaneous resource utilization of the target computing node of the map block at that time point as the resource utilization of the target computing node of the map block.

[0067] In some other embodiments, during the simulation process, the resource utilization rate of the target computing node of the map block at each simulation moment in the simulation period corresponding to the simulation process (the resource utilization rate at the moment refers to the instantaneous resource utilization rate at the simulation moment) can be counted as the resource utilization rate of the target computing node of the map block. The simulation period here is the specific period of the simulation process with the aforementioned duration as the simulation period. For example, if the simulation duration is 1h and the simulation process starts at 11 o'clock, the simulation period is from 11 o'clock to 12 o'clock.

[0068] Generally speaking, a simulation process is performed at each simulation moment, and simulation processes at multiple simulation moments are performed to realize simulation processes for the entire simulation period, wherein the simulation moment can be set based on demand, for example, simulation processes are performed every 1 second.

[0069] As mentioned above, the simulation process is implemented through the target computing node. Therefore, the resource utilization rate of the target computing node of the map block at each simulation moment may include the utilization rate of each resource object in the target computing node at the simulation moment.

[0070] Exemplarily, the resource objects include the aforementioned central processing unit, memory and network bandwidth. For the four simulation times t1, simulation time point t2, simulation time point t3 and simulation time point t4 within the simulation period, the resource utilization rates collected for the target computing node of a certain map block are shown in Table 3. Table 3 is as follows:

[0071] Table 3

[0072]

[0073]

[0074] S130 , determining the load rate of the target computing node of each map block according to the resource utilization rate of the target computing node of each map block.

[0075] In this embodiment, for each target computing node, the load rate of the target computing node may be determined according to the resource utilization rate of the target computing node.

[0076] As an implementation method, when the resource utilization of the target computing node of the map block is the instantaneous resource utilization at a certain point in time during the simulation process, the utilization of each resource object of the target computing node can be weighted and summed to obtain the load rate of the target computing node of the map block; wherein, the weight of the utilization of each resource object can be set based on demand, and this application does not limit it.

[0077] In another embodiment, the resource utilization of the target computing node of the map block includes the momentary resource utilization of the target computing node of the map block at each simulation moment within the simulation period corresponding to the simulation processing; the load rate of the target computing node of the map block can be determined based on the momentary resource utilization of the target computing node of the map block at each simulation moment involved in the entire simulation period, so as to indicate the resource usage of the target computing node in the entire simulation period through the load rate.

[0078] Among them, when the resource objects include central processing unit, memory and network bandwidth, for each map block, the load rate of the target computing node of the map block can be determined according to the utilization rate of the central processing unit, memory and network bandwidth of the target computing node of the map block at each simulation time.

[0079] In some implementations, when the resource utilization of the target computing node of the map block includes the resource utilization of the target computing node of the map block at each simulation moment within the simulation period corresponding to the simulation process, S140 may include A1-A3, where:

[0080] A1: For each simulation moment, the moment load score of the target computing node of each map block at the simulation moment is determined according to the moment resource utilization rate of the target computing node of each map block at the simulation moment.

[0081] In the present application, the target computing node is overloaded or underloaded. Therefore, the overload score and / or underload score of the i-th map block at the j-th simulation moment can be determined based on the momentary resource utilization of the i-th map block at the j-th simulation moment, as the momentary load score of the i-th map block at the j-th simulation moment.

[0082] In some embodiments, the resource utilization at a given moment includes the utilization of the central processing unit, the utilization of the memory, and the utilization of the network bandwidth of the target computing node; accordingly, A1 may include: determining the single load score of the target resource object of the target computing node of the map block at the simulation time according to the utilization of the target resource object of the target computing node of the map block at the simulation time and the utilization range corresponding to the target resource object; the target resource object is any one of the central processing unit, the memory, and the network bandwidth of the target computing node; summing the single load score of the central processing unit of the target computing node of the map block at the simulation time, the single load score of the memory at the simulation time, and the single load score of the network bandwidth at the simulation time to obtain the moment load score of the target computing node of the map block at the simulation time.

[0083] Among them, the utilization range corresponding to the target resource object can be set based on demand, and this application does not limit it. Since the target computing node has overload and underload situations, the utilization range corresponding to the target resource object of the target computing node can include a low-load utilization range in the underload situation and an overload utilization range in the overload situation.

[0084] For different target resource objects, the set overload utilization range and underload utilization range may be different. For example, the overload utilization range of memory and network bandwidth is different, and the underload utilization range of CPU and memory is different.

[0085] First, the utilization rate of the target resource object of the target computing node for map partitioning at the simulation moment and the overload utilization rate range corresponding to the target resource object can be calculated to determine the overload single-item load score of the target resource object of the target computing node for map partitioning at the simulation moment. Then, the overload single-item load scores of the central processing unit (CPU), memory, and network bandwidth of the target computing node for map partitioning at the simulation moment are summed up to obtain the overload score of the target computing node for map partitioning at the simulation moment. Similarly, the low-load single-item load score of the target resource object of the target computing node for map partitioning at the simulation moment can be determined based on the utilization rate of the target resource object of the target computing node for map partitioning at the simulation moment and the low-load utilization rate range corresponding to the target resource object. Then, the low-load single-item load scores of the CPU, memory, and network bandwidth of the target computing node for map partitioning at the simulation moment are summed up to obtain the low-load score of the target computing node for map partitioning at the simulation moment. Finally, the low-load score and overload score of the target computing node for map partitioning at the simulation moment are used as the moment load score of the target computing node for map partitioning at the simulation moment.

[0086] Of course, the low-load score of the target computing node for map partitioning at the simulation moment can also be used as the moment load score of the target computing node for map partitioning at the simulation moment. Or, the overload score of the target computing node for map partitioning at the simulation moment can be used as the moment load score of the target computing node for map partitioning at the simulation moment.

[0087] Specifically, the moment overload score SG of the target computing node for the i-th map partitioning at the j-th simulation moment can be determined in the following way ij :

[0088] sg_cpu = w_cpu * (u_cpu / h_cpu). If u_cpu < h_cpu, then sg_cpu = 0; otherwise, sg_cpu remains as it is. Here, sg_cpu is the cpu score of the target computing node for the i-th map partitioning at the j-th simulation moment (i.e., the overload single-item load score of the cpu), w_cpu is the weight of the cpu, which can be 0.6, h_cpu is the overload threshold corresponding to the cpu (equivalent to the overload utilization rate range of the cpu), and u_cpu is the utilization rate of the cpu of the target computing node for the i-th map partitioning at the j-th simulation moment. Here, i and j are non-zero natural numbers.

[0089] sg_mem = w_mem * (u_mem / h_mem). If u_mem < h_mem, then sg_mem = 0; otherwise, it is sg_mem. Here, sg_mem is the memory score of the target computing node of the i-th map block at the j-th simulation moment (i.e., the single overload load score of the memory), w_mem is the weight of the memory, which can be 0.2, h_mem is the overload threshold corresponding to the memory (equivalent to the overload utilization range corresponding to the memory), and u_mem is the utilization rate of the memory of the target computing node of the i-th map block at the j-th simulation moment.

[0090] sg_net = w_net * (u_net / h_net). If u_net < h_net, then sg_net = 0; otherwise, it is sg_net. Here, sg_net is the network bandwidth score of the network bandwidth of the target computing node of the i-th map block at the j-th simulation moment (i.e., the single overload load score of the network bandwidth), w_net is the weight of the network bandwidth, which can be 0.2, h_net is the overload threshold corresponding to the network bandwidth (equivalent to the overload utilization range corresponding to the network bandwidth), and u_net is the utilization rate of the network bandwidth of the target computing node of the i-th map block at the j-th simulation moment.

[0091] Finally, calculate the overload score SG of the target computing node of the i-th map block at the j-th simulation moment. ij―g :

[0092] SG ij = sg_cpu + sg_mem + sg_net.

[0093] Similarly, the underload score SD of the target computing node of the i-th map block at the j-th simulation moment can be determined in the following way. ij :

[0094] sd_cpu = w_cpu * (u_cpu / l_cpu). If u_cpu < l_cpu, then sd_cpu = 1; otherwise, it is sd_cpu. Here, sd_cpu is the cpu score of the cpu of the target computing node of the i-th map block at the j-th simulation moment (i.e., the single underload load score of the cpu), l_cpu is the underload threshold corresponding to the cpu (equivalent to the underload utilization range corresponding to the cpu), and u_cpu is the utilization rate of the cpu of the target computing node of the i-th map block at the j-th simulation moment.

[0095] sd_mem = w_mem * (u_mem / l_mem). If u_mem < l_mem, then sd_mem = 1; otherwise, it is sd_mem. Here, sd_mem is the memory score of the memory of the target computing node of the i-th map block at the j-th simulation moment (i.e., the low-load single-item load score of the memory), l_mem is the low-load threshold corresponding to the memory (equivalent to the low-load utilization range corresponding to the memory), and u_mem is the utilization rate of the memory of the target computing node of the i-th map block at the j-th simulation moment.

[0096] sd_net = w_net * (u_net / l_net). If u_net < l_net, then sd_net = 1; otherwise, it is sd_net. Here, sd_net is the network bandwidth score of the network bandwidth of the target computing node of the i-th map block at the j-th simulation moment (i.e., the low-load single-item load score of the network bandwidth), l_net is the low-load threshold corresponding to the network bandwidth (equivalent to the low-load utilization range corresponding to the network bandwidth), and u_net is the utilization rate of the network bandwidth of the target computing node of the i-th map block at the j-th simulation moment.

[0097] Finally, calculate the low-load score SD of the target computing node of the i-th map block at the j-th simulation moment. ij―g :

[0098] SD ij = sd_cpu + sd_mem + sd_net.

[0099] A2: Determine the load scores of the target computing nodes of each map block during the simulation period according to the moment load scores of the target computing nodes of each map block at each simulation moment.

[0100] As described above, the moment load scores of the target computing nodes of the map blocks at each simulation moment can include the overload scores and / or low-load scores of the target computing nodes of the map blocks at each simulation moment. Therefore, in the case where the moment load scores of the target computing nodes of the map blocks at each simulation moment include the overload scores of the target computing nodes of the map blocks at each simulation moment, the overload scores of the target computing nodes of each map block during the simulation period can be determined by combining the overload scores of the target computing nodes of the map blocks at each simulation moment. Similarly, in the case where the moment load scores of the target computing nodes of the map blocks at each simulation moment include the low-load scores of the target computing nodes of the map blocks at each simulation moment, the low-load scores of the target computing nodes of each map block during the simulation period can be determined by combining the low-load scores of the target computing nodes of the map blocks at each simulation moment.

[0101] In some embodiments, A2 may include: summing the load scores of the target computing nodes of the map blocks at the target simulation time to obtain the load scores of the target computing nodes of the map blocks during the simulation period; wherein the target simulation time refers to the simulation time during the simulation period when the load scores of the target computing nodes of the map blocks are higher than the load score threshold. The load score threshold can be set based on demand, for example, 1.

[0102] That is to say, for the target computing node of the ith map block at the jth simulation time, if the load score of the target computing node of the ith map block at the jth simulation time is higher than the load score threshold, the jth simulation time is determined as the target simulation time.

[0103] As mentioned above, the target computing node involves underload and overload, and accordingly, the moment load score involves overload score and underload score. The same moment load score threshold can be set for the overload score and the underload score. At this time, for the overload situation, for any simulation moment, if the overload score of the target computing node of the map block at the simulation moment is higher than the moment load score threshold, the simulation moment is determined as a target simulation moment, and then the overload scores of all target simulation moments in the simulation period are summed up as the overload score of the target computing node of the map block in the simulation period; for the underload situation, for any simulation moment, if the underload score of the target computing node of the map block at the simulation moment is higher than the moment load score threshold, the simulation moment is determined as a target simulation moment, and then the underload scores of all target simulation moments in the simulation period are summed up as the underload score of the target computing node of the map block in the simulation period. Finally, the overload score and underload score of the target computing node of the map block in the simulation period are obtained as the load score of the target computing node of the map block in the simulation period.

[0104] Of course, in some implementations, the overload score of the target computing node of the map block during the simulation period can also be obtained as the load score of the target computing node of the map block during the simulation period. Similarly, the underload score of the target computing node of the map block during the simulation period can be obtained as the load score of the target computing node of the map block during the simulation period.

[0105] A3: Determine the load rate of the target computing node of each map block according to the load score of the target computing node of each map block during the simulation period.

[0106] The load rate of the target computing node of the map block can be determined according to the load score of the target computing node of the map block in the simulation period and the duration of the simulation period. For example, the ratio of the load score of the target computing node of the map block in the simulation period and the duration of the simulation period can be calculated as the load rate of the target computing node of the map block.

[0107] As mentioned above, the target computing node involves overload and underload situations. In the overload situation, the ratio of the overload score of the target computing node of the map block in the simulation period and the duration of the simulation period can be calculated as the overload load rate of the target computing node of the map block; similarly, in the underload situation, the ratio of the underload score of the target computing node of the map block in the simulation period and the duration of the simulation period can be calculated as the underload load rate of the target computing node of the map block.

[0108] Afterwards, the underload rate and / or overload rate of the target computing node of the map block may be obtained as the load rate of the target computing node of the map block.

[0109] S140. According to the load rate of each target computing node, multiple map blocks are adjusted to obtain a target map block corresponding to each target computing node.

[0110] As mentioned above, if the target computing node is overloaded or underloaded, and the load rate of the target computing node includes an overload rate, if the overload rate of the target computing node is higher than the overload rate threshold, the map block processed by the target computing node is split, and the split map block is obtained as the target map block corresponding to the target computing node. The overload rate threshold can be set based on demand, for example, the overload rate threshold is 0.85.

[0111] When the overload rate of the target computing node is higher than the overload rate threshold, it means that the target computing node is seriously overloaded, and the computing resources of the target computing node are difficult to stably support the simulation processing of the map blocks. At this time, the map blocks that the target computing node is responsible for processing need to be split to obtain multiple map blocks after the split, and any one of the split map blocks can be selected as the target map block corresponding to the target computing node. At the same time, for the other map blocks except the target map block in the multiple split map blocks, an auxiliary computing node can be configured for each of the other map blocks, wherein the computing resource difference between the auxiliary computing node and the target computing node is less than the preset difference range, thereby reducing the amount of data processed by the target computing node and the auxiliary computing node, thereby reducing the overload of the target computing node, so that the computing resources of the target computing node can stably support the simulation processing of the map blocks.

[0112] In some embodiments, when the overload rate of the target computing node is higher than the overload rate threshold, the map block processed by the target computing node can be directly split into two blocks evenly, and one of the split map blocks can be used as the target map block of the target computing node.

[0113] When the overload rate of the target computing node is not higher than the overload rate threshold, it means that the overload situation of the target computing node is not serious, and the computing resources of the target computing node can stably support the simulation processing of the map block. At this time, the map block processed by the target computing node can be directly used as the target map block corresponding to the target computing node. That is, when the overload rate of the target computing node is not higher than the overload rate threshold, the map block processed by the target computing node is not adjusted.

[0114] When the load rate of the target computing node includes a low load rate, a target computing node to be merged whose low load rate is lower than a low load rate threshold can be determined; the map blocks processed by at least two target computing nodes to be merged are merged to obtain the merged map block as the target map block corresponding to any one of the at least two target computing nodes to be merged. The low load rate threshold can be set based on demand, for example, the low load rate threshold is 0.4.

[0115] When the low-load rate of the target computing node of the map block is lower than the low-load rate threshold, it means that the target computing node is seriously under-loaded, the computing resources of the target computing node are relatively idle, and there is a serious waste of resources. At this time, it is necessary to increase the amount of data processed by the target computing node, that is, to increase the total area of ​​the map blocks processed by the target computing node.

[0116] In some embodiments, each time a map block is merged, the map blocks processed by multiple target computing nodes to be merged whose total load rate does not exceed a specified threshold value may be merged into one map block, and the merged map block is used as the target map block corresponding to any one of the target computing nodes to be merged. In this way, the amount of data processed by the target computing node is increased, thereby reducing the underload situation of the target computing node and reducing the waste of computing resources of the target computing node. The specified threshold value may be consistent with the aforementioned overload rate threshold value.

[0117] Correspondingly, for other target computing nodes to be merged that have not been assigned target map blocks, the resources of other target computing nodes to be merged can be released so that the other target computing nodes to be merged can be used for other data processing matters. For example, other target computing nodes to be merged can be used as auxiliary computing nodes for processing a certain split map block.

[0118] When the low load rate of the target computing node of the map block is not lower than the low load rate threshold, it means that the low load situation of the target computing node is not serious, and there is no waste of computing resources of the target computing node. At this time, the map block processed by the target computing node can be directly used as the target map block corresponding to the target computing node. That is, when the low load rate of the target computing node is not lower than the low load rate threshold, the map block processed by the target computing node is not adjusted.

[0119] For example, Figure 2 After assigning respective target computing nodes to the map blocks according to the method shown in Table 2, the map blocks are simulated by the target computing nodes, and the load rates obtained are shown in Table 4. Table 4 is as follows:

[0120] Table 4

[0121] Target compute node n1 n2 n3 n4 Map Tiles p1 p2 p3 p4 Overload rate 0.8 0.95 0 0 Low load factor 0 0 0.34 0.25

[0122] In this example, the overload rate threshold involved is 0.85, and the underload rate threshold involved is 0.4.

[0123] Based on Table 4, it can be seen that the overload rate of the target computing node n1 does not exceed the overload rate threshold, and the computing resources of the map block p1 remain unchanged;

[0124] The overload rate of the target computing node n2 exceeds the overload rate threshold, and the map block p2 can be evenly divided into two map blocks p21 and p22, and the map block p21 is used as the target map block corresponding to the target computing node n2, and the computing node n5 with the same computing resources as the target computing node n2 is obtained, and the map block p22 is used as the target map block corresponding to the computing node n5;

[0125] The low load rates of the target computing node n3 and the target computing node n4 are both lower than the low load rate threshold, indicating that too much computing power is allocated to the map blocks p3 and p4. The target computing node n3 and the target computing node n4 are both used as target computing nodes to be merged, so that the map blocks p3 and p4 are merged into map block p5, and the map block p5 is used as the target map block of the target computing node n3. At this time, the resources of the target computing node n4 can be released.

[0126] The map block p2 can be evenly divided into two map blocks p21 and p22 in such a way that the height is halved, so that in the map blocks p1, p2, p3, and p4, Figure 2 In the case shown in the figure, after adjusting the computing resources and redividing the map blocks, the structure of the map blocks is as follows Figure 3As shown, the corresponding reallocated computing resources are shown in Table 5. Table 5 is as follows:

[0127] Table 5

[0128] Map Tiles p1 p21 p5 p22 X coordinate of the upper left corner of the block 0 10 0 10 Y coordinate of the upper left corner of the block 0 0 10 5 Block width 10 10 20 10 Block height 10 5 10 5 Number of CPUs 1C 1C 1C 1C Memory 4G 4G 4G 4G Target compute node n1 n2 n3 n5

[0129] In this embodiment, the plurality of map blocks are obtained by evenly dividing the target map. Therefore, the computing resource difference of the target computing node allocated to each map block is small (that is, the computing resource difference is smaller than the preset difference range). First, the target computing node allocated to each map block is simulated based on the block scene characteristics of each map block, and the resource utilization rate of the target computing node of each map block is counted. Then, according to the resource utilization rate of the target computing node of each map block, the load rate of the target computing node of each map block is determined; according to the load rate of each target computing node, the plurality of map blocks are adjusted, thereby realizing the load according to the actual resource utilization of the indicated target computing node. The load rate is determined to determine the target map blocks of each target computing node, so that the computing resources required for the target map blocks allocated to each target computing node are consistent with the computing resources of the target computing node itself, thereby achieving the purpose of allocating map blocks to the target computing nodes on demand, and effectively avoiding the situation where the target computing nodes with small computing resource differences are evenly allocated map blocks. When the target computing nodes re-simulate the map blocks, the simulation process is slow due to insufficient computing resources of the target computing nodes or the computing resources of the target computing nodes are too large, which causes resource waste. This improves the utilization rate of computing resources, greatly reduces the waste of computing resources and the slow simulation process, and realizes the rational use of resources.

[0130] In addition, in the present application, the load rate of the target computing nodes of each map block is determined in combination with the momentary resource utilization of the target computing nodes of the map blocks at each simulation moment within the simulation period, so as to realize the resource utilization of the target computing nodes of the comprehensive map blocks in the entire simulation period and determine the load rate of the target computing nodes of the map blocks. This effectively avoids the situation where the load rate of the determined target computing nodes is low based on the resource utilization at a certain simulation moment, which is difficult to represent the resource utilization of the entire simulation period. This improves the accuracy of the load rate of the target computing node, and then improves the accuracy of the target computing resource utilization determined according to the load rate, and further improves the rationality of resource utilization.

[0131] In one embodiment, the plurality of map blocks are obtained by dividing the target map; the block scene features of the map blocks are scene features of the target map that are within the range of the map blocks; Figure 4 As shown, after S150, the method further includes:

[0132] S210: Determine a target division method for dividing the target map into target map blocks according to the positions of the target map blocks in the target map.

[0133] The block scene features of the map block are scene features of the target map that are within the map block range. The target map is divided into the target map blocks of the aforementioned target computing nodes as the target division method of the target map blocks.

[0134] Optionally, the target division method may be determined according to the position of each target map block in the target map. Figure 3 The positions of the map blocks p1, p21, p22 and p5 in the target map are shown, and the target division method is determined as follows: the target is divided into two equal parts horizontally, and the upper part is divided into two equal parts vertically, and then the right part after the division is divided into two equal parts horizontally.

[0135] In some embodiments, one or more dividing curves are used to divide the target map into target map blocks, and the position of each dividing curve in the target map can be recorded as the curve position, thereby indicating the target division method based on the curve position of the dividing curve.

[0136] S220: If the similarity between the scene features corresponding to the map to be processed and the scene features corresponding to the target map is higher than the similarity threshold, the map to be processed is divided according to the target division method to obtain a plurality of map blocks to be processed; one map block to be processed corresponds to one target map block.

[0137] Among them, one to-be-processed map block corresponds to one target map block. The method of obtaining the to-be-processed map and the scene features corresponding to the to-be-processed map refers to the method of obtaining the target map and the scene features of the target map, which will not be repeated here. The to-be-processed map can be a map different from the target map.

[0138] In the present application, the scene features of the map can be converted into a vector format through a neural network to obtain the scene feature vector of the map, and then the similarity between the scene feature vector corresponding to the map to be processed and the block scene feature vector corresponding to the target map is calculated as the similarity between the scene features corresponding to the map to be processed and the scene features corresponding to the target map.

[0139] Exemplarily, the cosine similarity, Euclidean distance or Euler distance between the scene feature vector corresponding to the map to be processed and the block scene feature vector corresponding to the target map can be calculated as the similarity between the scene features corresponding to the map to be processed and the scene features corresponding to the target map.

[0140] For example, the scene features of the map are shown in Table 7, which is as follows:

[0141] Table 7

[0142] Scene characteristics Traffic car pedestrian bike Cone barrel Tripod 1 1 3 4 5 8 2 10 5 7 3 3

[0143] The scene features shown in Table 7 are converted into a vector format to obtain preprocessed data as shown in Table 8, and then the preprocessed data are aggregated into vectors to obtain a scene feature vector. Table 8 is as follows:

[0144] Table 8

[0145] Scene characteristics Traffic car pedestrian bike Cone barrel Tripod 1 0.1 0.3 0.42 0.57 0.64 2 1 0.5 0.75 0.33 0.24

[0146] Exemplarily, when calculating the cosine similarity between the scene feature vector corresponding to the map to be processed and the block scene feature vector corresponding to the target map as the similarity between the scene features corresponding to the map to be processed and the scene features corresponding to the target map, the similarity between the scene features corresponding to the map to be processed and the scene features corresponding to the target map belongs to the interval [-1,1], and a similarity of -1 means that the scene features corresponding to the map to be processed and the scene features corresponding to the target map are completely different, and a similarity of 1 means that the scene features corresponding to the map to be processed and the scene features corresponding to the target map are exactly the same.

[0147] If the similarity between the scene features corresponding to the map to be processed and the scene features corresponding to the target map is higher than the similarity threshold, it is determined that the scene features of the target map are highly similar to the scene features of the map to be processed, and the allocation method of computing resources in the simulation process of the target map can be used to simulate the map to be processed. At this time, the map to be processed is divided according to the target division method of the target map to obtain multiple blocks of the map to be processed. Among them, the similarity threshold can be set based on demand, and this application does not limit it. For example, the similarity threshold is 0.8.

[0148] If the position of the to-be-processed map block in the to-be-processed map is consistent with the position of the target map block in the target map, it is determined that the to-be-processed map block and the target map block correspond to each other. Figure 3 The map blocks p1, p21, p22 and p5 shown in the figure, the obtained square map dt to be processed has a similarity with the target map that is higher than the similarity threshold. The map dt to be processed is divided in the same way to obtain four map blocks d1, d2, d3 and d3 to be processed, among which the position of d1 in dt is consistent with the position of p1 in the target map (both are a block in the upper left corner), and d1 is determined to be corresponding to p1. Similarly, the position of d2 in dt is determined to be consistent with the position of p21 in the target map (both are a block in the upper right corner), and d2 is determined to be corresponding to p21.

[0149] If the similarity between the scene features corresponding to the map to be processed and the scene features corresponding to the target map is not higher than the similarity threshold, it is determined that the scene features of the target map are not similar to the scene features of the map to be processed, then the map to be processed is simulated according to the simulation method of the target map. The simulation process is as described in S110-S140 above and will not be repeated here.

[0150] S230. Determine the scene features within the range of each map block to be processed from the scene features corresponding to the map to be processed, as the block scene features of each map block to be processed; for each map block to be processed, perform simulation processing based on the block scene features of the map block to be processed through the target computing node corresponding to the target map block corresponding to the map block to be processed.

[0151] That is, if the similarity between the scene features corresponding to the map to be processed and the scene features corresponding to the target map is higher than the similarity threshold, it is determined that the scene features of the target map are highly similar to the scene features of the map to be processed, and then the map to be processed is divided into multiple blocks to be processed by target division, and then the block scene features of each block to be processed are determined, and a block to be processed is assigned to each target computing node, and finally each block to be processed is simulated by the target computing node for each block to be processed.

[0152] For each target computing node, the map block to be processed corresponding to the target map block corresponding to the target computing node is used as the map block to be processed processed by the target computing node. Each target computing node performs simulation processing on the block scene features of the map block to be processed processed by each target computing node.

[0153] For example, after the target map is divided, the target map blocks are as follows Figure 3 As shown in the figure, the obtained square map to be processed dt has a similarity with the target map that is higher than the similarity threshold. Figure 3 The target division method shown divides the map to be processed dt into four map blocks to be processed d1, d2, d3 and d3, among which d1 corresponds to p1, d2 corresponds to p21, d3 corresponds to p21, and d4 corresponds to p5. Then, the target computing node corresponding to p1 is determined to be simulated based on the block scene characteristics of d1, the target computing node corresponding to p21 is determined to be simulated based on the block scene characteristics of d2, the target computing node corresponding to p22 is determined to be simulated based on the block scene characteristics of d3, and the target computing node corresponding to p5 is determined to be simulated based on the block scene characteristics of d4.

[0154] It can be understood that as the simulation proceeds, the database can store the map blocks after each map division and the target computing resources of the map blocks of each map. When each new map arrives, it can be determined whether there is a highly similar scene feature (the similarity is higher than the similarity threshold) according to the process of S210-S230. Then, for maps with high similarity, the target division method is reused. For dissimilar scene features, the target computing resources are determined according to the process of S110-S140.

[0155] In this embodiment, the target division method can be reused for scene features with highly similar scene features, avoiding the situation where the target computing node simulates the map blocks after the average division of the map to be processed, resulting in a slow simulation process due to insufficient computing resources of the target computing node or a waste of resources due to excessive computing resources of the target computing node, thereby improving the utilization rate of computing resources, greatly reducing the waste of computing resources and the slow simulation process, and realizing the rational use of resources.

[0156] At the same time, the target division method is reused for scene features with highly similar scene features, avoiding the need to redefine the division method for scene features of new maps, saving the time for redefining the division method, improving simulation efficiency and resource utilization, and thus achieving rational resource utilization.

[0157] See attached Figure 5 , Figure 5 The structure block diagram of a map segmentation device proposed in one embodiment of the present application is shown. For electronic equipment, the device 800 includes:

[0158] An acquisition module 810 is used to acquire scene features of each of a plurality of map blocks; the plurality of map blocks are obtained by evenly dividing the target map;

[0159] The simulation module 820 is used to perform simulation processing based on the block scene characteristics of each map block through the target computing nodes allocated to each map block, and in the process of simulation processing, count the resource utilization rate of the target computing nodes of each map block; the computing resource difference of each target computing node is less than the preset difference range;

[0160] A load rate determination module 830, for determining the load rate of the target computing node of each map block according to the resource utilization rate of the target computing node of each map block;

[0161] The adjustment module 840 is used to adjust the multiple map blocks according to the load rate of each target computing node to obtain the target map block corresponding to each target computing node.

[0162] Optionally, the load rate includes an overload rate; the adjustment module 840 is also used to split the map block processed by the target computing node if the overload rate of the target computing node is higher than the overload rate threshold, and obtain the split map block as the target map block corresponding to the target computing node.

[0163] Optionally, the load rate includes a low load rate; the adjustment module 840 is also used to determine a target computing node to be merged whose low load rate is lower than a low load rate threshold; merge the map blocks processed by at least two target computing nodes to be merged, and obtain the merged map block as the target map block corresponding to any one of the at least two target computing nodes to be merged.

[0164] Optionally, the block scene features of the map blocks are scene features in the target map that are located within the map block range; the device also includes a multiplexing module, which is used to determine a target division method for dividing the target map into target map blocks according to the positions of each target map block in the target map; if the similarity between the acquired scene features corresponding to the map to be processed and the scene features corresponding to the target map is higher than a similarity threshold, the map to be processed is divided according to the target division method to obtain a plurality of map blocks to be processed; one map block to be processed corresponds to one target map block; scene features located within the range of each map block to be processed are determined from the scene features corresponding to the map to be processed as the block scene features of each map block to be processed; for each map block to be processed, simulation processing is performed based on the block scene features of the map block to be processed through the target computing node corresponding to the target map block corresponding to the map block to be processed.

[0165] Optionally, the resource utilization of the target computing nodes of the map blocks includes the moment resource utilization of the target computing nodes of the map blocks at each simulation moment within the simulation period corresponding to the simulation processing; the load rate determination module 830 is also used to determine, for each simulation moment, the moment load score of the target computing nodes of each map block at the simulation moment according to the moment resource utilization of the target computing nodes of each map block at the simulation moment; determine the load score of the target computing nodes of each map block in the simulation period according to the moment load score of the target computing nodes of each map block at each simulation moment; determine the load rate of the target computing nodes of each map block according to the load score of the target computing nodes of each map block in the simulation period.

[0166] Optionally, the resource utilization at a given moment includes the utilization of the central processing unit, the utilization of the memory and the utilization of the network bandwidth of the target computing node; the load rate determination module 830 is further used to determine the single load score of the target resource object of the target computing node of the map block at the simulation time according to the utilization of the target resource object of the target computing node of the map block at the simulation time and the utilization range corresponding to the target resource object; the target resource object is any one of the central processing unit, the memory and the network bandwidth of the target computing node; the single load score of the central processing unit of the target computing node of the map block at the simulation time, the single load score of the memory at the simulation time and the single load score of the network bandwidth at the simulation time are summed to obtain the moment load score of the target computing node of the map block at the simulation time.

[0167] Optionally, the load rate determination module 830 is also used to sum the moment load scores of the target computing nodes of the map block at the target simulation time to obtain the load scores of the target computing nodes of the map block during the simulation period; wherein, the target simulation time refers to the simulation time during the simulation period when the moment load score corresponding to the target computing node of the map block is higher than the moment load score threshold.

[0168] Optionally, the load rate determination module 830 is further used to determine the load rate of the target computing node of the map block according to the load score of the target computing node of the map block in the simulation period and the duration of the simulation period.

[0169] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and modules can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here.

[0170] In addition, each function in each embodiment of the present application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above integrated module can be implemented in the form of hardware or in the form of software function module.

[0171] Please refer to Figure 6 , which shows a structural block diagram of an electronic device provided according to an embodiment of the present application. The electronic device 500 can be an electronic device capable of running applications, such as a smart phone, a tablet computer, an e-book, and a vehicle. The electronic device 500 in the present application may include one or more of the following components: a processor 510, a memory 520, and one or more applications. One or more applications may be stored in the memory 520 and configured to be executed by one or more processors 510, and one or more programs are configured to execute the method described in the aforementioned method embodiment.

[0172] The processor 510 may include one or more processing cores. The processor 510 uses various interfaces and lines to connect various parts of the entire electronic device 500, and executes various functions and processes data of the electronic device 500 by running or executing instructions, programs, code sets or instruction sets stored in the memory 520, and calling data stored in the memory 520. Optionally, the processor 510 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 510 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU), and a modem. Among them, the CPU mainly processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing display content; and the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 510, but may be implemented separately through a communication chip.

[0173] The memory 520 may include a random access memory (RAM) or a read-only memory (ROM). The memory 520 may be used to store instructions, programs, codes, code sets or instruction sets. The memory 520 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for implementing at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the following various method embodiments, etc. The data storage area may also store data (such as a phone book, audio and video data, chat record data) created by the electronic device 500 during use.

[0174] In addition, each function in each embodiment of the present application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above integrated module can be implemented in the form of hardware or in the form of software function module.

[0175] On the other hand, the present application also provides a computer-readable storage medium, in which program code is stored. The program code can be called by a processor to execute the method described in the above method embodiment.

[0176] The computer readable storage medium may be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read-only memory), an EPROM, a hard disk, or a cluster of ROMs. Optionally, the computer readable storage medium comprises a non-transitory computer-readable storage medium. The computer readable storage medium has storage space for program codes that execute any of the method steps in the above method. These program codes can be read from or written to one or more computer program products. The program code can be compressed, for example, in an appropriate form.

[0177] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A map segmentation method, characterized in that: The method comprises: Obtaining block scene features of each of a plurality of map blocks; the plurality of map blocks are obtained by evenly dividing the target map; By allocating target computing nodes to each of the map blocks, simulation processing is performed based on the block scene characteristics of each of the map blocks, and during the simulation processing, the resource utilization rate of the target computing nodes of each of the map blocks is counted; the computing resource difference of each of the target computing nodes is less than a preset difference range; Determining the load rate of the target computing node of each of the map blocks according to the resource utilization rate of the target computing node of each of the map blocks; According to the load rate of each of the target computing nodes, the multiple map blocks are adjusted to obtain the target map blocks corresponding to each of the target computing nodes.

2. The method according to claim 1, characterized in that The load rate includes an overload load rate; The step of adjusting the plurality of map blocks according to the load rate of each target computing node to obtain a target map block corresponding to each target computing node includes: If the overload rate of the target computing node is higher than the overload rate threshold, the map block processed by the target computing node is split to obtain the split map block as the target map block corresponding to the target computing node.

3. The method according to claim 1, characterized in that The load rate includes a low load rate; The step of adjusting the plurality of map blocks according to the load rate of each target computing node to obtain a target map block corresponding to each target computing node includes: Determine a target computing node to be merged whose low load rate is lower than a low load rate threshold; The map blocks processed by at least two target computing nodes to be merged are merged to obtain the merged map block as the target map block corresponding to any one of the at least two target computing nodes to be merged.

4. The method according to claim 1, characterized in that: The block scene features of the map block are scene features of the target map that are within the range of the map block; After the plurality of map blocks are adjusted according to the load rate of each target computing node to obtain the target map block corresponding to each target computing node, the method further includes: Determining a target division method for dividing the target map into the target map blocks according to positions of the target map blocks in the target map; If the similarity between the scene features corresponding to the map to be processed and the scene features corresponding to the target map is higher than the similarity threshold, the map to be processed is divided according to the target division method to obtain a plurality of map blocks to be processed; one map block to be processed corresponds to one target map block; Determine, from the scene features corresponding to the map to be processed, the scene features within the range of each of the blocks of the map to be processed as the block scene features of each of the blocks of the map to be processed; For each of the map blocks to be processed, simulation processing is performed based on the block scene characteristics of the map block to be processed through the target computing node corresponding to the target map block corresponding to the map block to be processed.

5. The method according to claim 1, characterized in that The resource utilization rate of the target computing node of the map block includes the resource utilization rate of the target computing node of the map block at each simulation time within the simulation period corresponding to the simulation process; Determining the load rate of the target computing node of each map block according to the resource utilization rate of the target computing node of each map block includes: For each simulation moment, determining a moment load score of the target computing node of each map block at the simulation moment according to the moment resource utilization rate of the target computing node of each map block at the simulation moment; Determining the load score of the target computing node of each map block in the simulation period according to the moment load score of the target computing node of each map block at each simulation moment; The load rate of the target computing node of each map block is determined according to the load score of the target computing node of each map block in the simulation period.

6. The method according to claim 5, characterized in that The resource utilization at the moment includes the utilization of the central processing unit, the utilization of the memory and the utilization of the network bandwidth of the target computing node; The step of determining, for each simulation moment, a load score of the target computing node of each map block at the simulation moment according to the resource utilization rate of the target computing node of each map block at the simulation moment, comprises: Determine the single load score of the target resource object of the target computing node of the map block at the simulation time according to the utilization rate of the target resource object of the target computing node of the map block at the simulation time and the utilization rate range corresponding to the target resource object; the target resource object is any one of the central processing unit, memory and network bandwidth of the target computing node; The single load score of the central processor of the target computing node of the map block at the simulation time, the single load score of the memory at the simulation time and the single load score of the network bandwidth at the simulation time are summed to obtain the moment load score of the target computing node of the map block at the simulation time.

7. The method according to claim 5, characterized in that Determining the load score of the target computing node of each map block in the simulation period according to the load score of the target computing node of each map block at each simulation time includes: The moment load scores of the target computing nodes of the map block at the target simulation time are summed to obtain the load scores of the target computing nodes of the map block in the simulation period; wherein the target simulation time refers to the simulation time within the simulation period when the moment load score corresponding to the target computing node of the map block is higher than the moment load score threshold.

8. The method according to claim 5, characterized in that Determining the load rate of the target computing node of each map block according to the load score of the target computing node of each map block in the simulation period includes: The load rate of the target computing node of the map block is determined according to the load score of the target computing node of the map block in the simulation period and the duration of the simulation period.

9. A map segmentation device, characterized in that: The device comprises: An acquisition module, used to acquire scene features of each of a plurality of map blocks; the plurality of map blocks are obtained by evenly dividing the target map; A simulation module, for performing simulation processing based on the block scene characteristics of each map block by assigning a target computing node to each map block, and during the simulation processing, counting the resource utilization rate of the target computing node of each map block; the computing resource difference of each target computing node is less than a preset difference range; A load rate determination module, used to determine the load rate of the target computing node of each map block according to the resource utilization rate of the target computing node of each map block; The adjustment module is used to adjust the multiple map blocks according to the load rate of each target computing node to obtain the target map block corresponding to each target computing node.

10. An electronic device, characterized in that: include: one or more processors; Memory; One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, and the one or more applications are configured to execute the method according to any one of claims 1-8.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program code executable by a processor, and when the program code is executed by the processor, the processor executes the method according to any one of claims 1 to 8.

Images