Parameter recommendation method and apparatus, and device

Abstract

The present application provides a parameter recommendation method and apparatus, and a device. The method comprises: identifying first performance state information by means of a parameter determination model corresponding to a vehicle control apparatus; and then obtaining M first parameter sets of the vehicle control apparatus and second performance state information corresponding to each first parameter set, wherein the second performance state information indicates the performance state of the vehicle control apparatus when operating on the basis of the first parameter sets, so as to improve the efficiency of parameter adjustment and effectively enhance the system performance after parameter adjustment. Furthermore, recommended information sent by a server comprises the second performance state information corresponding to each first parameter set among the M first parameter sets, in order to characterize the performance state that can be achieved by the vehicle control apparatus when using the first parameter set, thereby further ensuring that the ultimately used first parameter set enables the performance state achieved by the vehicle control apparatus to meet performance requirements.

Description

Parameter recommendation method, device and equipment Technical Field

[0001] The present application relates to the field of vehicle technology, and in particular to a parameter recommendation method, device, and apparatus. Background Art

[0002] The computing platform for intelligent driving services serves as the "brain" of the car, enabling operations such as perception, positioning, decision-making, planning, and control for intelligent driving. Therefore, the computing platform can be composed of multiple different types of processing units, such as units that perform image processing tasks, units that perform computing tasks, and units that perform artificial intelligence (AI) acceleration. These processing units can provide computing power for heterogeneous services and ensure the end-to-end operation of intelligent driving services. When these processing units provide computing power for heterogeneous services, they share memory bandwidth, which leads to the problem of memory bandwidth preemption. When multiple processing units compete for memory bandwidth, problems such as image data frame loss, excessive CPU usage, and excessive AI inference latency are likely to occur.

[0003] To alleviate the problem of memory bandwidth preemption, current methods often rely on manual observation of system operating status and manual adjustment of system parameters. However, this method is inefficient and cannot guarantee effective system performance improvement. Therefore, an intelligent solution for determining system parameters is urgently needed.

[0004] Summary of the Invention

[0005] The present application provides a parameter recommendation method, device and equipment, which provide an intelligent solution for determining system parameters to improve the adjustment efficiency of system parameters and effectively improve the system performance after parameter adjustment.

[0006] In a first aspect, the present application provides a parameter recommendation method, which may be implemented by a server, which may be an electronic device or a component in an electronic device, such as a chip or a chip system.

[0007] The method includes: a server identifying first performance status information of a vehicle control device using a parameter determination model corresponding to the vehicle control device, obtaining M first parameter groups of the vehicle control device and second performance status information corresponding to each first parameter group, wherein the second performance status information indicates the performance status of the vehicle control device when operating based on the corresponding first parameter group. The server sends recommendation information, which includes the M first parameter groups and the second performance status information corresponding to each first parameter group. The recommendation information can indicate the performance status that can be achieved by the vehicle control device using the first parameter group, and can ensure that the performance status of the vehicle control device using the first parameter group meets performance requirements, that is, it can improve the efficiency of parameter adjustment and effectively enhance the system performance after parameter adjustment.

[0008] In one possible implementation, the recommendation information may also include performance boundary information, and the performance boundary information may be determined based on the historical performance status information of the vehicle control device. The recommendation information is used to generate a recommendation interface. In the recommendation interface generated based on the recommendation information, the positional relationship between the second performance status information corresponding to each first parameter group and the performance boundary information in the coordinate system is included. By presenting the positional relationship between the second performance status information corresponding to each first parameter group and the performance boundary information in the coordinate system in the recommendation interface, the gap between the second performance status information corresponding to the first parameter group and the optimal performance status of the vehicle control device can be visualized, so that the user can determine the first parameter group adopted for parameter adjustment of the vehicle control device.

[0009] Optionally, the first performance status information may include one or more values ​​of each of N performance parameters, where N is a positive integer. The N performance parameters may include, but are not limited to, at least one of CPU processing performance, AI processing performance, and image signal processing performance.

[0010] Optionally, the second performance status information may include one or more values ​​of each of N performance parameters, where N is a positive integer. The N performance parameters may include, but are not limited to, at least one of CPU processing performance, AI processing performance, and image signal processing performance.

[0011] In one possible implementation, each of the M first parameter groups corresponds to a recommended parameter. The first parameter group optimizes the performance of the recommended parameters of the vehicle control device. The recommended parameter is one of the N performance parameters, or the recommended parameter is determined by at least two of the N performance parameters. The recommended parameter may reflect the recommended dimension of the corresponding first parameter group to indicate the performance advantage that can be achieved by the vehicle control device using the first parameter group.

[0012] In one possible implementation, the server can input the first performance status information into a parameter determination model to obtain K second parameter groups, send the K second parameter groups, and then receive M first parameter groups and the second performance status information corresponding to each first parameter group. The M first parameter groups are obtained by iteratively optimizing each of the K second parameter groups until the performance status of the vehicle control device meets a preset performance status threshold. This method further ensures that system performance is effectively improved after parameter adjustment by further screening and / or fine-tuning the K parameter groups output by the parameter determination model to obtain M parameter groups for adjusting the parameters of the vehicle control device.

[0013] In a possible implementation, the server may sort the K second parameter groups according to the Pareto algorithm, and send the sorted K second parameter groups to facilitate the receiving end to subsequently screen the K second parameter groups.

[0014] In one possible implementation, the performance boundary information includes multiple performance boundary values, and the performance state threshold is associated with the performance boundary information. For example, the performance state threshold can be a performance boundary value. For another example, the distance between the performance state threshold and the performance boundary value can be less than or equal to a preset value. The performance boundary information can be determined based on historical performance state information of the vehicle control device. The performance boundary information can indicate the optimal performance state that the vehicle control device can achieve. The association of the performance state threshold with the performance boundary information can ensure that the first parameter group selected can enable the vehicle control device to have a better performance state.

[0015] In one possible implementation, the server may send recommendation information for K second parameter groups, including information about each of the K second parameter groups. If the K second parameter groups are used as the final recommended parameters, a parameter group may be selected from the K second parameter groups and the M first parameter groups for parameter adjustment by the vehicle control device, thereby expanding the range of parameter group selection. If the K second parameter groups are not used as the final recommended parameters, data from the execution process of the parameter recommendation method may be presented to enhance the credibility of the recommendation results.

[0016] In a possible implementation, the server trains the parameter determination model based on the first performance status information and the M first parameter groups to obtain an updated parameter determination model, so as to continuously improve the recognition accuracy of the parameter determination model.

[0017] In a possible implementation, the server may determine the performance boundary information according to the historical performance status information of the vehicle control device through a Pareto algorithm.

[0018] In second aspect, the present application provides a control device, including: an acquisition module for acquiring first performance status information of a vehicle control device and a parameter determination model corresponding to the vehicle control device; a processing module for acquiring, through the parameter determination model, M first parameter groups of the vehicle control device and second performance status information corresponding to each first parameter group based on the first performance status information, the second performance status information indicating the performance status of the vehicle control device when operating based on the first parameter group; a sending module for sending recommendation information, the recommendation information including the second performance status information corresponding to each first parameter group in the M first parameter groups.

[0019] In one possible implementation, the recommendation information also includes performance boundary information, which is determined based on the historical performance status information of the vehicle control device. The recommendation information is used to generate a recommendation interface, which includes the positional relationship between the second performance status information and the performance boundary information corresponding to each first parameter group in the coordinate system.

[0020] In one possible implementation, the first performance status information includes one or more values ​​of each of N performance parameters, and / or the second performance status includes one or more values ​​of each of N performance parameters; the N performance parameters include: at least one of: central processing unit CPU processing performance, artificial intelligence AI processing performance and image signal processing performance.

[0021] In one possible implementation, each of the M first parameter groups corresponds to a recommended parameter, and the first parameter group optimizes the performance of the recommended parameters of the vehicle control device. The recommended parameter is one of the N performance parameters, or the recommended parameter is determined by at least two of the N performance parameters.

[0022] In one possible implementation, the processing module is specifically used to: input the first performance status information into a parameter determination model to obtain K second parameter groups; send the K second parameter groups; receive M first parameter groups and the second performance status information corresponding to each first parameter group, where the M first parameter groups are obtained by iteratively optimizing each second parameter group in the K second parameter groups until the performance status of the vehicle control device meets a preset performance status threshold.

[0023] In a possible implementation manner, the processing module is specifically configured to: sort the K second parameter groups according to a Pareto algorithm; and send the sorted K second parameter groups.

[0024] In a possible implementation, the performance status threshold is associated with performance boundary information.

[0025] In a possible implementation, the sending module is further configured to: send recommendation information of the K second parameter groups, where the recommendation information includes information of each parameter group in the K second parameter groups.

[0026] In a possible implementation, the processing module is further configured to: train a parameter determination model according to the first performance status information and the M first parameter groups to obtain an updated parameter determination model.

[0027] In a possible implementation, the processing module is further configured to: determine the performance boundary information according to the historical performance status information of the vehicle control device through a Pareto algorithm.

[0028] In a third aspect, the present application provides an electronic device comprising: a processor and a memory, the memory being used to store a computer program, the processor being used to call and run the computer program stored in the memory to implement the method in the first aspect or various possible implementations.

[0029] In a fourth aspect, the present application provides a chip, comprising: a processor, configured to call and execute computer instructions from a memory to implement the method in the first aspect or various possible implementations.

[0030] In a fifth aspect, the present application provides a computer-readable storage medium for storing computer program instructions, which enable a computer to execute the method in the first aspect or each possible implementation manner.

[0031] In a sixth aspect, the present application provides a computer program product, comprising computer program instructions, which enable a computer to execute the method in the first aspect or any possible implementation manner.

[0032] The beneficial effects of the contents involved in the above-mentioned second to sixth aspects and each possible implementation method can be referred to the beneficial effects brought about by the above-mentioned first aspect and each possible implementation method of the first aspect, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] FIG1 is a schematic diagram of the structure of a computing platform provided in an embodiment of the present application;

[0034] FIG2 is a schematic structural diagram of a vehicle control system provided in an embodiment of the present application;

[0035] FIG3 is a schematic structural diagram of an electronic device provided in an embodiment of the present application;

[0036] FIG4 is a flow chart of a parameter recommendation method provided in an embodiment of the present application;

[0037] FIG5a is a schematic diagram of a Pareto frontier provided in this application;

[0038] FIG5 b is a schematic diagram of a Pareto-based performance frontier provided in an embodiment of the present application;

[0039] FIG6 is a schematic diagram of a performance boundary provided in an embodiment of the present application;

[0040] FIG7 is a schematic block diagram of a parameter recommendation device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application.

[0042] The vehicle in the embodiments of the present application may be an intelligent vehicle, such as an autonomous driving vehicle that automatically controls all functions, or an assisted driving vehicle that automatically controls some functions to provide driving assistance, or may be an ordinary vehicle. For ease of description, the vehicle will be collectively referred to as a vehicle below. When the vehicle in the embodiments of the present application is an ordinary vehicle, the vehicle may be automatically controlled by external devices of the vehicle.

[0043] The present application can realize vehicle control for any type of vehicle, such as fuel vehicles, new energy vehicles or extended-range vehicles. When the vehicle is a fuel vehicle, the fuel is used to provide kinetic energy for the vehicle and the control logic functions of the equipment; when the vehicle is a new energy vehicle, the new energy is used to provide kinetic energy for the vehicle and the control logic functions of the equipment, wherein the new energy may include electric energy, methane, etc.; when the vehicle is an extended-range vehicle, the fuel and electric energy are used to provide kinetic energy for the vehicle and the control logic functions of the equipment. The above energy that provides kinetic energy for the vehicle, or the energy provided for the vehicle power system, and the energy provided for the control logic functions of the vehicle can be collectively referred to as vehicle energy supply. For ease of understanding, the following explanation is given by taking the example of providing kinetic energy for the vehicle and the control logic functions of the equipment through electric energy, which should not be understood as a restrictive description.

[0044] Intelligent driving uses artificial intelligence to assist or replace human drivers in driving, potentially addressing the shortcomings of human drivers. Through extensive research and development, the performance and technology of the various sensors and computers required for intelligent driving have significantly improved, while costs are also gradually decreasing. As a result, intelligent driving has been widely adopted in a variety of areas, including automated parking, adaptive cruise control (ACC) following, automatic emergency braking, lane departure warning, and autonomous driving.

[0045] The computing platform for intelligent driving services, deployed with software that runs various intelligent driving functions, serves as the "brain" of the intelligent driving system. Various sensors are connected to the computing platform via different transmission methods. The algorithms running on the computing platform process these signals, generating control signals that are sent to downstream execution units. As shown in FIG1 , the computing platform 100 in a vehicle can include multiple heterogeneous processing units to implement intelligent driving services. These processing units can be implemented as processing chips or devices including processing chips, or as modules within a processing chip. These processing units can include, for example, a central processing unit (CPU), a graphics processing unit (GPU), an image sensor processor (ISP), a digital vision pre-processing (DVPP) module, a natural processing unit (NPU), or a microcontroller unit (MCU). These processing units can be integrated into the same system on a chip (SoC). Among them, CPU, GPU, ISP, DVPP, and NPU can be called computing units, and MCU can be called a control unit. The MCU generates a control signal based on the input of at least one computing unit and sends it to the downstream execution unit, such as the actuator, to achieve control of the vehicle. In addition, the computing platform 100 can also include a data exchange unit, such as an input and output (I / O) interface, a storage unit, etc. The units in the computing platform 100 can be connected through a bus. Different processing units can be used to perform different processing tasks, such as image processing tasks, computing tasks, AI reasoning tasks, etc. The above-mentioned computing platform 100 can be deployed in a vehicle control device in a vehicle. In some scenarios, in order to improve the management capabilities of each processing unit, the vehicle system can be divided into multiple domains (such as vehicle control domain, intelligent driving domain, cockpit domain, etc.), and the vehicle control device can be one or more domain controllers. For example, the vehicle control device can be an intelligent driving domain controller, or a mobile data center (MDC).

[0046] Different processing units in the computing platform 100 share memory bandwidth during operation. When bandwidth competition occurs between different processing units, serious system performance issues can arise, such as image data frame loss, excessive CPU usage, and excessive AI inference latency. To alleviate the memory bandwidth bottleneck, system parameters can be adjusted to prevent system performance issues caused by memory bandwidth competition when the computing platform 100 operates based on the adjusted system parameters.

[0047] Currently, with the help of some flow control mechanisms or optimization tools, the performance status of each processing unit in the computing platform can be monitored. The performance status of the processing unit may include, for example, the CPU occupancy rate, GPU usage, memory bandwidth occupancy, etc. By observing the performance status of the computing platform, users can analyze the services that occupy the memory bandwidth and optimize the services that occupy the memory bandwidth by adjusting the system parameters. However, determining the adjustment amount of system parameters through manual analysis is inefficient and it is difficult to ensure that the system performance can be effectively improved after the parameter adjustment. In particular, when there are many system parameters, this parameter adjustment method will be even more inefficient and low-precision.

[0048] In response to the above technical problems, the present application identifies the current performance status of the computing platform through a parameter determination model to determine the recommended parameter groups (such as the M first parameter groups mentioned below), and determines the performance status of the computing platform after parameter adjustment based on each parameter group. The user can adjust the parameters of the computing platform based on the determined parameter group, thereby improving the efficiency of parameter adjustment and ensuring that the system performance after parameter adjustment is effectively improved. Furthermore, the user can determine whether to select the corresponding parameter group for parameter adjustment based on the performance status of the computing platform after parameter adjustment, so that the system performance after parameter adjustment meets the user's expectations.

[0049] To facilitate understanding, the terms involved in this application are first explained below.

[0050] 1. System parameters. The system parameters in the embodiments of the present application include system parameters corresponding to different processing units in the vehicle control device. The system parameters corresponding to each processing unit may include, but are not limited to, configuration parameters under the memory system resource partitioning and monitoring mechanism (MPAM), configuration parameters for heterogeneous scheduling, and other flow control configuration parameters. The configuration parameters under MPAM can be shown in Table 1 below.

[0051] Table 1

[0052] The above system parameters may be included in a configuration file, and the vehicle control device may read the configuration file and adopt the corresponding system parameters during operation. In the embodiment of the present application, the adjustment operation of the system parameters may be achieved by modifying the system parameters in the configuration file.

[0053] In the embodiments of the present application, a parameter group may include one or more parameters, which may include the aforementioned system parameters. Hereinafter, the first parameter group and the second parameter group are used to distinguish different parameter groups, and both may be collectively referred to as parameter groups. When referring to parameter groups below, the number of parameter groups is not limited, and for example, one or more parameter groups may be included.

[0054] 2. Performance status information, used to indicate the performance status of the vehicle control device. The performance status information includes but is not limited to at least one of CPU processing performance information, AI processing (such as NPU) performance information, and image signal processing (such as ISP, DVPP) performance information. For example, CPU processing performance may include the number of times the business running on the CPU is executed per second. AI processing performance is similar. For the sake of brevity, it will not be repeated. Image signal processing performance may include an execution frequency similar to CPU processing performance, or may also include a loss rate (such as a frame loss rate). In order to distinguish the performance status of the vehicle control device under different system parameters, in an embodiment of the present application, the performance status of the vehicle control device before parameter adjustment is indicated by the first performance status information, and the performance status of the vehicle control device after parameter adjustment is indicated by the second performance status information. The first performance status information and the second performance status information can be summarized as performance status information.

[0055] FIG2 is a schematic diagram of the structure of a vehicle control system 200 provided in an embodiment of the present application. Referring to FIG2 , vehicle control system 200 may include a client 210, a server 220, and a tool 230. Client 210 and server 220 may be connected wirelessly, client 210 and tool 230 may be connected wirelessly, and server 220 and tool 230 may be connected via wired or wireless communication.

[0056] It should be understood that the parameter recommendation method provided in the embodiments of the present application can be applied during a vehicle's testing phase, maintenance phase, or normal driving phase. When applied during a vehicle's testing phase or maintenance phase, the method can provide the vehicle control system 200 with a simulated vehicle pressure, or a simulated vehicle configuration. The simulated vehicle pressure can be a preset fixed pressure.

[0057] The client 210 can be deployed in a vehicle. For example, the client 210 can be connected to the vehicle control device via a wired or wireless connection. The client 210 can obtain system parameter information, performance status information, and the like from the vehicle control device, and can also adjust the system parameters of the vehicle control device to optimize its performance.

[0058] Exemplarily, the client 210 may include a parameter adjustment module 211, a performance acquisition module 212, and a parameter optimization module 213. When the vehicle is tested or maintained based on a fixed pressure, the parameter adjustment module 211 may set the default system parameters based on the fixed pressure, such as setting the MPAM configuration parameters, heterogeneous scheduling configuration parameters, and other flow control configuration parameters to the default values. The performance acquisition module 212 obtains performance status information (such as the first performance status information below) from the vehicle control device, for example, including CPU processing performance, NPU processing performance, ISP processing performance, and DVPP processing performance. The performance acquisition module 212 may send the obtained performance status information of the vehicle control device (such as the first performance status information below) to the server 220. The parameter optimization module 213 can receive the parameter group sent by the server end 220 (such as the K second parameter groups mentioned below), and then cooperate with the parameter adjustment module 211 to optimize the parameter group to obtain the optimized parameter group (such as the M first parameter groups mentioned below) and the performance status information corresponding to the optimized parameter group (such as the second performance status information mentioned below), and send the optimized parameter group and the second performance status information to the tool end 230 for presentation.

[0059] The server 220 can analyze the performance status information of the vehicle control device obtained by the client 210 to determine a parameter group for optimizing the performance status of the vehicle control device, and send the parameter group for optimizing the performance status of the vehicle control device to the client 210, so that the client 210 can adjust the parameters of the vehicle control device based on the parameter group to optimize the performance status of the vehicle control device.

[0060] Exemplarily, the server 220 may include a data storage module 221, a model matching module 222, a model processing module 223, and a parameter group sorting module 224. The data storage module 221 may receive and store performance status information (such as the first performance status information described below) from the performance acquisition module 212. This application does not limit the storage mode of the data storage module 221. The model matching module 222 is used to obtain a parameter determination model that matches the vehicle control device. Optionally, the model matching module 222 may determine the corresponding parameter determination model based on the identifier of the vehicle control device. The identifier of the vehicle control device may be sent from the client 210 to the server 220 and stored in the data storage module 221. The model matching module 222 may obtain the identifier of the vehicle control device from the data storage module 221. Furthermore, the model matching module 222 can send the parameter determination model to the model processing module 223, and the model processing module 223 inputs the performance status information from the performance acquisition module 212 (such as the first performance status information below) into the parameter determination model to obtain at least one parameter group. Furthermore, the parameter sorting module 224 sorts the at least one parameter group output by the parameter determination model, and obtains at least one parameter group (such as the K second parameter groups below) based on the sorting result, and then sends the at least one parameter group (such as the K second parameter groups below) to the parameter optimization module 213.

[0061] The tool end 230 can be used to perform parameter group recommendations. Exemplarily, the tool end 230 may include a visualization module 231 and a parameter recommendation module 232. The visualization module 231 can visualize the acquired data. For example, the visualization module 231 renders and displays the performance status information (such as the second performance status information corresponding to the M first parameter groups described below) sent by the client 210 or the server 220. The user can select one of the second performance status information through human-computer interaction. The parameter recommendation module 232 can render and display the first parameter group corresponding to the second performance status information based on the user's selection operation to implement parameter recommendation.

[0062] The server 220 and the tool 230 can be deployed on a server, a cloud server, a server cluster, or a cloud server cluster. The server 220 and the tool 230 can be deployed independently or integrated. When the server 220 and the tool 230 are deployed integrated, the server 220 and the tool 230 can be considered different software modules or the same software module, which is not limited in this application.

[0063] It should be noted that the division of the units / modules in the above devices is only a division of logical functions, and in actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated.

[0064] In some embodiments, the client 210 in the above-mentioned vehicle control system 200 can be implemented as an electronic device 300 as shown in FIG. 3 .

[0065] In some embodiments, the server 220 in the vehicle control system 200 may be implemented as an electronic device 300 as shown in FIG. 3 .

[0066] In some embodiments, the tool end 230 in the vehicle control system 200 may be implemented as an electronic device 300 as shown in FIG. 3 .

[0067] When the server end 220 and the tool end 230 are deployed in an integrated manner, the server end 220 and the tool end 230 can be integrated into the electronic device 300 as shown in FIG. 3 .

[0068] Figure 3 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. As shown in Figure 3, the electronic device 300 may include: a processor 310 and a memory 320. The processor 310 and the memory 320 communicate with each other through an internal connection path. Optionally, the memory 320 may include a read-only memory and a random access memory, and provide instructions and data to the processor 310. A portion of the memory 320 may also include a non-volatile random access memory. The memory 320 may be a separate device or integrated in the processor 310. The processor 310 may be used to execute instructions stored in the memory 320, and when the processor 310 executes instructions stored in the memory 320, the processor 310 is used to execute the various steps and / or processes of the method embodiment of the present application.

[0069] During implementation, each step of the method in the embodiment of the present application can be performed by the hardware processor in the processor 310, or by a combination of hardware and software modules in the processor 310. The software module can be located in a storage medium well-known in the art. The storage medium is located in the memory 320. The processor 310 reads the information in the memory 320 and, in conjunction with its hardware, completes the steps of the following method.

[0070] In some embodiments, the apparatus 300 may further include an input interface 330. The processor 310 may control the input interface 330 to communicate with other devices or chips, and specifically, may obtain information or data sent by other devices or chips.

[0071] In some embodiments, the apparatus 300 may further include an output interface 340. The processor 310 may control the output interface 340 to communicate with other devices or chips, and specifically, may output information or data to other devices or chips.

[0072] In some embodiments, the device 300 can implement the corresponding processes of each method in the embodiments of the present application. For the sake of brevity, they will not be repeated here.

[0073] It should be noted that the processor 310 in the embodiment of the present application can be an integrated circuit chip with signal processing capabilities. The above-mentioned processor 310 can be a microprocessor, a central processor unit (CPU), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component. The various methods, steps, and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed.

[0074] It is understood that the memory in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct RAM bus RAM (DR RAM). It should be noted that the memory of the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0075] The parameter recommendation method provided in the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0076] The following is only for the convenience of understanding and explanation, and the method provided by the embodiment of the present application is described with the server as the execution subject. The server can be, for example, the server 220 in Figure 2 above. This application does not limit the execution subject of the method. As long as a program that can implement the code of the method provided by the embodiment of the present application can be run, it can serve as the execution subject of the method provided by the embodiment of the present application. For example, the server can be implemented as a server or a control module in the server, or as a chip or chip system in the server, or as other functional modules that can call and execute programs.

[0077] In some embodiments, the present application takes the interaction between the client, the server and the tool as an example to illustrate the method provided in the embodiment of the present application. The client can be, for example, the client 210 in Figure 2 above, and the client can be implemented as a vehicle or a controller in a vehicle, or can be implemented as a chip or chip system in a vehicle, or can be implemented as other functional modules that can call programs and execute programs; the tool side can be, for example, the tool side 230 in Figure 2 above, and the tool side can be implemented as a server or a control module in a server, or can be implemented as a chip or chip system in a server, or can be implemented as other functional modules that can call programs and execute programs. As mentioned above, the tool side and the server side can be deployed independently or in an integrated manner. For the sake of convenience, the following description will be made using the integrated deployment of the tool side and the server side as an example.

[0078] Fig. 4 is a flow chart of a parameter recommendation method 400 provided in an embodiment of the present application. As shown in Fig. 4 , the method 400 may include part or all of the following processes.

[0079] S410, obtaining first performance status information of a vehicle control device and a parameter determination model corresponding to the vehicle control device;

[0080] S420, obtaining, using a parameter determination model and based on the first performance status information, M first parameter groups of the vehicle control device and second performance status information corresponding to each first parameter group, wherein the second performance status information indicates a performance status of the vehicle control device when operating based on the first parameter group, where M is a positive integer;

[0081] S430: Send recommendation information, where the recommendation information includes second performance status information corresponding to each of the M first parameter groups.

[0082] In S410 above, the first performance status information may indicate the performance status of the vehicle control device before parameter adjustment. In other words, the first performance status information may be performance status information obtained from the vehicle control device before parameter adjustment. As previously mentioned, vehicle parameters may be adjusted during a testing phase, maintenance phase, or normal driving phase. After enabling parameter adjustment mode, the client may obtain the first performance status information of the vehicle control device and transmit the first performance status information to the server.

[0083] Optionally, the first performance status information may include one or more values ​​of each of N performance parameters, where N is a positive integer. The N performance parameters may include, but are not limited to, at least one of CPU processing performance, AI processing performance, and image signal processing performance.

[0084] Generally speaking, the performance state of the vehicle control device indicated by the first performance state information does not meet the performance requirements of the vehicle control device, and therefore it is necessary to adjust the parameters to make the performance state of the vehicle control device meet the performance requirements. The performance requirements can be determined based on user needs or based on application scenarios. Exemplarily, the first performance state information does not meet the performance requirements of the vehicle control device, which may mean that the performance state of the vehicle control device indicated by the first performance state information does not meet the preset performance state threshold. For example, when the performance requirement is optimal AI processing performance, the value of the AI ​​processing performance of the vehicle control device indicated by the first performance state information is less than or equal to the preset AI processing performance threshold; for another example, when the performance requirement is optimal CPU processing performance, the value of the CPU processing performance of the vehicle control device indicated by the first performance state information is less than or equal to the preset CPU processing performance threshold.

[0085] It should be noted that different vehicle control devices may correspond to different parameter determination models. Different vehicle control devices may be, for example, at least one of vehicle control devices from different manufacturers, vehicle control devices of different models, and vehicle control devices with different configurations (including but not limited to software configuration and / or hardware configuration). The server may obtain the corresponding parameter determination model based on the matching of the vehicle control device. For example, the server may preset a correspondence between the vehicle control device and the parameter determination model. The server may determine the parameter determination model corresponding to the vehicle control device in the correspondence based on the identifier of the vehicle control device. Optionally, the identifier of the vehicle control device may be obtained by the server from the client. For example, when the server and the client establish a connection, the server receives the identifier of the vehicle control device sent by the client. It should be noted that establishing a correspondence between the identifier of the vehicle control device and the parameter determination model is only an example, and this application does not limit this. For example, in the case where vehicle control devices with different configurations correspond to different parameter determination models, a correspondence may be established between the configuration of the vehicle control device and the parameter determination model. The server may determine the parameter determination model corresponding to the vehicle control device in the correspondence based on the configuration of the vehicle control device.

[0086] The parameter determination model is used to analyze the input first performance status information to obtain at least one parameter set for optimizing the performance status of the vehicle control device. In this embodiment of the present application, the parameter determination model corresponding to the vehicle control device can be preset or trained by the server. The following is an exemplary description of a parameter determination model corresponding to the vehicle control device trained by the server.

[0087] For example, the server may perform iterative training based on multiple training data to obtain a parameter determination model. The training data may be obtained based on historical data of the vehicle control device, or the training data may be based on experimental data or empirical settings. Each training data may include a performance status information sample and a parameter group sample of the vehicle control device. It should be understood that the performance status information sample may indicate the performance status of the vehicle control device before parameter adjustment, and the performance status of the vehicle control device after the vehicle control device operates using the parameter group sample so that the performance status of the vehicle control device meets the performance requirements.

[0088] The server can train the parameter determination model corresponding to the vehicle control device according to the pre-training method in the above example, or the offline training method. It should be noted that this application does not limit the executor of the model training. For example, the server can receive the parameter determination model corresponding to the vehicle control device sent by other devices or equipment. In addition, the parameter determination model corresponding to the vehicle control device can also be preset on the server.

[0089] In the above S420, the above first parameter group is used to implement parameter adjustment of the vehicle control device. Specifically, after the vehicle side adjusts the parameters of the vehicle control device based on the obtained first parameter group, the performance status of the vehicle control device can meet the performance requirements, that is, the value of the performance status of the vehicle control device indicated by the second performance status information corresponding to the first parameter group is greater than or equal to the preset performance status threshold.

[0090] Optionally, the second performance state information may include one or more values ​​of each of N performance parameters, where N is a positive integer. The N performance parameters may include, but are not limited to, at least one of CPU processing performance, AI processing performance, and image signal processing performance. Generally speaking, the second performance state information and the first performance state information may correspond to the same performance parameters, but this application is not limited to this. For example, the second performance state information may correspond to more or fewer performance parameters than the first performance state information.

[0091] The performance status threshold may be different in different user requirements or different application scenarios. For example, when the performance requirement is optimal AI processing performance, the value of the AI ​​processing performance of the vehicle control device indicated by the second performance status information is greater than or equal to the preset first AI processing performance threshold; for another example, when the performance requirement is optimal CPU processing performance, the value of the CPU processing performance of the vehicle control device indicated by the second performance status information is greater than or equal to the preset first CPU processing performance threshold; for another example, when the performance requirement is optimal image signal processing performance, the value of the image signal processing performance of the vehicle control device indicated by the second performance status information is greater than or equal to the preset image signal processing performance threshold; for another example, when the performance requirement is balanced configuration of AI and CPU, the value of the AI ​​processing performance of the vehicle control device indicated by the second performance status information is greater than or equal to the preset second AI processing performance threshold, and the value of the CPU processing performance of the vehicle control device indicated by the second performance status information is greater than or equal to the preset second CPU processing performance threshold; and so on.

[0092] In the first example of S420 above, any parameter group output by the server through the parameter determination model can be directly used to implement parameter adjustment of the vehicle control device. In this case, at least one parameter group output by the parameter determination model is the above-mentioned M first parameter groups; in the second example of S420 above, any parameter group output by the server through the parameter determination model needs to be processed (such as screening and / or parameter fine-tuning) to obtain a parameter group that can act on the vehicle control device. In this case, at least one parameter group output by the parameter determination model can be the K second parameter groups mentioned below, where K is a positive integer, and the K second parameter groups are processed to obtain M first parameter groups.

[0093] In the first example mentioned above, the server directly obtains the parameter group for adjusting the parameters of the vehicle control device through the parameter determination model, which has high processing efficiency; in the second example mentioned above, the parameter group output by the parameter determination model is further screened and / or fine-tuned to obtain the parameter group for adjusting the parameters of the vehicle control device, which can further ensure that the system performance after the parameter adjustment is effectively improved.

[0094] Based on the first example above, the server can send M first parameter groups to the client. For each first parameter group, the client can configure the first parameter group through the parameter adjustment module, and obtain the second performance status information corresponding to the first parameter group through the performance acquisition module when the vehicle control device is running based on the first parameter group. The client sends the second performance status information corresponding to each first parameter group in the M first parameter groups to the server.

[0095] Based on the second example described above, the server may send K second parameter groups to the client. For each second parameter group, the client may iteratively optimize the second parameter group until the performance state of the vehicle control device satisfies a preset performance state threshold, thereby obtaining one or more second parameter groups corresponding to the second parameter group. It should be understood that not every second parameter group can be iteratively optimized to obtain the corresponding first parameter group. For example, if the performance state of the vehicle control device still does not meet the preset performance state threshold after the client performs iterative optimization based on the second parameter group for a preset number of iterations, then the second parameter group cannot be iteratively optimized to obtain any first parameter group.

[0096] Exemplarily, the client may send the second parameter group to the parameter adjustment module through the parameter optimization module, configure the second parameter group through the parameter adjustment module, and obtain the current performance status information through the performance acquisition module when the vehicle control device operates based on the second parameter group. When the current performance status of the vehicle control device meets the preset performance status threshold, the client uses the current performance status information of the vehicle control device as the second performance status information, and uses the parameter group currently participated in by the vehicle control device as the first parameter group, and sends the first parameter group and the second performance status information to the server. When the current performance status of the vehicle control device does not meet the preset performance status threshold, the client fine-tunes the parameters of the second parameter group through the parameter optimization module, configures the adjusted second parameter group through the parameter adjustment module, and obtains the current performance status information through the performance acquisition module when the vehicle control device operates based on the second parameter group, and then determines whether the current performance status meets the preset performance status threshold. The above process is repeated until the current performance status of the vehicle control device meets the preset performance status threshold, and the current performance status information of the vehicle control device is used as the second performance status information, and the parameter group currently participated in by the vehicle control device is used as the first parameter group, and the client sends the first parameter group and the second performance status information to the server.

[0097] Exemplarily, before the client sends the second parameter group to the parameter adjustment module through the parameter optimization module, the second parameters can be preliminarily screened and / or fine-tuned to improve the applicability of the second parameters in the vehicle control device and reduce the complexity of iterative optimization.

[0098] In the second example above, the server can sort the K second parameter groups through the parameter group sorting module to implement the screening of the K second parameter groups. Exemplarily, the server can sort the K second parameter groups based on at least one recommended parameter. Among them, at least one recommended parameter can be a recommended parameter on a different dimension. The recommended parameter can be one of the N performance parameters mentioned above. For example, the recommended parameter can be CPU processing performance, AI processing performance or image signal processing performance, etc.; or the recommended parameter can be determined by at least two performance parameters among the N performance parameters. For example, the recommended parameter can be a balanced configuration of CPU processing performance and AI processing performance, or a balanced configuration of CPU processing performance and image signal processing performance, or a balanced configuration of AI processing performance and image signal processing performance, or a balanced configuration of CPU processing performance, AI processing performance and image signal processing performance. Exemplarily, after sorting the K second parameter groups based on at least one recommended parameter, each recommended parameter may correspond to at least one sorted second parameter group. For example, the at least one second parameter group includes at least one second parameter group sorted from high to low according to CPU processing performance, at least one second parameter group sorted from high to low according to AI processing performance, and at least one second parameter group sorted from high to low according to a balanced configuration of CPU processing performance and AI processing performance.

[0099] Optionally, when considering the combined optimization of multiple performance parameters, it can be understood as solving the Pareto optimal problem. Based on this, the server can sort the K second parameter groups according to the Pareto algorithm. As shown in Figure 5a, taking two performance parameters as an example, the horizontal axis of the coordinate is CPU processing performance and the vertical axis is AI processing performance. The dotted line indicates the Pareto optimal boundary, also known as the Pareto optimal frontier. Points on the Pareto optimal boundary, such as point A and point B, have better CPU processing performance than point A, and better AI processing performance than point A. However, under the balanced configuration of CPU processing performance and AI processing performance, point A and point B are both optimal solutions, such as point A and point B are ranked better than point C.

[0100] Continuing with the above example, after the server sorts the K second parameter groups, it can send the K second parameter groups to the client, and the client can filter the K second parameter groups based on the sorting of the K second parameter groups, for example, selecting one or more second parameter groups that are ranked higher under each recommended parameter; or after the server sorts the K second parameter groups, it can filter the K second parameter groups, for example, selecting one or more second parameter groups that are ranked higher under each recommended parameter, and send at least one of the filtered second parameter groups to the client.

[0101] In some embodiments, in order to continuously improve the accuracy of the parameter determination model and ensure that the system performance after parameter adjustment is effectively improved, the server can update the parameter determination model based on the execution process of the above-mentioned S420. For example, in the second example of the above-mentioned S420, the server outputs K second parameter groups based on the first performance status information through the parameter determination model, and obtains M first parameter groups determined based on the K second parameter groups. Then, the parameter determination model is trained based on the first performance status information and the obtained M first parameter groups to obtain an updated parameter determination model.

[0102] In some embodiments, the above-mentioned preset performance state threshold is associated with performance boundary information. For example, the performance state threshold may be a performance boundary value. For another example, the distance between the performance state threshold and the performance boundary value may be less than or equal to a preset value. The performance boundary information may be determined based on the historical performance state information of the vehicle control device. The performance boundary information may indicate the optimal performance state that the vehicle control device can achieve. For example, the performance boundary information may indicate the optimal performance boundary that the vehicle control device can achieve under N performance parameters.

[0103] See Figure 6, the horizontal axis is the CPU processing performance, the vertical axis is the AI ​​processing performance, and the grayscale value indicates the ISP performance (such as ISP loss rate, the smaller the grayscale value, the lower the ISP loss rate, and the larger the grayscale value, the higher the ISP loss rate). Each point in Figure 6 corresponds to a parameter group, and the coordinates of the point are determined based on the performance state of the vehicle control device when it is running with the parameter group. The broken line in Figure 6 (in some examples, it can also be a curve or a straight line) represents the boundary of the vehicle control device, and the points on the broken line (such as point 77, point 48, and point 8) can respectively correspond to a first parameter group. The points on the broken line (such as point 77, point 48, and point 8) can respectively be the first parameter groups corresponding to different recommended parameters. For example, see Table 2:

[0104] Table 2

[0105] For example, the performance boundary in Figure 6 can be determined based on the Pareto algorithm, the implementation of which has been described in the previous example and will not be repeated for the sake of brevity. Referring to Figure 5b, the horizontal axis represents CPU processing performance, the vertical axis represents AI processing performance, and the grayscale value indicates ISP performance. Considering the Pareto optimal boundary of the three performance parameters, this optimal boundary can be used to screen out "pseudo-optimal" values. For example, when the parameter group corresponding to point 86 in the figure acts on the vehicle control device, although the CPU processing performance and AI performance are good, the ISP frame loss is severe and does not meet the requirements.

[0106] It should be noted that, for ease of understanding, only three performance parameters are used as an example in the embodiments of this application, but this application is not limited to this. The server can also determine the performance boundaries and sort the parameter groups based on more or fewer performance parameters. For example, the server can determine the performance boundary information based on historical performance status information based on CPU processing performance, AI processing performance, ISP performance, and DVPP performance.

[0107] In the above S430, the server side can send recommendation information to the tool side, and the recommendation information includes the second performance status information corresponding to each first parameter group in the M first parameter groups. After receiving the recommendation information, the tool side can render the recommendation information, generate a recommendation interface, and present the recommendation interface through the display screen. When the server side and the tool side are deployed in an integrated manner, such as the tool side is integrated and deployed on the server side, the server side can generate a recommendation interface based on the recommendation information and present the recommendation interface through the display screen. It should be noted that this application does not limit the recommendation information to be presented through the display screen. For example, it can also be played through a speaker.

[0108] Through the recommendation information, the performance status of the vehicle control device when running based on each first parameter group can be presented, so that the user can select the first parameter group that meets the performance requirements. The embodiment of the present application does not limit the presentation method of the recommendation information. For example, the second performance status information corresponding to the M first parameter groups can be presented in the form of coordinates. See Figure 7. The horizontal axis of the coordinate is CPU processing performance, the vertical axis is AI processing performance, and the grayscale value indicates ISP performance. The AI ​​processing performance of point 77 is the best among the three second performance status information, the CPU processing performance of point 8 is the best among the three second performance status information, and the performance of point 48 is the most balanced among the three second performance status information. Assuming that the performance requirement is the best AI processing performance, the user can select point 77 in the recommendation interface, and then select the first parameter group corresponding to point 77. The tool end can send the selected first parameter group to the client, so that the client can adjust the parameters of the vehicle control device based on the selected first parameter group, so that the vehicle control device achieves the performance state indicated by the second performance status information reached by point 77.

[0109] Optionally, in response to the user selecting the second performance status information, the tool may display a first parameter group corresponding to the second performance status information, thereby presenting to the user the first parameter group to be used for adjusting parameters of the vehicle control device. Optionally, the user may determine to adopt the first parameter group or modify the parameters in the first parameter group.

[0110] In some embodiments, in addition to the second performance state information, the recommendation information may also include M first parameter groups. Optionally, the recommendation information may include an identifier of each of the M first parameter groups (such as the aforementioned numbers 77, 48, and 8), and / or the recommendation information may include a value for a parameter in each of the M first parameter groups.

[0111] In some embodiments, the recommendation information may include the content shown in Table 2 above.

[0112] In some embodiments, the recommendation information may also include performance boundary information. As mentioned above, the performance boundary information may be determined based on the historical performance status information of the vehicle control device. For the description of the performance boundary information, please refer to the above example and will not be repeated for the sake of brevity. In the recommendation interface generated based on the recommendation information, the positional relationship between the second performance status information and the performance boundary information corresponding to each first parameter group in the coordinate system is included. It should be understood that by presenting the positional relationship between the second performance status information and the performance boundary information corresponding to each first parameter group in the recommendation interface in the coordinate system, it can be clarified whether the second performance status information corresponding to the first parameter group can achieve the optimal performance state of the vehicle control device. The smaller the difference between the second performance status information corresponding to the first parameter group and the performance boundary information, the better the performance state of the vehicle control device can be achieved after the vehicle control device adopts the first parameter group. For example, if the second performance status information is on the performance boundary information in the coordinate system, it means that the first parameter group corresponding to the second performance status information can achieve a better performance state. In one example, the second performance state information and performance boundary information corresponding to the M first parameter groups presented in the recommendation interface can be seen in Figure 6. The second performance states corresponding to the three first parameter groups are respectively presented as points 77, 48 and 8. Points 77, 48 and 8 are all on the performance boundary of the vehicle control device. Compared with other points in Figure 6, the AI ​​processing performance of point 77 is optimal, the CPU processing performance of point 8 is optimal, and the performance of point 48 is the most balanced.

[0113] In some embodiments, the above S430 can be replaced by the server determining a first parameter group from M first parameter groups according to preset performance requirements, sending the first parameter group to the client, and allowing the client to configure the first parameter group to the vehicle control device to achieve automatic adjustment of parameters, thereby further improving the efficiency of parameter adjustment.

[0114] In some embodiments, the server may send recommendation information for K second parameter groups, and the recommendation information may include information about each of the K second parameter groups, such as the identifier of each parameter group (such as No. 77, No. 48, No. 8, etc.), and the parameters included in each parameter group. The tool may present information about each of the K second parameter groups. In the case where the K second parameter groups are used as the final recommended parameters, the parameter group used when the vehicle control device performs parameter adjustment may be selected from the K second parameter groups and the M first parameter groups to expand the range of parameter group selection; in the case where the K second parameter groups are not used as the final recommended parameters, the execution process data of the parameter recommendation method may be presented to improve the credibility of the recommendation results.

[0115] Therefore, the server in the embodiment of the present application identifies the first performance status information through the parameter determination model corresponding to the vehicle control device, and then obtains M first parameter groups of the vehicle control device and the second performance status information corresponding to each first parameter group. The second performance status information indicates the performance status of the vehicle control device when operating based on the first parameter group, thereby improving the efficiency of parameter adjustment and effectively improving the system performance after parameter adjustment. Furthermore, the recommendation information sent by the server includes the second performance status information corresponding to each first parameter group in the M first parameter groups, so as to characterize the performance status that can be achieved by the vehicle control device using the first parameter group, further ensuring that the performance status achieved by the vehicle control device by the ultimately adopted first parameter group meets the performance requirements.

[0116] In the various embodiments of the present application, unless otherwise specified or there is a logical conflict, the terms and / or descriptions between the various embodiments are consistent and can be referenced by each other. The technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationships.

[0117] FIG7 is a schematic block diagram of a parameter recommendation device 700 according to an embodiment of the present application. As shown in FIG7 , the device 700 at least includes: an acquisition module 710 , a processing module 720 , and a sending module 730 .

[0118] Among them, the acquisition module 710 is used to obtain the first performance status information of the vehicle control device and the parameter determination model corresponding to the vehicle control device; the processing module 720 is used to obtain the M first parameter groups of the vehicle control device and the second performance status information corresponding to each first parameter group based on the first performance status information through the parameter determination model, and the second performance status information indicates the performance status of the vehicle control device when it is running based on the first parameter group; the sending module 730 is used to send recommendation information, and the recommendation information includes the second performance status information corresponding to each first parameter group in the M first parameter groups.

[0119] Exemplarily, the acquisition module 710 may include the data storage module 221 in FIG. 2 , configured to acquire the first performance status information from the client 210 and store the first performance status information.

[0120] Exemplarily, the acquisition module 710 may include the model matching module 222 in FIG. 2 , configured to acquire a parameter determination model that matches the vehicle control device.

[0121] Exemplarily, the processing module 720 may include the model processing module 223 in FIG. 2 , which is used to input the first performance status information of the vehicle control device into a parameter determination model to obtain at least one parameter group (such as K second parameter groups or M first parameter groups).

[0122] Exemplarily, the processing module 720 may include the parameter group sorting module 224 in FIG. 2 , configured to sort at least one parameter group (eg, K second parameter groups or M first parameter groups) output by the parameter determination model.

[0123] Exemplarily, the acquisition module 710 may be implemented as the input interface 330 in FIG. 3 , the sending module 730 may be implemented as the output interface 340 in FIG. 3 , and the processing module 720 may be implemented as the processor 310 in FIG. 3 .

[0124] It should be understood that the specific process of each unit executing the above corresponding steps has been described in detail in the above method embodiment, and for the sake of brevity, it will not be repeated here.

[0125] The division of the modules in the above devices is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated.

[0126] An embodiment of the present application also provides a computer-readable storage medium for storing a computer program.

[0127] In some embodiments, the computer-readable storage medium can be applied to the server, client or tool side in the embodiments of the present application, and the computer program enables the computer to execute the corresponding processes in the various methods of the embodiments of the present application. For the sake of brevity, they will not be repeated here.

[0128] An embodiment of the present application also provides a computer program product, including computer program instructions.

[0129] In some embodiments, the computer program product can be applied to the server, client or tool side in the embodiments of the present application, and the computer program instructions enable the computer to execute the corresponding processes in the various methods of the embodiments of the present application. For the sake of brevity, they will not be repeated here.

[0130] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0131] The above are merely specific embodiments of the present application, but the scope of protection of this application is not limited thereto. Any modifications or substitutions that can be readily conceived by a person skilled in the art within the technical scope disclosed in this application are intended to be encompassed by the scope of protection of this application. Therefore, the scope of protection of this application shall be subject to the scope of protection of the claims.

Claims

1. A parameter recommendation method, characterized in that: include: Acquiring first performance status information of a vehicle control device and a parameter determination model corresponding to the vehicle control device; obtaining, using the parameter determination model and based on the first performance status information, M first parameter groups of the vehicle control device and second performance status information corresponding to each first parameter group, wherein the second performance status information indicates a performance status of the vehicle control device when operating based on the first parameter group; Recommendation information is sent, where the recommendation information includes second performance status information corresponding to each of the M first parameter groups.

2. The method according to claim 1, characterized in that The recommendation information also includes performance boundary information, which is determined based on the historical performance status information of the vehicle control device. The recommendation information is used to generate a recommendation interface, which includes the second performance status information corresponding to each first parameter group and the positional relationship of the performance boundary information in the coordinate system.

3. The method according to claim 1 or 2, characterized in that The first performance state information includes one or more values of each performance parameter of N performance parameters, and / or the second performance state includes one or more values of each performance parameter of N performance parameters; The N performance parameters include: at least one of central processing unit (CPU) processing performance, artificial intelligence (AI) processing performance, and image signal processing performance.

4. The method according to claim 3, characterized in that Each of the M first parameter groups corresponds to a recommended parameter, and the first parameter group optimizes the performance of the recommended parameter of the vehicle control device. The recommended parameter is one of the N performance parameters, or the recommended parameter is determined by at least two of the N performance parameters.

5. The method according to any one of claims 1 to 4, characterized in that The obtaining, by the parameter determination model and based on the first performance status information, M first parameter groups of the vehicle control device and second performance status information corresponding to each first parameter group, includes: Inputting the first performance state information into the parameter determination model to obtain K second parameter groups; Sending the K second parameter groups; Receive the M first parameter groups and second performance status information corresponding to each first parameter group, where the M first parameter groups are obtained by iteratively optimizing each second parameter group in the K second parameter groups until the performance status of the vehicle control device meets a preset performance status threshold.

6. The method according to claim 5, characterized in that The sending the K second parameter groups includes: Sorting the K second parameter groups according to the Pareto algorithm; The sorted K second parameter groups are sent.

7. The method according to claim 5 or 6, characterized in that The performance status threshold is associated with performance boundary information.

8. The method according to any one of claims 5 to 7, characterized in that Also includes: Send recommendation information of the K second parameter groups, where the recommendation information includes information of each parameter group in the K second parameter groups.

9. The method according to any one of claims 1 to 8, characterized in that Also includes: The parameter determination model is trained according to the first performance state information and the M first parameter groups to obtain an updated parameter determination model.

10. The method according to any one of claims 2 to 9, characterized in that Also includes: The performance boundary information is determined according to the historical performance status information of the vehicle control device through the Pareto algorithm.

11. A parameter recommendation device, characterized in that: include: an acquisition module, configured to acquire first performance status information of a vehicle control device and a parameter determination model corresponding to the vehicle control device; a processing module, configured to obtain, using the parameter determination model and based on the first performance status information, M first parameter groups of the vehicle control device and second performance status information corresponding to each first parameter group, wherein the second performance status information indicates a performance status of the vehicle control device when operating based on the first parameter group; The sending module is configured to send recommendation information, where the recommendation information includes second performance status information corresponding to each of the M first parameter groups.

12. The device according to claim 11, characterized in that The recommendation information also includes performance boundary information, which is determined based on the historical performance status information of the vehicle control device. The recommendation information is used to generate a recommendation interface, which includes the second performance status information corresponding to each first parameter group and the positional relationship of the performance boundary information in the coordinate system.

13. The device according to claim 11 or 12, characterized in that The first performance state information includes one or more values of each performance parameter of N performance parameters, and / or the second performance state includes one or more values of each performance parameter of N performance parameters; The N performance parameters include: at least one of central processing unit (CPU) processing performance, artificial intelligence (AI) processing performance, and image signal processing performance.

14. The device according to claim 13, characterized in that Each of the M first parameter groups corresponds to a recommended parameter, and the first parameter group optimizes the performance of the recommended parameter of the vehicle control device. The recommended parameter is one of the N performance parameters, or the recommended parameter is determined by at least two of the N performance parameters.

15. The device according to any one of claims 11 to 14, characterized in that The processing module is specifically used for: Inputting the first performance state information into the parameter determination model to obtain K second parameter groups; Sending the K second parameter groups; Receive the M first parameter groups and second performance status information corresponding to each first parameter group, where the M first parameter groups are obtained by iteratively optimizing each second parameter group in the K second parameter groups until the performance status of the vehicle control device meets a preset performance status threshold.

16. The device according to claim 15, characterized in that The processing module is specifically used for: Sorting the K second parameter groups according to the Pareto algorithm; The sorted K second parameter groups are sent.

17. The device according to claim 15 or 16, characterized in that The performance status threshold is associated with performance boundary information.

18. The device according to any one of claims 15 to 17, characterized in that The sending module is further used for: Send recommendation information of the K second parameter groups, where the recommendation information includes information of each parameter group in the K second parameter groups.

19. The device according to any one of claims 11 to 18, characterized in that The processing module is further configured to: The parameter determination model is trained according to the first performance state information and the M first parameter groups to obtain an updated parameter determination model.

20. The device according to any one of claims 12 to 19, characterized in that The processing module is further configured to: The performance boundary information is determined according to the historical performance status information of the vehicle control device through the Pareto algorithm.

21. A chip, characterized in that: The method comprises: a processor, configured to call and execute computer instructions from a memory to implement the method according to any one of claims 1 to 10.

22. An electronic device, characterized in that: The system comprises a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program to implement the method according to any one of claims 1 to 10.

23. A computer-readable storage medium, characterized in that Used to store computer program instructions, wherein the computer program causes a computer to execute the method according to any one of claims 1 to 10.

24. A computer program product, characterized in that The method comprises computer program instructions, which cause a computer to execute the method according to any one of claims 1 to 10.

Images