A computing power matching method, device and management platform

Abstract

The present invention provides a computing power matching method, device and management platform, the computing power matching method comprising: obtaining the maximum computing power consumption value of each sensor among multiple sensors, and the minimum remaining computing power value of each MEC device among multiple multi-access edge computing MEC devices; performing computing power matching between the sensor and the MEC device according to the maximum computing power consumption value of each sensor and the minimum remaining computing power value of each MEC device, and obtaining a computing power matching result; adjusting the matching relationship between the sensor and the MEC device according to the computing power matching result. In this scheme, computing power matching between the sensor and the MEC device is performed according to the maximum computing power consumption value of each sensor and the minimum remaining computing power value of each MEC device, and computing power matching can be completed. According to the computing power matching result, the matching relationship between the sensor and the MEC device is adjusted, and load balancing or fault replacement functions of unused services can be realized, thereby avoiding the problem of insufficient computing power or computing power waste.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a computing power matching method, device and management platform. Background Art

[0002] The cellular vehicle-to-everything (C-V2X) vehicle-road cooperative system is mainly composed of three subsystems: the perception subsystem, the cloud control subsystem, and the communication subsystem. Figure 1 As shown in the figure, the perception subsystem is mainly composed of sensors and Multi-Access Edge Computing (MEC). Sensors mainly include cameras, millimeter-wave radars, laser radars, etc., which are responsible for collecting raw data from traffic participants and sending the raw data to MEC. Based on the raw data, MEC completes calculations such as target recognition, event detection, and perception fusion, and sends the calculation results to the cloud control platform in the cloud control subsystem and the road side unit (RSU) in the communication subsystem. The cloud control subsystem is responsible for business data aggregation, data analysis, intelligent decision-making, business management, equipment management and other functions. The communication subsystem is responsible for broadcasting information such as safety warnings and information services, as well as reporting and sending business data. Among them, the cloud control subsystem also includes a MEC management platform for managing MEC, and the communication subsystem also includes an on-board unit (OBU) for receiving information and business data such as safety warnings and information services.

[0003] In the existing C-V2X vehicle-road cooperative system, MEC and sensors are configured in a 1:N ratio. Each MEC is configured to access multiple sensors, which are pre-planned and set based on the empirical data of MEC computing power consumption of various sensors. The MEC computing power consumed by the same type of sensor will vary greatly for different intersections and sections, and different time periods. Based on the static ratio of sensors and MEC, it is impossible to flexibly respond to such differences, which easily leads to insufficient computing power or waste of computing power. Summary of the invention

[0004] The embodiments of the present invention provide a computing power matching method, device and management platform to solve the problem of insufficient computing power or computing power waste in the prior art under a static sensor and MEC matching relationship.

[0005] In order to solve the above technical problems, the embodiments of the present invention provide the following technical solutions:

[0006] An embodiment of the present invention provides a computing power matching method, including:

[0007] Obtain the maximum computing power consumption value of each sensor among multiple sensors, and the minimum remaining computing power value of each MEC device among multiple multi-access edge computing MEC devices;

[0008] According to the maximum computing power consumption value of each of the sensors and the minimum remaining computing power value of each of the MEC devices, computing power matching is performed between the sensors and the MEC devices to obtain a computing power matching result;

[0009] According to the computing power matching result, the matching relationship between the sensor and the MEC device is adjusted.

[0010] Optionally, the maximum computing power consumption value of each sensor among the multiple sensors is obtained, including:

[0011] Periodically obtaining the number of targets identified by each of the sensors sent by the MEC device;

[0012] The maximum computing power consumption value of each sensor is determined according to the maximum number of targets identified by each sensor in multiple cycles and the predetermined correspondence between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor.

[0013] Optionally, before determining the maximum computing power consumption value of each of the sensors, the method further includes:

[0014] Periodically obtaining a computing power consumption value of each of the sensors;

[0015] A corresponding relationship between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor is established.

[0016] Optionally, establishing a corresponding relationship between the number of targets identified by each sensor and a computing power consumption value of the corresponding sensor includes:

[0017] Obtaining a computing power consumption value of a first sensor;

[0018] Establishing a corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor according to the number of targets identified by the first sensor and the computing power consumption value of the first sensor in a plurality of cycles;

[0019] Determine the computing power consumption value of the second sensor in the current period according to the correspondence between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, the number of targets identified by the first sensor in the current period, the initial computing power value of the first MEC device, and the first period computing power value of the first MEC device in the current period;

[0020] Determining a corresponding relationship between the number of targets identified by the second sensor and the computing power consumption value of the second sensor according to the number of targets identified by the second sensor and the computing power consumption value of the second sensor;

[0021] The first cycle computing power value of the first MEC device refers to the total MEC computing power consumption value when the first MEC device processes the sensor data of the first sensor and the sensor data of the second sensor according to the preset target detection algorithm in the current cycle;

[0022] The initial computing power value of the first MEC device indicates the computing power consumption value of the first MEC device in a no-load state;

[0023] The first MEC device processes the sensor data of the first sensor and the sensor data of the second sensor respectively;

[0024] The first sensor is one of each of the sensors, and the second sensor is another of each of the sensors;

[0025] The first MEC device is one of the MEC devices of each of the MEC devices.

[0026] Optionally, determining the computing power consumption value of the second sensor in the current period according to the correspondence between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, the number of targets identified by the first sensor in the current period, the initial computing power value of the first MEC device, and the first period computing power value of the first MEC device in the current period includes:

[0027] Determine the computing power consumption value of the first sensor in the current period according to the corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, and the number of targets identified by the first sensor in the current period;

[0028] The computing power consumption value of the second sensor in the current period is obtained by subtracting the computing power consumption value of the first sensor in the current period from the first period computing power value of the first MEC device in the current period and subtracting the initial computing power value of the first MEC device.

[0029] Optionally, obtaining the computing power consumption value of the first sensor includes:

[0030] When the first MEC device processes the sensor data of the first sensor, periodically obtaining a second period computing power value of the first MEC device;

[0031] Determine a computing power consumption value of the first sensor in the current period according to the second period computing power value of the first MEC device in the current period and the initial computing power value of the first MEC device;

[0032] The second cycle computing power value of the first MEC device refers to the total MEC computing power consumption value when the first MEC device processes the sensor data of the first sensor according to the preset target detection algorithm in the current cycle.

[0033] Optionally, determining the computing power consumption value of the first sensor in the current period according to the second period computing power value of the first MEC device in the current period and the initial computing power value of the first MEC device includes:

[0034] The computing power consumption value of the first sensor in the current period is obtained by subtracting the initial computing power value of the first MEC device from the second period computing power value of the first MEC device in the current period.

[0035] Optionally, establishing a correspondence between the number of targets identified by the first sensor and the computing power consumption value of the first sensor according to the number of targets identified by the first sensor in a plurality of cycles and the computing power consumption value of the first sensor includes:

[0036] Determine, according to the number of targets identified by the first sensor in different periods and the corresponding computing power consumption values, a tuple of the number of targets identified by the first sensor and the computing power consumption values ​​in different periods;

[0037] Mapping the binary pairs of the number of targets identified by the first sensor and the computing power consumption value in different periods on the coordinate system to obtain a relationship curve between the number of targets identified by the first sensor and the computing power consumption value;

[0038] A polynomial fitting is performed on the relationship curve to obtain a polynomial function of the corresponding relationship between the number of targets recognized by the first sensor and the computing power consumption value.

[0039] Optionally, obtaining the minimum remaining computing power value of each MEC device in multiple multi-access edge computing MEC devices includes:

[0040] Periodically obtaining the remaining computing power value of each of the MEC devices;

[0041] According to the remaining computing power value of each of the MEC devices, the minimum remaining computing power value of each of the MEC devices is determined.

[0042] Optionally, periodically obtaining the remaining computing power value of each MEC device includes:

[0043] Periodically obtaining a periodic computing power value of each MEC device sent by each MEC device;

[0044] According to the periodic computing power value of each of the MEC devices and the preset computing power threshold of each of the MEC devices, the remaining computing power value of each of the MEC devices in the current period is determined.

[0045] Optionally, determining the remaining computing power value of each MEC device in the current period according to the periodic computing power value of each MEC device and a preset computing power threshold of each MEC device includes:

[0046] The remaining computing power value of the corresponding MEC device in the current period is obtained by subtracting the periodic computing power value of the corresponding MEC device in the current period from the preset computing power threshold of each of the MEC devices.

[0047] Optionally, the computing power value of the MEC device is related to the first parameter;

[0048] The first parameter includes at least one of the following:

[0049] Network bandwidth;

[0050] Central processing unit CPU processing power;

[0051] Graphics processor GPU processing power;

[0052] The memory capacity of the graphics processor GPU;

[0053] Memory storage capacity;

[0054] Hard disk storage capacity;

[0055] The computing power value is one of the following:

[0056] Computing power consumption value;

[0057] Remaining computing power value;

[0058] Initial hashrate value;

[0059] Periodic computing power value.

[0060] Optionally, the measurement unit of the network bandwidth is one of the following:

[0061] Percentage; kilobits per second kbps; megabits per second Mbps; gigabit per second Gbps;

[0062] The CPU processing capability is measured in one of the following units:

[0063] Percentage; Hertz Hz; Kilohertz kHz; Megahertz MHz; Gigahertz Ghz;

[0064] The unit of measurement of the processing capability of the graphics processor GPU is one of the following:

[0065] Percentage; Hertz Hz; Kilohertz kHz; Megahertz MHz; Gigahertz Ghz;

[0066] The unit of measurement of the video memory capacity of the graphics processor GPU is one of the following:

[0067] Percentage; Kilobyte KB; Megabyte MB; Gigabyte GB; Terabyte TB;

[0068] The memory storage capacity is measured in one of the following units:

[0069] Percentage; Kilobyte KB; Megabyte MB; Gigabyte GB; Terabyte TB;

[0070] The unit of measurement for the hard disk storage capacity is one of the following:

[0071] Percentage; kilobyte KB; megabyte MB; gigabyte GB; terabyte TB; petabyte PB.

[0072] An embodiment of the present invention also provides a management platform, including a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements the steps of the computing power matching method as described above.

[0073] The embodiment of the present invention further provides a computing power matching device, including:

[0074] An acquisition module, used to acquire a maximum computing power consumption value of each of a plurality of sensors and a minimum remaining computing power value of each of a plurality of multi-access edge computing MEC devices;

[0075] A matching module, configured to perform computing power matching between the sensor and the MEC device according to the maximum computing power consumption value of each sensor and the minimum remaining computing power value of each MEC device, and obtain a computing power matching result;

[0076] An adjustment module is used to adjust the matching relationship between the sensor and the MEC device according to the computing power matching result.

[0077] Optionally, the acquisition module includes:

[0078] A first acquisition unit, configured to periodically acquire the number of targets identified by each sensor sent by the MEC device;

[0079] The first determination unit is used to determine the maximum computing power consumption value of each sensor based on the maximum number of targets recognized by each sensor in multiple cycles and the predetermined correspondence between the number of targets recognized by each sensor and the computing power consumption value of the corresponding sensor.

[0080] Optionally, the acquisition module further includes:

[0081] A second acquisition unit, used for periodically acquiring a computing power consumption value of each of the sensors;

[0082] The corresponding relationship establishing unit is used to establish a corresponding relationship between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor.

[0083] Optionally, the correspondence establishing unit is specifically configured to:

[0084] Obtaining a computing power consumption value of a first sensor;

[0085] Establishing a corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor according to the number of targets identified by the first sensor and the computing power consumption value of the first sensor in a plurality of cycles;

[0086] Determine the computing power consumption value of the second sensor in the current period according to the correspondence between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, the number of targets identified by the first sensor in the current period, the initial computing power value of the first MEC device, and the first period computing power value of the first MEC device in the current period;

[0087] Determining a corresponding relationship between the number of targets identified by the second sensor and the computing power consumption value of the second sensor according to the number of targets identified by the second sensor and the computing power consumption value of the second sensor;

[0088] The first cycle computing power value of the first MEC device refers to the total MEC computing power consumption value when the first MEC device processes the sensor data of the first sensor and the sensor data of the second sensor according to the preset target detection algorithm in the current cycle;

[0089] The initial computing power value of the first MEC device indicates the computing power consumption value of the first MEC device in a no-load state;

[0090] The first MEC device processes the sensor data of the first sensor and the sensor data of the second sensor respectively;

[0091] The first sensor is one of each of the sensors, and the second sensor is another of each of the sensors;

[0092] The first MEC device is one of the MEC devices of each of the MEC devices.

[0093] Optionally, the correspondence establishing unit is specifically configured to:

[0094] Determine the computing power consumption value of the first sensor in the current period according to the corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, and the number of targets identified by the first sensor in the current period;

[0095] The computing power consumption value of the second sensor in the current period is obtained by subtracting the computing power consumption value of the first sensor in the current period from the first period computing power value of the first MEC device in the current period and subtracting the initial computing power value of the first MEC device.

[0096] Optionally, the correspondence establishing unit is specifically configured to:

[0097] When the first MEC device processes the sensor data of the first sensor, periodically obtaining a second period computing power value of the first MEC device;

[0098] Determine a computing power consumption value of the first sensor in the current period according to the second period computing power value of the first MEC device in the current period and the initial computing power value of the first MEC device;

[0099] The second cycle computing power value of the first MEC device refers to the total MEC computing power consumption value when the first MEC device processes the sensor data of the first sensor according to the preset target detection algorithm in the current cycle.

[0100] Optionally, the correspondence establishing unit is specifically configured to:

[0101] The computing power consumption value of the first sensor in the current period is obtained by subtracting the initial computing power value of the first MEC device from the second period computing power value of the first MEC device in the current period.

[0102] Optionally, the correspondence establishing unit is specifically configured to:

[0103] Determine, according to the number of targets identified by the first sensor in different periods and the corresponding computing power consumption values, a tuple of the number of targets identified by the first sensor and the computing power consumption values ​​in different periods;

[0104] Mapping the binary pairs of the number of targets identified by the first sensor and the computing power consumption value in different periods on the coordinate system to obtain a relationship curve between the number of targets identified by the first sensor and the computing power consumption value;

[0105] A polynomial fitting is performed on the relationship curve to obtain a polynomial function of the corresponding relationship between the number of targets recognized by the first sensor and the computing power consumption value.

[0106] Optionally, obtaining the minimum remaining computing power value of each MEC device in multiple multi-access edge computing MEC devices includes:

[0107] Periodically obtaining the remaining computing power value of each of the MEC devices;

[0108] According to the remaining computing power value of each of the MEC devices, the minimum remaining computing power value of each of the MEC devices is determined.

[0109] Optionally, the acquisition module includes:

[0110] A third acquisition unit is used to periodically acquire a periodic computing power value of each MEC device sent by each MEC device;

[0111] The second determining unit is used to determine the remaining computing power value of each MEC device in the current period according to the periodic computing power value of each MEC device and the preset computing power threshold of each MEC device.

[0112] Optionally, the second determining unit is specifically configured to:

[0113] The remaining computing power value of the corresponding MEC device in the current period is obtained by subtracting the periodic computing power value of the corresponding MEC device in the current period from the preset computing power threshold of each of the MEC devices.

[0114] Optionally, the computing power value of the MEC device is related to the first parameter;

[0115] The first parameter includes at least one of the following:

[0116] Network bandwidth;

[0117] Central processing unit CPU processing power;

[0118] Graphics processor GPU processing power;

[0119] The memory capacity of the graphics processor GPU;

[0120] Memory storage capacity;

[0121] Hard disk storage capacity;

[0122] The computing power value is one of the following:

[0123] Computing power consumption value;

[0124] Remaining computing power value;

[0125] Initial hashrate value;

[0126] Periodic computing power value.

[0127] Optionally, the measurement unit of the network bandwidth is one of the following:

[0128] Percentage; kilobits per second kbps; megabits per second Mbps; gigabit per second Gbps;

[0129] The CPU processing capability is measured in one of the following units:

[0130] Percentage; Hertz Hz; Kilohertz kHz; Megahertz MHz; Gigahertz Ghz;

[0131] The unit of measurement of the processing capability of the graphics processor GPU is one of the following:

[0132] Percentage; Hertz Hz; Kilohertz kHz; Megahertz MHz; Gigahertz Ghz;

[0133] The unit of measurement of the video memory capacity of the graphics processor GPU is one of the following:

[0134] Percentage; Kilobyte KB; Megabyte MB; Gigabyte GB; Terabyte TB;

[0135] The memory storage capacity is measured in one of the following units:

[0136] Percentage; Kilobyte KB; Megabyte MB; Gigabyte GB; Terabyte TB;

[0137] The unit of measurement for the hard disk storage capacity is one of the following:

[0138] Percentage; kilobyte KB; megabyte MB; gigabyte GB; terabyte TB; petabyte PB.

[0139] An embodiment of the present invention further provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the computing power matching method as described above are implemented.

[0140] The beneficial effects of the present invention are:

[0141] The scheme of the present invention obtains the maximum computing power consumption value of each sensor among multiple sensors and the minimum remaining computing power value of each MEC device among multiple multi-access edge computing MEC devices, and performs computing power matching between sensors and MEC devices according to the maximum computing power consumption value of each sensor and the minimum remaining computing power value of each MEC device to obtain a computing power matching result, thereby completing computing power matching, and by adjusting the matching relationship between the sensor and the MEC device according to the computing power matching result, load balancing or fault replacement functions of different services can be realized, thereby avoiding the problem of insufficient computing power or waste of computing power. BRIEF DESCRIPTION OF THE DRAWINGS

[0142] Figure 1 A schematic diagram showing the structure of a C-V2X vehicle-road cooperative system provided by an embodiment of the present invention;

[0143] Figure 2 One of the flowcharts of the computing power matching method provided by an embodiment of the present invention;

[0144] Figure 3 A second flowchart showing the computing power matching method provided by an embodiment of the present invention;

[0145] Figure 4 A schematic diagram showing the structure of a MEC management platform provided in an embodiment of the present invention;

[0146] Figure 5 A schematic diagram showing a computing power matching device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0147] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0148] First, some concepts involved in the embodiments of the present invention are explained.

[0149] In the C-V2X vehicle-road cooperative system, a method for dynamically evaluating the processing capabilities of MEC and sensors (including central processing unit (CPU) processing capability, graphics processing unit (GPU) processing capability, GPU video memory processing capability, memory processing capability and other indicators, hereinafter referred to as computing power) is used to more accurately evaluate the business load capacity of MEC and the computing power requirements of sensors, providing a basis for business strategies such as business load balancing, business fault succession, and business computing power prediction.

[0150] The present invention aims to solve the problem of insufficient computing power or waste of computing power under a static sensor and MEC matching relationship in the prior art, and provides a computing power matching method, device and management platform.

[0151] like Figure 2 As shown, an embodiment of the present invention provides a computing power matching method, including:

[0152] Step 201: Obtain a maximum computing power consumption value of each sensor among a plurality of sensors, and a minimum remaining computing power value of each MEC device among a plurality of multi-access edge computing MEC devices;

[0153] It should be noted that the computing power matching method provided in the embodiment of the present invention is executed by the MEC management platform in the C-V2X vehicle-road cooperative system.

[0154] The multiple sensors in the embodiment of the present invention refer to two or more sensors.

[0155] In an embodiment of the present invention, a list of maximum computing power consumption values ​​of all sensors is formed based on the maximum computing power consumption value of each sensor among the multiple sensors obtained. The list of maximum computing power consumption values ​​includes items such as the identity document (ID) of each sensor, the MEC to which each sensor belongs, and the computing power consumption value of each sensor.

[0156] According to the minimum remaining computing power value of each MEC device among the obtained MEC devices, a minimum remaining computing power value list of all MEC devices is formed, and the minimum remaining computing power value list includes items such as the MEC ID and MEC remaining computing power value of each MEC device.

[0157] Step 202: According to the maximum computing power consumption value of each sensor and the minimum remaining computing power value of each MEC device, computing power matching is performed between the sensor and the MEC device to obtain a computing power matching result.

[0158] That is, the MEC management platform can complete computing power matching according to the above-mentioned maximum computing power consumption list and minimum remaining computing power list according to the sensor granularity, providing a basis for subsequent adjustment of the matching relationship between sensors and MEC devices.

[0159] Step 203: According to the computing power matching result, adjust the matching relationship between the sensor and the MEC device.

[0160] It should be noted that, according to the computing power matching results, the matching relationship between different sensors and different MEC devices can be adjusted, thereby realizing functions such as service load balancing or service fault alternation. Exemplarily, the maximum computing power consumption value list includes the sensor IDs and maximum computing power consumption values ​​of four sensors (sensor 1, sensor 2, sensor 3, and sensor 4), and the minimum remaining computing power value list includes the MEC IDs and minimum remaining computing power values ​​of three MEC devices (MEC 1, MEC 2, and MEC 3), wherein the maximum computing power consumption value of sensor 1 is 5%, the maximum computing power consumption value of sensor 2 is 20%, the maximum computing power consumption value of sensor 3 is 10%, the maximum computing power consumption value of sensor 4 is 40%, the minimum remaining computing power value of MEC 1 is 55%, the minimum remaining computing power value of MEC 2 is 50%, and the minimum remaining computing power value of MEC 3 is 75%. Before computing power matching, if sensor 1 is matched with MEC 1, sensor 2 and sensor 3 are matched with MEC 2 respectively, and sensor 4 is matched with MEC 3, in order to achieve business balance, according to the computing power matching result, the matching relationship between the sensors and the MEC devices is adjusted as follows: sensor 2 is matched with MEC 1, sensor 1 and sensor 3 are matched with MEC 2 respectively, and sensor 4 is matched with MEC 3 matches; if MEC1 fails, then according to the computing power matching result, sensor 1 that matches MEC 1 can be adjusted to match MEC 2, that is, the matching relationship between the sensor and the MEC device is adjusted to: sensor 1, sensor 2 and sensor 3 are matched with MEC 2 respectively, and sensor 4 is matched with MEC 3, that is, the service failure alternation function can be realized.

[0161] As a preferred implementation, obtaining the maximum computing power consumption value of each sensor among the multiple sensors includes:

[0162] Periodically obtaining the number of targets identified by each of the sensors sent by the MEC device;

[0163] The maximum computing power consumption value of each sensor is determined according to the maximum number of targets identified by each sensor in multiple cycles and the predetermined correspondence between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor.

[0164] In an embodiment of the present invention, the MEC management platform obtains and records the number of targets identified by each sensor sent by the MEC device within multiple cycles to form a historical number of identified targets. Based on the historical number of identified targets corresponding to each sensor within multiple cycles, the maximum number of targets identified by each sensor is selected, and the maximum computing power consumption value of each sensor is determined according to the predetermined correspondence between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor. Specifically, the correspondence between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor is a polynomial function. By substituting the maximum number of targets identified by each sensor into the corresponding polynomial function, the maximum computing power consumption value of each sensor can be determined to form a list of maximum computing power consumption values.

[0165] It should also be noted that the number of targets identified by each sensor is obtained and counted by the MEC device that matches the sensor according to the preset period T1, based on the sensor data sent by the sensor in the current period. Each MEC device reports the counted number of targets identified by each sensor to the MEC management platform according to the period T1.

[0166] Furthermore, before determining the maximum computing power consumption value of each of the sensors, the method further includes:

[0167] Periodically obtaining a computing power consumption value of each of the sensors;

[0168] A corresponding relationship between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor is established.

[0169] That is, the MEC management platform obtains and records the maximum computing power consumption value of each sensor in multiple cycles, determines the number and computing power consumption values ​​of the identified targets corresponding to the same sensor and in the same cycle, and establishes a corresponding relationship between the number of targets identified by the sensor and the computing power consumption value of the corresponding sensor based on the number and computing power consumption values ​​of multiple corresponding identified targets. The corresponding relationship is a polynomial function: P = f(n), where P is the computing power consumption value of the sensor and n is the number of targets identified by the sensor.

[0170] Through this polynomial function, the computing power consumption value of the corresponding sensor can be calculated when the number of targets recognized by the sensor is any value.

[0171] Optionally, establishing a corresponding relationship between the number of targets identified by each sensor and a computing power consumption value of the corresponding sensor includes:

[0172] Obtaining a computing power consumption value of a first sensor;

[0173] Establishing a corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor according to the number of targets identified by the first sensor and the computing power consumption value of the first sensor in a plurality of cycles;

[0174] Determine the computing power consumption value of the second sensor in the current period according to the correspondence between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, the number of targets identified by the first sensor in the current period, the initial computing power value of the first MEC device, and the first period computing power value of the first MEC device in the current period;

[0175] Determining a corresponding relationship between the number of targets identified by the second sensor and the computing power consumption value of the second sensor according to the number of targets identified by the second sensor and the computing power consumption value of the second sensor;

[0176] The first cycle computing power value of the first MEC device refers to the total MEC computing power consumption value when the first MEC device processes the sensor data of the first sensor and the sensor data of the second sensor according to the preset target detection algorithm in the current cycle;

[0177] The initial computing power value of the first MEC device indicates the computing power consumption value of the first MEC device in a no-load state;

[0178] The first MEC device processes the sensor data of the first sensor and the sensor data of the second sensor respectively;

[0179] The first sensor is one of each of the sensors, and the second sensor is another of each of the sensors;

[0180] The first MEC device is one of the MEC devices of each of the MEC devices.

[0181] The following specifically describes the process of establishing a corresponding relationship between the number of targets recognized by each sensor and the computing power consumption value of the corresponding sensor.

[0182] When the first MEC device (EMC 1) processes the sensor data sent by the first sensor (sensor 1) and the sensor data sent by the second sensor (sensor 2) respectively, the computing power consumption value of sensor 1 is obtained in advance, and according to the above-mentioned step of establishing the corresponding relationship between the number of targets recognized by each sensor and the computing power consumption value of the corresponding sensor, a polynomial function between the number of targets recognized by sensor 1 and the computing power consumption value of sensor 1 is established, that is, the number of targets recognized by sensor 1 in multiple cycles is obtained, the number and computing power consumption value of the recognized targets corresponding to sensor 1 in the same cycle are determined, and according to the number and computing power consumption value of the multiple corresponding recognized targets, a polynomial function between the number of targets recognized by sensor 1 and the computing power consumption value of the corresponding sensor is established. Obtain the first cycle computing power value reported by EMC 1 according to the cycle (the first cycle computing power value is the total computing power consumption value of MEC when MEC 1 processes the sensor data sent by sensor 1 and sensor 2 respectively according to the preset target detection algorithm within one cycle), and obtain the initial computing power value of MEC 1 reported by MEC 1, according to the polynomial function between the number of targets identified by sensor 1 and the computing power consumption value of the corresponding sensor, the number of targets identified by sensor 1 in the corresponding cycle, the initial computing power value of MEC 1, and the MEC 1, determine the computing power consumption value of sensor 2 in the corresponding period, calculate the computing power consumption value of sensor 2 in each period, and then establish a polynomial function between the number of targets identified by sensor 2 and the computing power consumption value of sensor 2 according to the obtained number of targets identified by sensor 2 and the above-mentioned step of establishing the corresponding relationship between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor, that is, obtain the number of targets identified by sensor 2 in multiple periods, determine the number and computing power consumption value of the identified targets corresponding to sensor 2 in the same period, and establish a polynomial function between the number of targets identified by sensor 2 and the computing power consumption value of the corresponding sensor according to the number and computing power consumption value of the multiple corresponding identified targets. Among them, the initial computing power value of MEC 1 is the computing power consumption value of MEC 1 when MEC 1 is in an unloaded state, and the initial computing power value is the same in each period.

[0183] The device type of sensor 1 and the device type of sensor 2 may be the same or different.

[0184] Specifically, the process of EMC 1 reporting the first cycle computing power value periodically is as follows: MEC 1 processes the sensor data sent by sensor 1 and sensor 2 at the same time, and MEC 1 reports the first cycle computing power value periodically, and the first cycle computing power value includes the computing power consumption value of sensor 1 and the computing power consumption value of sensor 2 in the current period.

[0185] By repeating the above steps, a corresponding relationship between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor can be established.

[0186] Further, according to the correspondence between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, the number of targets identified by the first sensor in the current cycle, the initial computing power value of the first MEC device, and the first cycle computing power value of the first MEC device in the current cycle, determining the computing power consumption value of the second sensor in the current cycle includes:

[0187] Determine the computing power consumption value of the first sensor in the current period according to the corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, and the number of targets identified by the first sensor in the current period;

[0188] The computing power consumption value of the second sensor in the current period is obtained by subtracting the computing power consumption value of the first sensor in the current period from the first period computing power value of the first MEC device in the current period and subtracting the initial computing power value of the first MEC device.

[0189] That is, according to the polynomial function between the number of targets identified by sensor 1 and the computing power consumption value of the corresponding sensor, the number of targets identified by sensor 1 in the corresponding period, the initial computing power value of MEC 1, and the computing power value of the first period of MEC 1, the process of determining the computing power consumption value of sensor 2 in the corresponding period is:

[0190] Substitute the number of targets recognized by sensor 1 corresponding to each cycle into the polynomial function between the number of targets recognized by sensor 1 and the computing power consumption value of sensor 1 to obtain the computing power consumption value of sensor 1 in the corresponding cycle, and calculate the computing power consumption value of sensor 2 in each cycle according to the formula: computing power consumption value of sensor 2 in the current cycle = computing power value of the first cycle of MEC 1 in the current cycle - computing power consumption value of sensor 1 in the current cycle - initial computing power value of MEC 1.

[0191] As a preferred implementation, obtaining the computing power consumption value of the first sensor includes:

[0192] When the first MEC device processes the sensor data of the first sensor, periodically obtaining a second period computing power value of the first MEC device;

[0193] Determine a computing power consumption value of the first sensor in the current period according to the second period computing power value of the first MEC device in the current period and the initial computing power value of the first MEC device;

[0194] The second cycle computing power value of the first MEC device refers to the total MEC computing power consumption value when the first MEC device processes the sensor data of the first sensor according to the preset target detection algorithm in the current cycle.

[0195] The process of obtaining the computing power consumption value of sensor 1 is described in detail below.

[0196] When MEC 1 only processes the sensor data of sensor 1, it receives the second cycle computing power value reported periodically by MEC 1 (the first cycle computing power value is the total computing power consumption value of MEC when MEC 1 processes the sensor data sent by sensor 1 separately according to the preset target detection algorithm within one cycle), and determines the computing power consumption value of sensor 1 in the corresponding cycle according to the second cycle computing power value of the corresponding cycle and the initial computing power value of MEC 1.

[0197] Further, determining the computing power consumption value of the first sensor in the current period according to the second period computing power value of the first MEC device in the current period and the initial computing power value of the first MEC device includes:

[0198] The computing power consumption value of the first sensor in the current period is obtained by subtracting the initial computing power value of the first MEC device from the second period computing power value of the first MEC device in the current period.

[0199] That is, according to the second period computing power value of the corresponding period and the initial computing power value of MEC 1, the process of determining the computing power consumption value of sensor 1 in the corresponding period is:

[0200] According to the formula: the computing power consumption value of sensor 1 in the current cycle = the second cycle computing power value of MEC 1 in the current cycle - the initial computing power value of MEC 1, the computing power consumption value of sensor 1 in each cycle is calculated.

[0201] As a preferred implementation, establishing a corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor according to the number of targets identified by the first sensor in multiple cycles and the computing power consumption value of the first sensor includes:

[0202] Determine, according to the number of targets identified by the first sensor in different periods and the corresponding computing power consumption values, a tuple of the number of targets identified by the first sensor and the computing power consumption values ​​in different periods;

[0203] Mapping the binary pairs of the number of targets identified by the first sensor and the computing power consumption value in different periods on the coordinate system to obtain a relationship curve between the number of targets identified by the first sensor and the computing power consumption value;

[0204] A polynomial fitting is performed on the relationship curve to obtain a polynomial function of the corresponding relationship between the number of targets recognized by the first sensor and the computing power consumption value.

[0205] Specifically, the process by which the MEC management platform establishes a functional relationship between the number of targets identified by sensor 1 and the computing power consumption value of MEC 1 is as follows:

[0206] According to the number of targets recognized by the sensor 1 in multiple corresponding periods and the corresponding computing power consumption values, a pair of the number of targets recognized by the sensor 1 and the computing power consumption values ​​corresponding to different periods <number of recognized targets, computing power consumption value> is formed, and the pair of <number of recognized targets, computing power consumption value> with different values ​​is mapped on the cross coordinate, where the horizontal axis is the number of recognized targets and the vertical axis is the computing power consumption value. A curve reflecting the relationship between the number of targets recognized by the sensor 1 and the computing power consumption value can be obtained, and the relationship curve is fitted to form a polynomial function.

[0207] As a preferred implementation, obtaining the minimum remaining computing power value of each MEC device in a plurality of multi-access edge computing MEC devices includes:

[0208] Periodically obtaining the remaining computing power value of each of the MEC devices;

[0209] According to the remaining computing power value of each of the MEC devices, the minimum remaining computing power value of each of the MEC devices is determined.

[0210] In an embodiment of the present invention, the MEC management platform selects the minimum value of the remaining computing power of each MEC device, that is, the minimum remaining computing power value of each MEC device, based on the calculated remaining computing power value of the MEC device in each cycle, to form a minimum remaining computing power value list.

[0211] Furthermore, the remaining computing power value of each MEC device is periodically obtained, including:

[0212] Periodically obtaining a periodic computing power value of each MEC device sent by each MEC device;

[0213] According to the periodic computing power value of each of the MEC devices and the preset computing power threshold of each of the MEC devices, the remaining computing power value of each of the MEC devices in the current period is determined.

[0214] The periodic computing power value of each MEC device is the total computing power consumption value of the MEC when each MEC device processes the sensor data of the corresponding sensor according to the preset target detection algorithm in the current period.

[0215] The following takes the first MEC device (MEC 1) as an example to illustrate the process of periodically obtaining the remaining computing power value of each MEC device.

[0216] When MEC 1 only processes the sensor data of sensor 1, sensor 1 sends the sensor data to MEC 1, MEC 1 processes the sensor data of sensor 1, starts the target algorithm, the computing power changes, MEC 1 records the periodic computing power value (first periodic computing power value) periodically, and MEC 1 reports the periodic computing power value periodically; when MEC 1 processes the sensor data of sensor 1 and sensor 2 at the same time, sensor 1 and sensor 2 send the sensor data value MEC 1 at the same time, MEC 1 processes the sensor data of sensor 1 and sensor 2 at the same time, starts the target algorithm, the computing power changes, MEC 1 records the periodic computing power value (second periodic computing power value) periodically, and MEC 1 reports the periodic computing power value periodically.

[0217] Since the computing power value of the MEC device is related to multiple computing power indicators such as network bandwidth, central processing unit CPU processing power, graphics processing unit GPU processing power, graphics processing unit GPU memory capacity, memory storage capacity and hard disk storage capacity, the MEC management platform sets a computing power threshold for MEC 1 based on the system operation strategy, wherein multiple computing power indicators such as network bandwidth, central processing unit CPU processing power, graphics processing unit GPU processing power, graphics processing unit GPU memory capacity, memory storage capacity and hard disk storage capacity are set separately, and the computing power threshold is the upper limit of the computing power utilization rate to ensure the stable operation of the system. For example, the computing power threshold of the CPU processing power is set to 85%, which means that the CPU utilization rate of MEC 1 cannot exceed 85%, and the excess part is the system redundant computing power, which cannot be allocated and used by the MEC management platform. The MEC management platform calculates the remaining computing power value of MEC 1 according to the periodic computing power value of MEC 1 and the preset computing power threshold of MEC1.

[0218] Repeat the above steps to obtain the remaining computing power value of each MEC device periodically.

[0219] Further, according to the periodic computing power value of each MEC device and the preset computing power threshold of each MEC device, determining the remaining computing power value of each MEC device in the current period includes:

[0220] The remaining computing power value of the corresponding MEC device in the current period is obtained by subtracting the periodic computing power value of the corresponding MEC device in the current period from the preset computing power threshold of each of the MEC devices.

[0221] That is, the MEC management platform calculates the remaining computing power value of MEC 1 periodically based on the periodic computing power value of MEC 1 and the preset computing power threshold of MEC 1. The specific process is:

[0222] According to the formula: the remaining computing power value of MEC 1 in the current cycle = the computing power threshold of MEC 1 - the periodic computing power value of MEC 1 in the current cycle, the remaining computing power value of MEC 1 in each cycle is calculated.

[0223] Optionally, the computing power value of the MEC device is related to the first parameter;

[0224] The first parameter includes at least one of the following:

[0225] Network bandwidth;

[0226] Central processing unit CPU processing power;

[0227] Graphics processor GPU processing power;

[0228] The memory capacity of the graphics processor GPU;

[0229] Memory storage capacity;

[0230] Hard disk storage capacity;

[0231] The computing power value is one of the following:

[0232] Computing power consumption value;

[0233] Remaining computing power value;

[0234] Initial hashrate value;

[0235] Periodic computing power value.

[0236] It should be noted that the comprehensive computing power of MEC is based on parameters such as network bandwidth, CPU processing power, GPU processing power, GPU memory capacity, memory storage capacity and hard disk storage capacity. Among them, the comprehensive computing power value includes computing power consumption value, remaining computing power value, initial computing power value and periodic computing power value.

[0237] Optionally, the measurement unit of the network bandwidth is one of the following:

[0238] Percentage; kilobits per second kbps; megabits per second Mbps; gigabit per second Gbps;

[0239] The CPU processing capability is measured in one of the following units:

[0240] Percentage; Hertz Hz; Kilohertz kHz; Megahertz MHz; Gigahertz Ghz;

[0241] The unit of measurement of the processing capability of the graphics processor GPU is one of the following:

[0242] Percentage; Hertz Hz; Kilohertz kHz; Megahertz MHz; Gigahertz Ghz;

[0243] The unit of measurement of the video memory capacity of the graphics processor GPU is one of the following:

[0244] Percentage; Kilobyte KB; Megabyte MB; Gigabyte GB; Terabyte TB;

[0245] The memory storage capacity is measured in one of the following units:

[0246] Percentage; Kilobyte KB; Megabyte MB; Gigabyte GB; Terabyte TB;

[0247] The unit of measurement for the hard disk storage capacity is one of the following:

[0248] Percentage; kilobyte KB; megabyte MB; gigabyte GB; terabyte TB; petabyte PB.

[0249] Exemplarily, the initial computing power value of each MEC is an MEC initial computing power value table generated by the MEC management platform by aggregating the initial computing power values ​​reported by each MEC. The MEC initial computing power value table includes indicators such as MEC ID, CPU occupancy, GPU occupancy, GPU video memory occupancy, and memory occupancy. The MEC initial computing power value table is shown in Table 1 below.

[0250] Table 1 MEC initial computing power value table

[0251]

[0252] It should be noted that the measurement units of the CPU occupancy, GPU occupancy, GPU video memory occupancy, and memory occupancy in Table 1 above are all percentages, and can also correspond to other measurement units.

[0253] The maximum computing power consumption list also includes indicators such as sensor ID, MEC ID, CPU occupancy, GPU occupancy, GPU memory occupancy, and memory occupancy. The maximum computing power consumption list is shown in Table 2 below.

[0254] Table 2 Maximum computing power consumption value list

[0255]

[0256] It should be noted that the measurement units of the CPU consumption value, GPU consumption value, GPU video memory consumption value, and memory consumption value in Table 2 above are all percentages, and can also correspond to other measurement units.

[0257] The minimum remaining computing power value list also includes indicators such as MEC ID, CPU occupancy, GPU occupancy, GPU memory occupancy, and memory occupancy. The minimum remaining computing power value list is shown in Table 3 below.

[0258] Table 3 Minimum remaining computing power value list

[0259]

[0260] It should be noted that the measurement units of the CPU remaining computing power value, GPU remaining computing power value, GPU video memory remaining computing power value, and memory remaining computing power value in Table 3 above are all percentages, and can also be converted to other units of measurement.

[0261] Combine the following Figure 3 , specifically illustrate the computing power matching method provided by an embodiment of the present invention.

[0262] Step 1: MEC 1 collects various computing power values ​​that reflect its own computing power. Step 2: MEC 1 reports the collected computing power values. Step 3: Generate the MEC initial computing power set based on the computing power values ​​collected by each MEC. The MEC initial computing power value set includes multiple computing power value indicators such as CPU occupancy, GPU occupancy, GPU video memory occupancy, and memory occupancy. The MEC initial computing power value set is used to reflect the current EMC computing power load. Step 4: Sensor 1 sends raw data (sensor data) to MEC1. Step 5: MEC 1 processes the raw data sent by sensor 1, starts the target detection algorithm, and the computing power changes. MEC1 records its own computing power set (first cycle computing power value) according to period T1. Step 6: MEC 1 reports its own computing power set recorded by MEC 1 to the MEC management platform according to period T1. Step 7: MEC 1 counts the number of targets identified by sensor 1 in the current period according to period T1; the more targets identified, the more computing power the algorithm consumes, and the higher the values ​​of each indicator item of the current MEC computing power set. Step 8: MEC 1 reports the number of targets identified by sensor 1 to the MEC management platform according to period T1. Step 9: The MEC management platform calculates the set of computing power consumption data values ​​of sensor 1 (hereinafter referred to as computing power consumption set) according to the period, that is, calculates the computing power requirement set of sensor 1. The computing power consumption value set of the sensor includes computing power consumption values ​​of multiple dimensions such as CPU occupancy, GPU occupancy, GPU video memory occupancy, and memory occupancy. Each type of computing power consumption value needs to be calculated separately. According to the following calculation formula, the CPU computing power consumption value, GPU computing power consumption value, GPU video memory computing power consumption value, and memory computing power consumption value are obtained respectively, and finally the computing power consumption set of sensor 1 is obtained. The computing power consumption value of the sensor in the current period = the computing power value of the first period of the current period - the initial computing power value of MEC 1. Step 10: On the MEC management platform, a binary tuple <number of identified targets, computing power consumption value> is formed based on the computing power consumption set of sensor 1 calculated in step 9 and the number of targets identified by sensor 1 corresponding to the computing power consumption period (reported by step 8). The binary tuple <number of identified targets, computing power consumption value> with different values ​​is mapped in a cross coordinate system, with the horizontal axis being the number of identified targets and the vertical axis being the computing power consumption value. A curve reflecting the computing power consumption value (defined as P) and the number of identified targets (defined as n) can be obtained. A polynomial is fitted to the curve to form a polynomial function of P = f(n). This polynomial function can be used to calculate the computing power consumption value of sensor 1 when the number of identified targets is any value. Step 11: Sensor 1 and sensor 2 send raw data (sensor data) to MEC 1 at the same time. At this time, MEC 1 reports the periodic computing power value (second period computing power value) periodically. The second period computing power value includes both the computing power consumption value of sensor 1 and the computing power consumption value of sensor 2. The MEC management platform completes the computing power consumption value of sensor 2 and the computing power consumption set calculation periodically.The calculation formula is: the computing power consumption value of sensor 2 in the current cycle = the computing power value of the second cycle of the current cycle - the computing power consumption value of sensor 1 in the current cycle - the initial computing power value of MEC 1. The computing power consumption value of sensor 1 in the current cycle is calculated by the MEC management platform based on the number of targets identified by sensor 1 in the current cycle and the polynomial function P = f(n). Step 12: The MEC management platform sets a computing power threshold for MEC 1 based on the system operation strategy, where the CPU occupancy rate, GPU occupancy rate, GPU memory occupancy rate, and memory occupancy rate are set separately. The computing power threshold is the upper limit of the computing power usage rate set to ensure the stable operation of the system. The MEC management platform calculates the set of MEC remaining computing power values ​​(MEC remaining computing power set) on a periodic basis. The MEC remaining computing power set includes multiple computing power value indicators such as CPU occupancy rate, GPU occupancy rate, GPU memory occupancy rate, and memory occupancy rate. Each computing power indicator needs to be calculated separately, and finally the MEC remaining computing power set is obtained. MEC remaining computing power value = MEC computing power threshold - the periodic computing power value of the current cycle. Step 13: MEC 2 joins the MEC management platform, and the MEC management platform repeats steps 4 to 12 to complete the calculation of the computing power consumption set of each sensor of MEC 2 and the remaining computing power set of MEC 2. Step 14: The MEC management platform selects the maximum number of recognized targets of each sensor based on the historical data of the number of targets recognized by the sensor, and uses the polynomial function P = f(n) to calculate the maximum computing power consumption value of each sensor, and generates a list of the maximum computing power consumption values ​​of all sensors. Step 15: Based on the MEC remaining computing power value calculated in each cycle, the MEC management platform selects the minimum remaining computing power value of each MEC as the minimum remaining computing power value of the MEC, and finally generates a list of the minimum remaining computing power values ​​of all MECs. Step 16: The MEC management platform completes computing power matching according to the maximum computing power consumption value of each sensor and the minimum remaining computing power value of each MEC, and adjusts the matching relationship between the sensor and the MEC according to the sensor granularity. According to different adjustment strategies, functions such as service load balancing and service failure succession are realized.

[0263] The computing power matching method provided in the embodiment of the present invention can more finely adjust the configuration relationship between the sensor and the MEC through quantitative calculation of the sensor computing power consumption value and the MEC remaining computing power value, realize business load balancing and smooth succession of business failures, and can also provide computing power prediction for newly launched businesses.

[0264] like Figure 4As shown, an embodiment of the present invention further provides a management platform, which is a MEC management platform, including a processor 400, a transceiver 410, a memory 420, and a program stored in the memory 420 and executable on the processor 400; wherein the transceiver 410 is connected to the processor 400 and the memory 420 through a bus interface, wherein the processor 400 performs the following process for reading the program in the memory:

[0265] Obtain the maximum computing power consumption value of each sensor among multiple sensors, and the minimum remaining computing power value of each MEC device among multiple multi-access edge computing MEC devices;

[0266] According to the maximum computing power consumption value of each of the sensors and the minimum remaining computing power value of each of the MEC devices, computing power matching is performed between the sensors and the MEC devices to obtain a computing power matching result;

[0267] According to the computing power matching result, the matching relationship between the sensor and the MEC device is adjusted.

[0268] Optionally, the processor 400 is specifically configured to:

[0269] Periodically obtaining the number of targets identified by each of the sensors sent by the MEC device;

[0270] The maximum computing power consumption value of each sensor is determined according to the maximum number of targets identified by each sensor in multiple cycles and the predetermined correspondence between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor.

[0271] Optionally, the processor 400 is further configured to:

[0272] Periodically obtaining a computing power consumption value of each of the sensors;

[0273] A corresponding relationship between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor is established.

[0274] Optionally, the processor 400 is specifically configured to:

[0275] Obtaining a computing power consumption value of a first sensor;

[0276] Establishing a corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor according to the number of targets identified by the first sensor and the computing power consumption value of the first sensor in a plurality of cycles;

[0277] Determine the computing power consumption value of the second sensor in the current period according to the correspondence between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, the number of targets identified by the first sensor in the current period, the initial computing power value of the first MEC device, and the first period computing power value of the first MEC device in the current period;

[0278] Determining a corresponding relationship between the number of targets identified by the second sensor and the computing power consumption value of the second sensor according to the number of targets identified by the second sensor and the computing power consumption value of the second sensor;

[0279] The first cycle computing power value of the first MEC device refers to the total MEC computing power consumption value when the first MEC device processes the sensor data of the first sensor and the sensor data of the second sensor according to the preset target detection algorithm in the current cycle;

[0280] The initial computing power value of the first MEC device indicates the computing power consumption value of the first MEC device in a no-load state;

[0281] The first MEC device processes the sensor data of the first sensor and the sensor data of the second sensor respectively;

[0282] The first sensor is one of each of the sensors, and the second sensor is another of each of the sensors;

[0283] The first MEC device is one of the MEC devices of each of the MEC devices.

[0284] Optionally, the processor 400 is specifically configured to:

[0285] Determine the computing power consumption value of the first sensor in the current period according to the corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, and the number of targets identified by the first sensor in the current period;

[0286] The computing power consumption value of the second sensor in the current period is obtained by subtracting the computing power consumption value of the first sensor in the current period from the first period computing power value of the first MEC device in the current period and subtracting the initial computing power value of the first MEC device.

[0287] Optionally, the processor 400 is specifically configured to:

[0288] When the first MEC device processes the sensor data of the first sensor, periodically obtaining a second period computing power value of the first MEC device;

[0289] Determine a computing power consumption value of the first sensor in the current period according to the second period computing power value of the first MEC device in the current period and the initial computing power value of the first MEC device;

[0290] The second cycle computing power value of the first MEC device refers to the total MEC computing power consumption value when the first MEC device processes the sensor data of the first sensor according to the preset target detection algorithm in the current cycle.

[0291] Optionally, the processor 400 is specifically configured to:

[0292] The computing power consumption value of the first sensor in the current period is obtained by subtracting the initial computing power value of the first MEC device from the second period computing power value of the first MEC device in the current period.

[0293] Optionally, the processor 400 is specifically configured to:

[0294] Determine, according to the number of targets identified by the first sensor in different periods and the corresponding computing power consumption values, a tuple of the number of targets identified by the first sensor and the computing power consumption values ​​in different periods;

[0295] Mapping the binary pairs of the number of targets identified by the first sensor and the computing power consumption value in different periods on the coordinate system to obtain a relationship curve between the number of targets identified by the first sensor and the computing power consumption value;

[0296] A polynomial fitting is performed on the relationship curve to obtain a polynomial function of the corresponding relationship between the number of targets recognized by the first sensor and the computing power consumption value.

[0297] Optionally, the processor 400 is specifically configured to:

[0298] Periodically obtaining the remaining computing power value of each of the MEC devices;

[0299] According to the remaining computing power value of each of the MEC devices, the minimum remaining computing power value of each of the MEC devices is determined.

[0300] Optionally, the processor 400 is specifically configured to:

[0301] Periodically obtaining a periodic computing power value of each MEC device sent by each MEC device;

[0302] According to the periodic computing power value of each of the MEC devices and the preset computing power threshold of each of the MEC devices, the remaining computing power value of each of the MEC devices in the current period is determined.

[0303] Optionally, the processor 400 is specifically configured to:

[0304] The remaining computing power value of the corresponding MEC device in the current period is obtained by subtracting the periodic computing power value of the corresponding MEC device in the current period from the preset computing power threshold of each of the MEC devices.

[0305] Optionally, the computing power value of the MEC device is related to the first parameter;

[0306] The first parameter includes at least one of the following:

[0307] Network bandwidth;

[0308] Central processing unit CPU processing power;

[0309] Graphics processor GPU processing power;

[0310] The memory capacity of the graphics processor GPU;

[0311] Memory storage capacity;

[0312] Hard disk storage capacity;

[0313] The computing power value is one of the following:

[0314] Computing power consumption value;

[0315] Remaining computing power value;

[0316] Initial hashrate value;

[0317] Periodic computing power value.

[0318] Optionally, the measurement unit of the network bandwidth is one of the following:

[0319] Percentage; kilobits per second kbps; megabits per second Mbps; gigabit per second Gbps;

[0320] The CPU processing capability is measured in one of the following units:

[0321] Percentage; Hertz Hz; Kilohertz kHz; Megahertz MHz; Gigahertz Ghz;

[0322] The unit of measurement of the processing capability of the graphics processor GPU is one of the following:

[0323] Percentage; Hertz Hz; Kilohertz kHz; Megahertz MHz; Gigahertz Ghz;

[0324] The unit of measurement of the video memory capacity of the graphics processor GPU is one of the following:

[0325] Percentage; Kilobyte KB; Megabyte MB; Gigabyte GB; Terabyte TB;

[0326] The memory storage capacity is measured in one of the following units:

[0327] Percentage; Kilobyte KB; Megabyte MB; Gigabyte GB; Terabyte TB;

[0328] The unit of measurement for the hard disk storage capacity is one of the following:

[0329] Percentage; kilobyte KB; megabyte MB; gigabyte GB; terabyte TB; petabyte PB.

[0330] Among them, Figure 4 In the embodiment, the bus architecture may include any number of interconnected buses and bridges, specifically one or more processors represented by processor 400 and various circuits of memory represented by memory 420 are linked together. The bus architecture may also link together various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface. The transceiver 410 may be a plurality of components, namely, a transmitter and a receiver, providing a unit for communicating with various other devices on a transmission medium, including transmission media such as wireless channels, wired channels, and optical cables. For different user devices, the user interface 430 may also be an interface capable of externally and internally connecting required devices, and the connected devices include but are not limited to a keypad, a display, a speaker, a microphone, a joystick, and the like.

[0331] The processor 400 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or a complex programmable logic device (CPLD). The processor may also adopt a multi-core architecture.

[0332] like Figure 5 As shown, an embodiment of the present invention further provides a computing power matching device, including:

[0333] An acquisition module 501 is used to acquire a maximum computing power consumption value of each of a plurality of sensors and a minimum remaining computing power value of each of a plurality of multi-access edge computing MEC devices;

[0334] A matching module 502 is used to perform computing power matching between the sensor and the MEC device according to the maximum computing power consumption value of each sensor and the minimum remaining computing power value of each MEC device to obtain a computing power matching result;

[0335] The adjustment module 503 is used to adjust the matching relationship between the sensor and the MEC device according to the computing power matching result.

[0336] In an embodiment of the present invention, by obtaining the maximum computing power consumption value of each sensor among multiple sensors and the minimum remaining computing power value of each MEC device among multiple multi-access edge computing MEC devices, computing power matching between sensors and MEC devices is performed according to the maximum computing power consumption value of each sensor and the minimum remaining computing power value of each MEC device to obtain a computing power matching result, thereby completing computing power matching, and by adjusting the matching relationship between the sensor and the MEC device according to the computing power matching result, load balancing or fault replacement functions of different services can be realized, thereby avoiding the problem of insufficient computing power or waste of computing power.

[0337] Optionally, the acquisition module 501 includes:

[0338] A first acquisition unit, configured to periodically acquire the number of targets identified by each sensor sent by the MEC device;

[0339] The first determination unit is used to determine the maximum computing power consumption value of each sensor based on the maximum number of targets recognized by each sensor in multiple cycles and the predetermined correspondence between the number of targets recognized by each sensor and the computing power consumption value of the corresponding sensor.

[0340] Optionally, the acquisition module 501 further includes:

[0341] A second acquisition unit, used for periodically acquiring a computing power consumption value of each of the sensors;

[0342] The corresponding relationship establishing unit is used to establish a corresponding relationship between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor.

[0343] Optionally, the correspondence establishing unit is specifically configured to:

[0344] Obtaining a computing power consumption value of a first sensor;

[0345] Establishing a corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor according to the number of targets identified by the first sensor and the computing power consumption value of the first sensor in a plurality of cycles;

[0346] Determine the computing power consumption value of the second sensor in the current period according to the correspondence between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, the number of targets identified by the first sensor in the current period, the initial computing power value of the first MEC device, and the first period computing power value of the first MEC device in the current period;

[0347] Determining a corresponding relationship between the number of targets identified by the second sensor and the computing power consumption value of the second sensor according to the number of targets identified by the second sensor and the computing power consumption value of the second sensor;

[0348] The first cycle computing power value of the first MEC device refers to the total MEC computing power consumption value when the first MEC device processes the sensor data of the first sensor and the sensor data of the second sensor according to the preset target detection algorithm in the current cycle;

[0349] The initial computing power value of the first MEC device indicates the computing power consumption value of the first MEC device in a no-load state;

[0350] The first MEC device processes the sensor data of the first sensor and the sensor data of the second sensor respectively;

[0351] The first sensor is one of each of the sensors, and the second sensor is another of each of the sensors;

[0352] The first MEC device is one of the MEC devices of each of the MEC devices.

[0353] Optionally, the correspondence establishing unit is specifically configured to:

[0354] Determine the computing power consumption value of the first sensor in the current period according to the corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, and the number of targets identified by the first sensor in the current period;

[0355] The computing power consumption value of the second sensor in the current period is obtained by subtracting the computing power consumption value of the first sensor in the current period from the first period computing power value of the first MEC device in the current period and subtracting the initial computing power value of the first MEC device.

[0356] Optionally, the correspondence establishing unit is specifically configured to:

[0357] When the first MEC device processes the sensor data of the first sensor, periodically obtaining a second period computing power value of the first MEC device;

[0358] Determine a computing power consumption value of the first sensor in the current period according to the second period computing power value of the first MEC device in the current period and the initial computing power value of the first MEC device;

[0359] The second cycle computing power value of the first MEC device refers to the total MEC computing power consumption value when the first MEC device processes the sensor data of the first sensor according to the preset target detection algorithm in the current cycle.

[0360] Optionally, the correspondence establishing unit is specifically configured to:

[0361] The computing power consumption value of the first sensor in the current period is obtained by subtracting the initial computing power value of the first MEC device from the second period computing power value of the first MEC device in the current period.

[0362] Optionally, the correspondence establishing unit is specifically configured to:

[0363] Determine, according to the number of targets identified by the first sensor in different periods and the corresponding computing power consumption values, a tuple of the number of targets identified by the first sensor and the computing power consumption values ​​in different periods;

[0364] Mapping the binary pairs of the number of targets identified by the first sensor and the computing power consumption value in different periods on the coordinate system to obtain a relationship curve between the number of targets identified by the first sensor and the computing power consumption value;

[0365] A polynomial fitting is performed on the relationship curve to obtain a polynomial function of the corresponding relationship between the number of targets recognized by the first sensor and the computing power consumption value.

[0366] Optionally, obtaining the minimum remaining computing power value of each MEC device in multiple multi-access edge computing MEC devices includes:

[0367] Periodically obtaining the remaining computing power value of each of the MEC devices;

[0368] According to the remaining computing power value of each of the MEC devices, the minimum remaining computing power value of each of the MEC devices is determined.

[0369] Optionally, the acquisition module 501 includes:

[0370] A third acquisition unit is used to periodically acquire a periodic computing power value of each MEC device sent by each MEC device;

[0371] The second determining unit is used to determine the remaining computing power value of each MEC device in the current period according to the periodic computing power value of each MEC device and the preset computing power threshold of each MEC device.

[0372] Optionally, the second determining unit is specifically configured to:

[0373] The remaining computing power value of the corresponding MEC device in the current period is obtained by subtracting the periodic computing power value of the corresponding MEC device in the current period from the preset computing power threshold of each of the MEC devices.

[0374] Optionally, the computing power value of the MEC device is related to the first parameter;

[0375] The first parameter includes at least one of the following:

[0376] Network bandwidth;

[0377] Central processing unit CPU processing power;

[0378] Graphics processor GPU processing power;

[0379] The memory capacity of the graphics processor GPU;

[0380] Memory storage capacity;

[0381] Hard disk storage capacity;

[0382] The computing power value is one of the following:

[0383] Computing power consumption value;

[0384] Remaining computing power value;

[0385] Initial hashrate value;

[0386] Periodic computing power value.

[0387] Optionally, the measurement unit of the network bandwidth is one of the following:

[0388] Percentage; kilobits per second kbps; megabits per second Mbps; gigabit per second Gbps;

[0389] The CPU processing capability is measured in one of the following units:

[0390] Percentage; Hertz Hz; Kilohertz kHz; Megahertz MHz; Gigahertz Ghz;

[0391] The unit of measurement of the processing capability of the graphics processor GPU is one of the following:

[0392] Percentage; Hertz Hz; Kilohertz kHz; Megahertz MHz; Gigahertz Ghz;

[0393] The unit of measurement of the video memory capacity of the graphics processor GPU is one of the following:

[0394] Percentage; Kilobyte KB; Megabyte MB; Gigabyte GB; Terabyte TB;

[0395] The memory storage capacity is measured in one of the following units:

[0396] Percentage; Kilobyte KB; Megabyte MB; Gigabyte GB; Terabyte TB;

[0397] The unit of measurement for the hard disk storage capacity is one of the following:

[0398] Percentage; kilobyte KB; megabyte MB; gigabyte GB; terabyte TB; petabyte PB.

[0399] It should be noted that the computing power matching device provided in the embodiment of the present invention is a device capable of executing the above-mentioned computing power matching method. All embodiments of the above-mentioned computing power matching method are applicable to the device and can achieve the same or similar technical effects.

[0400] An embodiment of the present invention further provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the computing power matching method as described above are implemented.

[0401] The readable storage medium can be any available medium or data storage device that can be accessed by the processor, including but not limited to magnetic storage (such as floppy disks, hard disks, magnetic tapes, magneto-optical disks (MO), etc.), optical storage (such as CD, DVD, BD, HVD, etc.), and semiconductor storage (such as ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid-state drive (SSD)), etc.

[0402] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) that contain computer-usable program code.

[0403] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer executable instructions. These computer executable instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0404] These processor executable instructions may also be stored in a processor readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the processor readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0405] These processor-executable instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0406] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary personnel in this technical field, several improvements and modifications can be made without departing from the principles described in the present invention. These improvements and modifications are also within the scope of protection of the present invention.

Claims

1. A computing power matching method, characterized in that: include: Obtain the maximum computing power consumption value of each sensor among multiple sensors, and the minimum remaining computing power value of each MEC device among multiple multi-access edge computing MEC devices; According to the maximum computing power consumption value of each of the sensors and the minimum remaining computing power value of each of the MEC devices, computing power matching is performed between the sensors and the MEC devices to obtain a computing power matching result; According to the computing power matching result, adjusting the matching relationship between the sensor and the MEC device; The maximum computing power consumption value of each sensor among multiple sensors is obtained, including: Periodically obtaining the number of targets identified by each of the sensors sent by the MEC device; The maximum computing power consumption value of each sensor is determined according to the maximum number of targets identified by each sensor in multiple cycles and the predetermined correspondence between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor.

2. The computing power matching method according to claim 1, characterized in that: Before determining the maximum computing power consumption value of each of the sensors, the method further includes: Periodically obtaining a computing power consumption value of each of the sensors; A corresponding relationship between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor is established.

3. The computing power matching method according to claim 2, characterized in that: Establishing a corresponding relationship between the number of targets identified by each sensor and the computing power consumption value of the corresponding sensor includes: Obtaining a computing power consumption value of a first sensor; Establishing a corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor according to the number of targets identified by the first sensor and the computing power consumption value of the first sensor in a plurality of cycles; Determine the computing power consumption value of the second sensor in the current period according to the correspondence between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, the number of targets identified by the first sensor in the current period, the initial computing power value of the first MEC device, and the first period computing power value of the first MEC device in the current period; Determining a corresponding relationship between the number of targets identified by the second sensor and the computing power consumption value of the second sensor according to the number of targets identified by the second sensor and the computing power consumption value of the second sensor; The first cycle computing power value of the first MEC device refers to the total MEC computing power consumption value when the first MEC device processes the sensor data of the first sensor and the sensor data of the second sensor according to the preset target detection algorithm in the current cycle; The initial computing power value of the first MEC device indicates the computing power consumption value of the first MEC device in a no-load state; The first MEC device processes the sensor data of the first sensor and the sensor data of the second sensor respectively; The first sensor is one of each of the sensors, and the second sensor is another of each of the sensors; The first MEC device is one of the MEC devices of each of the MEC devices.

4. The computing power matching method according to claim 3, characterized in that: Determining the computing power consumption value of the second sensor in the current period according to the correspondence between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, the number of targets identified by the first sensor in the current period, the initial computing power value of the first MEC device, and the first period computing power value of the first MEC device in the current period, including: Determine the computing power consumption value of the first sensor in the current period according to the corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor, and the number of targets identified by the first sensor in the current period; The computing power consumption value of the second sensor in the current period is obtained by subtracting the computing power consumption value of the first sensor in the current period from the first period computing power value of the first MEC device in the current period and subtracting the initial computing power value of the first MEC device.

5. The computing power matching method according to claim 3, characterized in that: Obtain the computing power consumption value of the first sensor, including: When the first MEC device processes the sensor data of the first sensor, periodically obtaining a second period computing power value of the first MEC device; Determine a computing power consumption value of the first sensor in the current period according to the second period computing power value of the first MEC device in the current period and the initial computing power value of the first MEC device; The second cycle computing power value of the first MEC device refers to the total MEC computing power consumption value when the first MEC device processes the sensor data of the first sensor according to the preset target detection algorithm in the current cycle.

6. The computing power matching method according to claim 5, characterized in that: Determining a computing power consumption value of the first sensor in the current period according to a second period computing power value of the first MEC device in the current period and an initial computing power value of the first MEC device includes: The computing power consumption value of the first sensor in the current period is obtained by subtracting the initial computing power value of the first MEC device from the second period computing power value of the first MEC device in the current period.

7. The computing power matching method according to claim 3, characterized in that: Establishing a corresponding relationship between the number of targets identified by the first sensor and the computing power consumption value of the first sensor according to the number of targets identified by the first sensor in a plurality of cycles and the computing power consumption value of the first sensor, including: Determine, according to the number of targets identified by the first sensor in different periods and the corresponding computing power consumption values, a tuple of the number of targets identified by the first sensor and the computing power consumption values ​​in different periods; Mapping the binary pairs of the number of targets identified by the first sensor and the computing power consumption value in different periods on the coordinate system to obtain a relationship curve between the number of targets identified by the first sensor and the computing power consumption value; A polynomial fitting is performed on the relationship curve to obtain a polynomial function of the corresponding relationship between the number of targets recognized by the first sensor and the computing power consumption value.

8. The computing power matching method according to claim 1, characterized in that: Get the minimum remaining computing power value of each MEC device among multiple multi-access edge computing MEC devices, including: Periodically obtaining the remaining computing power value of each of the MEC devices; According to the remaining computing power value of each of the MEC devices, the minimum remaining computing power value of each of the MEC devices is determined.

9. The computing power matching method according to claim 8, characterized in that: Periodically obtaining the remaining computing power value of each MEC device, including: Periodically obtaining a periodic computing power value of each MEC device sent by each MEC device; According to the periodic computing power value of each of the MEC devices and the preset computing power threshold of each of the MEC devices, the remaining computing power value of each of the MEC devices in the current period is determined.

10. The computing power matching method according to claim 9, characterized in that: Determining the remaining computing power value of each MEC device in the current cycle according to the periodic computing power value of each MEC device and the preset computing power threshold of each MEC device includes: The remaining computing power value of the corresponding MEC device in the current period is obtained by subtracting the periodic computing power value of the corresponding MEC device in the current period from the preset computing power threshold of each of the MEC devices.

11. The computing power matching method according to claim 1, characterized in that: The computing power value of the MEC device is related to the first parameter; The first parameter includes at least one of the following: Network bandwidth; Central processing unit CPU processing power; Graphics processor GPU processing power; The memory capacity of the graphics processor GPU; Memory storage capacity; Hard disk storage capacity; The computing power value is one of the following: Computing power consumption value; Remaining computing power value; Initial hashrate value; Periodic computing power value.

12. The computing power matching method according to claim 11, characterized in that: The network bandwidth is measured in one of the following units: Percentage; kilobits per second kbps; megabits per second Mbps; gigabit per second Gbps; The CPU processing capability is measured in one of the following units: Percentage; Hertz Hz; Kilohertz kHz; Megahertz MHz; Gigahertz Ghz; The unit of measurement of the processing capability of the graphics processor GPU is one of the following: Percentage; Hertz Hz; Kilohertz kHz; Megahertz MHz; Gigahertz Ghz; The unit of measurement of the video memory capacity of the graphics processor GPU is one of the following: Percentage; Kilobyte KB; Megabyte MB; Gigabyte GB; Terabyte TB; The memory storage capacity is measured in one of the following units: Percentage; Kilobyte KB; Megabyte MB; Gigabyte GB; Terabyte TB; The unit of measurement for the hard disk storage capacity is one of the following: Percentage; kilobyte KB; megabyte MB; gigabyte GB; terabyte TB; petabyte PB.

13. A management platform, characterized in that: It includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements the steps of the computing power matching method as described in any one of claims 1 to 12.

14. A computing power matching device, characterized in that: include: An acquisition module, used to acquire a maximum computing power consumption value of each of a plurality of sensors and a minimum remaining computing power value of each of a plurality of multi-access edge computing MEC devices; A matching module, configured to perform computing power matching between the sensor and the MEC device according to the maximum computing power consumption value of each sensor and the minimum remaining computing power value of each MEC device, and obtain a computing power matching result; An adjustment module, used to adjust the matching relationship between the sensor and the MEC device according to the computing power matching result; Wherein, the acquisition module includes: A first acquisition unit, configured to periodically acquire the number of targets identified by each sensor sent by the MEC device; The first determination unit is used to determine the maximum computing power consumption value of each sensor based on the maximum number of targets recognized by each sensor in multiple cycles and the predetermined correspondence between the number of targets recognized by each sensor and the computing power consumption value of the corresponding sensor.

15. A readable storage medium, characterized in that: The readable storage medium stores a program or instruction, and when the program or instruction is executed by the processor, the steps of the computing power matching method as described in any one of claims 1 to 12 are implemented.

Images