Edge-cloud cooperative positioning method and system based on composite sensor

Abstract

The invention belongs to the technical field of satellite positioning chips, and particularly relates to an edge-cloud cooperative positioning method and system based on a composite sensor, and the method comprises the steps: firstly obtaining a positioning task state, a power storage state and signal intensity through a positioning chip, carrying out the transmission discrimination of locally stored batch data through a configured storage-transmission regulation and control model, and carrying out the positioning of the positioning task state, the power storage state and the signal intensity; edge-cloud data transmission is realized; secondly, according to the current positioning task information, the calculation complexity of the task is determined through a calculation complexity model; thirdly, an edge-cloud cooperative positioning model is configured, and edge-cloud cooperative positioning judgment is carried out according to the calculation complexity of a positioning task, the chip power storage state and the signal intensity; and finally, according to a judgment result, selecting to carry out positioning calculation at the edge or cooperatively allocate calculation resources between the edge and the cloud so as to obtain high-precision positioning information. By optimizing data transmission and computing resource allocation, the response speed and energy consumption efficiency of the positioning system are remarkably improved, and the method is suitable for positioning application in various complex environments.

Description

Technical Field

[0001] The present invention belongs to the technical field of satellite positioning chips, and in particular relates to an edge-cloud collaborative positioning method and system based on a composite sensor. Background Art

[0002] Traditional positioning devices have significant deficiencies in coverage, real-time positioning capabilities, and performance in weak signal areas; these devices usually have narrow coverage and cannot provide comprehensive positioning services in remote or weak signal areas, resulting in extremely low cost-effectiveness in a variety of application scenarios; for example, in the fields of outdoor adventures, agricultural monitoring, and intelligent transportation, traditional positioning devices often cannot meet actual needs, especially in areas with poor signal coverage, where their positioning accuracy and reliability are greatly reduced. In addition, traditional positioning devices also have defects in data transmission and processing; the high power consumption caused by real-time data transmission makes the device life short, and the need for frequent charging increases maintenance costs; at the same time, the data processing capacity is limited, making it difficult to meet the needs of modern Internet of Things and intelligent applications for big data processing and analysis; existing solutions often rely on a single sensor and simple data processing methods, lacking comprehensive analysis and intelligent processing capabilities for multi-source data; therefore, positioning accuracy and reliability in complex environments are difficult to guarantee, limiting its wide application in high-precision positioning and intelligent applications.

[0003] For example, a Chinese patent application with publication number CN117761732A discloses a satellite navigation positioning and orientation chip and a signal processing method. The satellite navigation positioning device includes: a first-level functional chip, a second-level functional chip, a third-level functional chip and a substrate; the first-level functional chip is arranged on the substrate, and is used to realize full-frequency satellite navigation positioning and orientation; the second-level functional chip is arranged above the first-level functional chip, and is used to run a satellite navigation and positioning program; the third-level functional chip is arranged above the second-level functional chip, and is used to store the satellite navigation and positioning program.

[0004] The above existing technologies have the following problems: traditional positioning devices generally have narrow coverage, cannot be positioned in real time, and cannot fully cover remote areas or areas with weak signals. For this reason, the present invention provides an edge-cloud collaborative positioning method and system based on composite sensors. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention proposes an edge-cloud collaborative positioning method and system based on a composite sensor. The method first obtains the positioning task status data, power storage status and signal strength data through the positioning chip and stores them locally. At the same time, the configured storage-transmission control model is used to transmit and judge the batch data stored locally, and the edge-cloud data transmission is realized through the full network access module; secondly, according to the current positioning task information, the computational complexity of the task is determined by the computational complexity model; thirdly, the edge-cloud collaborative positioning model is configured, the edge positioning sub-model is built into the chip, and the collaborative positioning model is deployed to the cloud. According to the computational complexity of the positioning task, the chip power storage status and the signal strength, the edge-cloud collaborative positioning judgment is performed; finally, according to the judgment result, it is selected to perform positioning calculation at the edge or to collaboratively allocate computing resources between the edge and the cloud to obtain high-precision positioning information; this method significantly improves the response speed and energy efficiency of the positioning system by optimizing data transmission and computing resource allocation, and is suitable for positioning applications in various complex environments.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The edge-cloud collaborative positioning method based on composite sensors includes the following steps:

[0008] S1. Obtain the positioning task status data, power storage status and signal strength data of the positioning chip and store them locally on the chip. At the same time, the configured storage-transmission control model is used to identify the local storage batch data transmission, and the configured full network access module is used to perform edge-to-cloud data transmission.

[0009] S2. Obtain the current positioning task information, and obtain the computational complexity of the corresponding positioning task through the configured computational complexity model;

[0010] S3. Configure the edge-cloud collaborative positioning model, build the edge positioning sub-model in the edge-cloud collaborative positioning model into the chip, deploy the collaborative positioning sub-model to the cloud, and make edge-cloud collaborative positioning decisions based on the computational complexity of the corresponding positioning task, the chip power storage status, and the signal strength.

[0011] S4. Perform edge positioning calculation or edge-cloud collaborative allocation positioning calculation on the positioning task according to the determination result to obtain positioning information of the corresponding positioning task.

[0012] Specifically, the steps for constructing the storage-transmission regulation model in S1 include:

[0013] S101, obtaining the positioning task type and result time point, the size and timestamp of the positioning acquisition data packet corresponding to each positioning task, the current signal strength, the data transmission rate per unit time and the chip SOC state quantity at the current moment, and preprocessing the obtained data;

[0014] S102: Configure the priority model to obtain the positioning task processing priority according to the positioning task type, result time point, and positioning acquisition data packet timestamp. And sort the obtained task priorities in descending order; where d i represents the positioning task processing priority corresponding to the i-th positioning task, T e represents the result time point corresponding to the i-th positioning task, T c Indicates the timestamp of the positioning data packet corresponding to the i-th positioning task;

[0015] S103, according to the current signal strength, the data transmission rate per unit time and the network packet loss rate per unit time period, the network signal evaluation index corresponding to each moment is calculated by a support vector machine;

[0016] S104, constructing a signal transmission-power consumption curve using the acquired network signal evaluation index and chip SOC state quantity, obtaining a network signal evaluation index value corresponding to the minimum value of signal transmission power consumption as the data batch transmission signal index threshold value [q 1 …q k ], where q k Indicates the data batch transmission signal indicator threshold in the kth time period.

[0017] Specifically, the steps of constructing the storage-transmission regulation model in S1 also include:

[0018] S105, obtaining the size of the positioning data packet of the corresponding type of positioning task, the required computing resource data, the remaining available computing resources of the chip at the corresponding time, the computing complexity of each task and the local positioning task computing delay, and inputting the obtained data into the computing complexity model constructed by the support vector machine for training, to obtain a trained computing complexity model;

[0019] S106. Deploy the trained computational complexity model into the local database of the chip, input the positioning acquisition data packet size, the required computing resource data and the remaining available computing resources of the chip at the current moment of each priority type positioning task obtained in S102 into the computational complexity model, and obtain the computational complexity and computational delay of the corresponding positioning task.

[0020] Specifically, the steps of constructing the storage-transmission regulation model in S1 also include:

[0021] S107: construct the current input state of the storage-transmission control model according to the acquired positioning task processing priority, positioning acquisition data packet size, current signal strength, unit time data transmission rate and current chip SOC state quantity. Among them, soct Indicates the current storage capacity of the positioning chip. represents the computational complexity of the i-th positioning task at the current moment, represents the computational delay of the i-th positioning task at the current moment, q t Indicates the network signal evaluation index corresponding to the current moment, represents the positioning calculation accuracy of the jth local positioning task at the current moment, soc 0 Indicates the minimum working power storage capacity threshold corresponding to the positioning chip;

[0022] S108, according to the current input state, setting the current execution action trigger information, including:

[0023] At the current moment q t =q k And soc t >soc 0 , the first execution action trigger information is triggered;

[0024] In the state of triggering the first execution action trigger information, the corresponding subordinate execution action trigger information includes:

[0025] S1081, when d i Greater than or equal to the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is greater than or equal to the positioning calculation accuracy threshold, the second execution action trigger information is triggered;

[0026] S1082, when d i Greater than or equal to the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is less than the positioning calculation accuracy threshold, the third execution action trigger information is triggered.

[0027] Specifically, setting the trigger information for executing the action at the current moment also includes:

[0028] S1083, when d i Greater than or equal to the configured task processing priority threshold, Greater than or equal to the maximum complexity threshold of the locally executable calculation or is greater than the configured computational latency threshold and When it is less than the positioning calculation accuracy threshold, the fourth execution action trigger information is triggered;

[0029] S1084, when d iLess than the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is greater than or equal to the positioning calculation accuracy threshold, the fifth execution action trigger information is triggered;

[0030] S1085, when d i Less than the configured task processing priority threshold, Greater than or equal to the maximum complexity threshold of the locally executable calculation or When it is less than the configured calculation delay threshold, the sixth execution action trigger information is triggered;

[0031] S1086, setting a maximum throughput threshold corresponding to the network signal evaluation index, when the size of the positioning acquisition data packet set to be uploaded for the positioning task at the current moment is equal to the maximum throughput threshold corresponding to the network signal evaluation index, uploading the corresponding batch task to the cloud, and performing positioning calculation by calling the collaborative positioning sub-model;

[0032] Specifically, setting the trigger information for executing the action at the current moment also includes:

[0033] At the current moment q t ≠q k And soc t >soc 0 , the seventh execution action trigger information is triggered;

[0034] In the state of triggering the seventh execution action triggering information, the corresponding subordinate execution action triggering information includes:

[0035] S1087, when d i Greater than or equal to the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is greater than or equal to the positioning calculation accuracy threshold, the second execution action trigger information is triggered;

[0036] S1088, when d i Greater than or equal to the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is less than the positioning calculation accuracy threshold, the eighth execution action trigger information is triggered;

[0037] S1089, when d i Less than or equal to the configured task processing priority threshold, if When it is greater than or equal to the positioning calculation accuracy threshold, the fifth execution action trigger information is triggered. When it is less than the positioning calculation accuracy threshold, the eighth execution action trigger information is triggered;

[0038] S1090, when the current moment soc t ≤soc 0 When the second execution action trigger information is triggered, a low battery warning is issued.

[0039] Specifically, the calculation steps of performing edge positioning calculation or edge-cloud collaborative allocation positioning calculation on the positioning task according to the determination result in S4 include:

[0040] S401, real-time monitoring and collection of positioning data packets corresponding to the positioning task, performing initial positioning calculations through the deployed edge positioning sub-model, and directly feeding back the corresponding results to the corresponding user if the corresponding positioning accuracy meets the positioning calculation accuracy threshold;

[0041] S402. When the corresponding positioning accuracy does not meet the positioning calculation accuracy threshold and the corresponding data is uploaded to the cloud, the collaborative positioning sub-model is called to perform secondary calculation based on the positioning acquisition data packet of the corresponding positioning task and the initial positioning calculation result of the edge positioning sub-model to obtain the secondary positioning task calculation result.

[0042] The edge-cloud collaborative positioning system based on composite sensors includes: a transmission discrimination module, a complexity module and a discrimination calculation module;

[0043] A transmission discrimination module is used for data acquisition and data transmission discrimination; the transmission discrimination module includes a data acquisition unit and a transmission discrimination unit;

[0044] A data acquisition unit, used to acquire the positioning task status data, power storage status and signal strength data of the positioning chip and perform pre-processing and local chip storage;

[0045] The transmission identification unit is used to identify the local storage batch data transmission through the configured storage-transmission control model, and to perform edge-to-cloud data transmission through the configured full-network module;

[0046] The complexity module is used to obtain the computational complexity of the corresponding positioning task through the configured computational complexity model based on the current positioning task information.

[0047] Specifically, the discriminant computing module includes a deployment unit, a computation discriminant unit, and an edge-to-cloud computing unit;

[0048] A deployment unit, used to build an edge positioning sub-model in the configured edge-cloud co-positioning model into the chip, and deploy the co-positioning sub-model to the cloud;

[0049] A calculation and judgment unit is used to perform edge positioning calculation and edge-cloud collaborative positioning calculation judgment according to the calculation complexity of the corresponding positioning task, the chip power storage state and the signal strength, and obtain the judgment result;

[0050] The edge-cloud computing unit is used to perform edge positioning calculation or edge-cloud collaborative positioning calculation according to the judgment result of the calculation judgment unit and the corresponding collected positioning task status data, through the deployed edge positioning sub-model or collaborative positioning sub-model, to obtain the positioning result of the corresponding positioning task.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] In view of the deficiencies in the prior art, the present invention uses the status data of the locally stored positioning chip and the configuration storage-transmission control model to perform data transmission judgment and batch data transmission of the full-network module in real time according to the chip status, further reducing the chip's power loss under low power and weak signal conditions, and improving the chip's endurance and adaptability to different transmission networks; secondly, by utilizing the computational complexity model and the edge-cloud collaborative positioning model, the method can dynamically adjust the resource allocation of data calculation according to task requirements and environmental conditions, significantly improving the response speed and energy efficiency of the positioning system; especially in remote or signal-weak areas, through the edge positioning sub-model, computational complexity and locally stored positioning data in the edge-cloud collaborative positioning model, real-time low-complexity and low-power positioning services can be achieved, and support can be provided for subsequent in-depth analysis of large amounts of data; in addition, when the signal is good, the locally stored data can be quickly uploaded to the cloud for high-precision positioning, thereby improving the accuracy of positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a flow chart of the edge-cloud collaborative positioning method based on composite sensors according to Embodiment 1 of the present invention;

[0054] Figure 2 This is a chip body structure diagram of Example 1 of the present invention;

[0055] Figure 3 This is a module diagram of the edge-cloud collaborative positioning system based on composite sensors according to Example 2 of the present invention. DETAILED DESCRIPTION

[0056] Example 1

[0057] See also Figure 1 , an embodiment provided by the present invention: an edge-cloud collaborative positioning method based on a composite sensor, the steps comprising:

[0058] S1. Obtain the positioning task status data, power storage status and signal strength data of the positioning chip and store them locally on the chip. At the same time, the configured storage-transmission control model is used to identify the local storage batch data transmission, and the configured full network access module is used to perform edge-to-cloud data transmission.

[0059] Furthermore, the positioning chip of this embodiment is as follows Figure 2 As shown;

[0060] Furthermore, in this embodiment, the positioning chip housing is provided with an FPC antenna, a Beidou III positioning antenna and a CAT1 communication antenna for collecting and transmitting positioning data; by configuring different antennas, the positioning chip can support global navigation satellite system positioning and has the capability of single-frequency and 64-channel simultaneous tracking;

[0061] Furthermore, this embodiment adopts the NMEA 4.1 protocol when performing data transmission to ensure compatibility with other devices and achieve seamless connection and data exchange;

[0062] Furthermore, the full network access module in this embodiment is a communication module configured on the chip body, which can adapt to the network environment of different operators and ensure the communication connection of the device worldwide;

[0063] S2. Obtain current positioning task information, and obtain the computational complexity of the corresponding positioning task through the configured computational complexity model; further, in this embodiment, the computational complexity model is constructed by a linear function;

[0064] S3. Configure the edge-cloud collaborative positioning model, build the edge positioning sub-model in the edge-cloud collaborative positioning model into the chip, deploy the collaborative positioning sub-model to the cloud, and make edge-cloud collaborative positioning decisions based on the computational complexity of the corresponding positioning task, the chip power storage status, and the signal strength.

[0065] S4. Perform edge positioning calculation or edge-cloud collaborative allocation positioning calculation on the positioning task according to the determination result to obtain positioning information of the corresponding positioning task.

[0066] This process significantly improves the performance and efficiency of the positioning system by optimizing data management and computing resource allocation. First, by obtaining the positioning task status data, power storage status and signal strength data of the positioning chip and storing them locally, the system can intelligently determine when to perform batch data transmission, reducing unnecessary network transmission, reducing energy consumption and improving the reliability of data transmission. Secondly, through the configured computational complexity model (built based on linear functions), the computational complexity of each positioning task can be accurately evaluated, thus providing a scientific basis for subsequent task allocation. Furthermore, the edge-cloud collaborative positioning model builds the edge positioning sub-model into the chip and deploys the collaborative positioning sub-model to the cloud, making intelligent judgments based on computational complexity, chip power storage status and signal strength, and realizing reasonable task allocation. This flexible task allocation mechanism not only reduces the computing burden of a single device, but also extends battery life. In addition, with the support of the full network access module, the system can adapt to the network environment of different operators and ensure the stability of communication connections worldwide.

[0067] Furthermore, the steps of constructing the storage-transmission regulation model in S1 in this embodiment include:

[0068] S101, obtaining the positioning task type and result time point, the size and timestamp of the positioning acquisition data packet corresponding to each positioning task, the current signal strength, the data transmission rate per unit time and the chip SOC state quantity at the current moment, and preprocessing the obtained data;

[0069] S102: Configure the priority model to obtain the positioning task processing priority according to the positioning task type, result time point, and positioning acquisition data packet timestamp. And sort the obtained task priorities in descending order; where d i represents the positioning task processing priority corresponding to the i-th positioning task, T e represents the result time point corresponding to the i-th positioning task, T c Indicates the timestamp of the positioning data packet corresponding to the i-th positioning task;

[0070] S103, according to the current signal strength, the data transmission rate per unit time and the network packet loss rate per unit time period, the network signal evaluation index corresponding to each moment is calculated by a support vector machine;

[0071] S104, constructing a signal transmission-power consumption curve using the acquired network signal evaluation index and chip SOC state quantity, obtaining a network signal evaluation index value corresponding to the minimum value of signal transmission power consumption as the data batch transmission signal index threshold value [q 1 …q k ], where q kIndicates the data batch transmission signal indicator threshold in the kth time period;

[0072] S105, obtaining the size of the positioning data packet of the corresponding type of positioning task, the required computing resource data, the remaining available computing resources of the chip at the corresponding time, the computing complexity of each task and the local positioning task computing delay, and inputting the obtained data into the computing complexity model constructed by the support vector machine for training, to obtain a trained computing complexity model;

[0073] S106, deploying the trained computational complexity model to the local database of the chip, inputting the positioning acquisition data packet size, the required computing resource amount data and the remaining available computing resources of the chip at the current moment of each type of positioning task corresponding to each priority obtained in S102 into the computational complexity model, and obtaining the computational complexity and computational delay of the corresponding positioning task;

[0074] S107: construct the current input state of the storage-transmission control model according to the acquired positioning task processing priority, positioning acquisition data packet size, current signal strength, unit time data transmission rate and current chip SOC state quantity. Among them, soc t Indicates the current storage capacity of the positioning chip. represents the computational complexity of the i-th positioning task at the current moment, represents the computational delay of the i-th positioning task at the current moment, q t Indicates the network signal evaluation index corresponding to the current moment, represents the positioning calculation accuracy of the jth local positioning task at the current moment, soc 0 Indicates the minimum working power storage capacity threshold corresponding to the positioning chip;

[0075] S108, according to the current input state, setting the current execution action trigger information, including:

[0076] At the current moment q t =q k And soc t >soc 0 , the first execution action trigger information is triggered;

[0077] In the state of triggering the first execution action trigger information, the corresponding subordinate execution action trigger information includes:

[0078] S1081, when d i Greater than or equal to the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is greater than or equal to the positioning calculation accuracy threshold, the second execution action trigger information is triggered, that is, the i-th positioning task data is processed in the local chip;

[0079] S1082, when d i Greater than or equal to the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is less than the positioning calculation accuracy threshold, the third execution action trigger information is triggered;

[0080] That is, for d i If the task processing priority threshold is greater than or equal to the configured one, the edge positioning sub-model is used to perform initial positioning calculations, and the initial positioning results are fed back to the corresponding user. At the same time, the positioning data packet information and initial positioning results corresponding to the i-th positioning task are uploaded to the cloud, and the collaborative positioning sub-model is called to perform secondary positioning calculations.

[0081] S1083, when d i Greater than or equal to the configured task processing priority threshold, Greater than or equal to the maximum complexity threshold of the locally executable calculation or is greater than the configured computational latency threshold and When it is less than the positioning calculation accuracy threshold, the fourth execution action trigger information is triggered; that is, the i-th positioning task is uploaded to the cloud for positioning calculation;

[0082] S1084, when d i Less than the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is greater than or equal to the positioning calculation accuracy threshold, the fifth execution action trigger information is triggered;

[0083] That is, repeatedly triggering the second execution action triggering information, and performing the local positioning calculation according to the task processing priority order of all positioning tasks for performing the local positioning calculation;

[0084] S1085, when d i Less than the configured task processing priority threshold, Greater than or equal to the maximum complexity threshold of the locally executable calculation or When it is less than the configured calculation delay threshold, the sixth execution action trigger information is triggered, that is, the i-th positioning task is directly uploaded to the cloud for positioning calculation;

[0085] S1086, setting a maximum throughput threshold corresponding to the network signal evaluation index, when the size of the positioning acquisition data packet set to be uploaded for the positioning task at the current moment is equal to the maximum throughput threshold corresponding to the network signal evaluation index, uploading the corresponding batch task to the cloud, and performing positioning calculation by calling the collaborative positioning sub-model;

[0086] At the current moment q t ≠q k And soc t >soc 0 , the seventh execution action trigger information is triggered;

[0087] In the state of triggering the seventh execution action triggering information, the corresponding subordinate execution action triggering information includes:

[0088] S1087, when d i Greater than or equal to the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is greater than or equal to the positioning calculation accuracy threshold, the second execution action trigger information is triggered;

[0089] S1088, when d i Greater than or equal to the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is less than the positioning calculation accuracy threshold, the eighth execution action trigger information is triggered;

[0090] That is, firstly, the initial positioning calculation is performed through the edge positioning sub-model, and then the feasible transmission interval of the local network signal evaluation index in the corresponding time period is obtained through the signal transmission-power consumption curve, and the network signal evaluation index corresponding to the feasible transmission interval and the maximum data throughput of the corresponding signal index are used to fit the batch transmission fitting function through the support vector machine, and the batch transmission fitting function is built into the storage-transmission control model, and the network signal evaluation index and the size of the positioning acquisition data packet set of the positioning task to be uploaded at the corresponding moment are monitored in real time. When the size of the positioning acquisition data packet set of the positioning task to be uploaded is equal to the maximum throughput corresponding to the network signal evaluation index, the batch positioning task is uploaded, and the positioning calculation is performed by calling the collaborative positioning sub-model. Further, in this embodiment, the maximum data throughput of the corresponding signal index is obtained by the Iperf tool; Iperf is a widely used network performance testing tool that can measure the bandwidth performance of TCP and UDP, thereby obtaining the network throughput. When using it, the corresponding commands need to be run at the sending end and the receiving end respectively, and the throughput under different conditions is tested by setting different parameters, such as transmission time, data packet size, etc., to determine the maximum throughput.

[0091] Furthermore, in this embodiment, the feasible transmission interval of the local network signal evaluation index is an interval in which the network signal evaluation index value corresponding to data transmission can be constructed;

[0092] S1089, when d i Less than or equal to the configured task processing priority threshold, if When it is greater than or equal to the positioning calculation accuracy threshold, the fifth execution action trigger information is triggered. When the value is less than the positioning calculation accuracy threshold, the eighth execution action trigger information is triggered, and the process S1086 is repeated to upload batch data;

[0093] S1090, when the current moment soc t ≤soc 0 When the battery is low, the second execution action trigger information is triggered and a low battery warning is issued;

[0094] S1091, constructing the current moment execution action based on the current moment execution action trigger information, and inputting the current moment input state, the current moment execution action trigger information and the current moment execution action into the storage-transmission control model constructed by the SAC model for training, obtaining the trained storage-transmission control model, and deploying the trained model to the chip body storage database interface to perform data batch transmission judgment and transmission execution;

[0095] This process significantly improves the performance and efficiency of the positioning system through refined data management and intelligent task allocation. First, by obtaining the positioning task type, result time point, data packet size and timestamp, signal strength, data transmission rate and chip SOC state quantity, and performing preprocessing, the system can accurately determine the priority and processing order of the data. The configured priority model sorts the tasks according to the task type and timestamp to ensure that high-priority tasks are processed in a timely manner. The support vector machine (SVM) is used to calculate the network signal evaluation index, and the signal transmission-power consumption curve is constructed in combination with the chip SOC state quantity to determine the optimal transmission time, thereby reducing energy consumption and improving transmission efficiency. Through the trained computational complexity model, the system can accurately evaluate the computational complexity and delay of each task and realize intelligent task allocation. The edge-cloud collaborative positioning model dynamically determines whether the task is processed at the edge or in the cloud according to the task priority, computational complexity, delay and computational accuracy, thereby optimizing resource utilization. In addition, by real-time monitoring of network signal evaluation indicators and data packet size, the system can batch upload tasks at the best time, further improving the efficiency and reliability of data transmission. The low battery warning mechanism ensures the stable operation of the system.

[0096] Furthermore, in this embodiment, the calculation steps of performing edge positioning calculation or edge-cloud collaborative allocation positioning calculation on the positioning task according to the determination result include:

[0097] S401, real-time monitoring and collection of positioning data packets corresponding to the positioning task, performing initial positioning calculations through the deployed edge positioning sub-model, and directly feeding back the corresponding results to the corresponding user if the corresponding positioning accuracy meets the positioning calculation accuracy threshold;

[0098] S402. When the corresponding positioning accuracy does not meet the positioning calculation accuracy threshold and the corresponding data is uploaded to the cloud, the collaborative positioning sub-model is called to perform secondary calculation based on the positioning acquisition data packet of the corresponding positioning task and the initial positioning calculation result of the edge positioning sub-model to obtain the secondary positioning task calculation result.

[0099] This process significantly improves the response speed and efficiency of the positioning system by real-time monitoring of the collected data packets of the positioning task and using the edge positioning sub-model for initial positioning calculations. When the positioning accuracy of the edge calculation meets the threshold, the results are directly fed back to the user, reducing data transmission delays and the occupation of cloud computing resources. If the edge computing accuracy is insufficient, the data is uploaded to the cloud and the collaborative positioning sub-model is called for secondary calculations to ensure the high accuracy of the final positioning result. This edge-cloud collaborative computing mechanism not only improves positioning accuracy, but also optimizes resource allocation and reduces energy consumption.

[0100] Example 2

[0101] See also Figure 3,Another embodiment provided by the present invention: An edge-cloud collaborative positioning system based on a composite sensor, comprising: a transmission discrimination module, a complexity module, and a discrimination calculation module;

[0102] A transmission discrimination module is used for data acquisition and data transmission discrimination; the transmission discrimination module includes a data acquisition unit and a transmission discrimination unit;

[0103] A data acquisition unit, used to acquire the positioning task status data, power storage status and signal strength data of the positioning chip and perform pre-processing and local chip storage;

[0104] The transmission identification unit is used to identify the local storage batch data transmission through the configured storage-transmission control model, and to perform edge-to-cloud data transmission through the configured full-network module;

[0105] The complexity module is used to obtain the computational complexity of the corresponding positioning task through the configured computational complexity model according to the current positioning task information;

[0106] The discriminant computing module is used for the deployment of the model and the judgment of the edge-cloud computing of the positioning task; the discriminant computing module includes a deployment unit, a computing discriminant unit and an edge-cloud computing unit;

[0107] A deployment unit is used to build the edge positioning sub-model in the configured edge-cloud collaborative positioning model into the chip, and deploy the collaborative positioning sub-model to the cloud; a calculation and judgment unit is used to perform edge positioning calculation and edge-cloud collaborative positioning calculation judgment according to the calculation complexity of the corresponding positioning task, the chip power storage state and the signal strength, and obtain the judgment result;

[0108] The edge-cloud computing unit is used to perform edge positioning calculation or edge-cloud collaborative positioning calculation according to the judgment result of the calculation judgment unit and the corresponding collected positioning task status data, through the deployed edge positioning sub-model or collaborative positioning sub-model, to obtain the positioning result of the corresponding positioning task.

[0109] Example 3

[0110] Another embodiment provided by the present invention: an edge-cloud collaborative positioning device based on a composite sensor, comprising: a chip body, an FPC antenna on the chip body, a Beidou III positioning antenna and a CAT1 communication antenna, and a chip SOC module configured on the chip body, an SD card slot, a mechanical switch, a Type C interface, and LA780EG; the FPC antenna, a Beidou III positioning antenna and a CAT1 communication antenna are all in a ceramic antenna module;

[0111] The chip body first uses FPC (flexible printed circuit) soft board, PCB (printed circuit board), and minimizes the layout of the circuit board to reduce redundant components. Secondly, the chip body uses SMT (surface mount technology) to weld all components, reducing the cables and connectors in the assembly process and reducing the risk of failure.

[0112] The chip SOC (system on chip) module uses a vertical stacking method to stack power, positioning, and CAT1 communication on one module, making the chip body highly integrated, reducing the number of devices and reducing power consumption; CAT1 is a communication standard in the LTE network, providing a lower data transmission rate and is suitable for IoT applications that do not require high bandwidth;

[0113] The chip body adopts BGA (ball grid array) packaging and automotive-grade design, which optimizes the heat dissipation, stress and reliability of the chip after it is reduced in size. The power supply part of the chip body adopts a 3.7V gallium arsenide flexible battery, which can work for a long time and provide part of the battery life while being charged.

[0114] The comprehensive application of the above-mentioned chip components not only makes the chip body highly integrated and miniaturized, but also improves reliability and performance; for example, FPC and SMT reduce redundant devices and cables, reducing the risk of failure; the vertical stacking design of the SOC module further reduces the number of devices and reduces power consumption; BGA packaging and automotive-grade design optimize heat dissipation and stress management.

[0115] Example 4

[0116] An electronic device includes a memory and a processor, wherein the memory stores a computer program, and the processor implements an edge-cloud collaborative positioning method based on a composite sensor when executing the computer program.

[0117] A computer-readable storage medium stores computer instructions, which, when executed, execute an edge-cloud collaborative positioning method based on a composite sensor.

[0118] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the scope of protection of the purpose of the present invention and the claims, and all of these are within the protection of the present invention.

[0119] If the disclosed technical solution involves personal information, the product using the disclosed technical solution has clearly informed the personal information processing rules and obtained the individual's voluntary consent before processing the personal information. If the disclosed technical solution involves sensitive personal information, the product using the disclosed technical solution has obtained the individual's separate consent before processing the sensitive personal information, and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, clear and prominent signs are set to inform that the personal information collection scope has been entered and personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that he or she agrees to the collection of his or her personal information; or on the device that processes personal information, when the personal information processing rules are notified by obvious signs / information, the individual's authorization is obtained through pop-up information or by asking the individual to upload his or her personal information; among them, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

Claims

1. An edge-cloud collaborative positioning method based on composite sensors, characterized in that the steps include: S1. Obtain the positioning task status data, power storage status and signal strength data of the positioning chip and store them locally on the chip. At the same time, the configured storage-transmission control model is used to identify the local storage batch data transmission, and the configured full network access module is used to perform edge-to-cloud data transmission. S2. Obtain the current positioning task information, and obtain the computational complexity of the corresponding positioning task through the configured computational complexity model; S3. Configure the edge-cloud collaborative positioning model, build the edge positioning sub-model in the edge-cloud collaborative positioning model into the chip, deploy the collaborative positioning sub-model to the cloud, and make edge-cloud collaborative positioning decisions based on the computational complexity of the corresponding positioning task, the chip power storage status, and the signal strength. S4. Perform edge positioning calculation or edge-cloud collaborative allocation positioning calculation on the positioning task according to the determination result to obtain positioning information of the corresponding positioning task.

2. The edge-cloud collaborative positioning method based on composite sensors according to claim 1, characterized in that: The steps of constructing the storage-transmission regulation model in S1 include: S101, obtaining the positioning task type and result time point, the size and timestamp of the positioning acquisition data packet corresponding to each positioning task, the current signal strength, the data transmission rate per unit time and the chip SOC state quantity at the current moment, and preprocessing the obtained data; S102: Configure the priority model to obtain the positioning task processing priority according to the positioning task type, result time point, and positioning acquisition data packet timestamp. And the obtained task priorities are arranged in descending order; where di represents the positioning task processing priority corresponding to the i-th positioning task, Te represents the result time point corresponding to the i-th positioning task, and Tc represents the positioning acquisition data packet timestamp corresponding to the i-th positioning task; S103, according to the current signal strength, the data transmission rate per unit time and the network packet loss rate per unit time period, the network signal evaluation index corresponding to each moment is calculated by a support vector machine; S104. Use the acquired network signal evaluation index and chip SOC state quantity to construct a signal transmission-power consumption curve, and obtain the network signal evaluation index value corresponding to the minimum signal transmission power consumption value as the data batch transmission signal index threshold [q1...qk] in the corresponding time period, where qk represents the data batch transmission signal index threshold in the kth time period.

3. The edge-cloud collaborative positioning method based on composite sensors according to claim 2, characterized in that: The step of constructing the storage-transmission regulation model in S1 further includes: S105, obtaining the size of the positioning data packet of the corresponding type of positioning task, the required computing resource data, the remaining available computing resources of the chip at the corresponding time, the computing complexity of each task and the local positioning task computing delay, and inputting the obtained data into the computing complexity model constructed by the support vector machine for training, to obtain a trained computing complexity model; S106. Deploy the trained computational complexity model into the local database of the chip, input the positioning acquisition data packet size, the required computing resource data and the remaining available computing resources of the chip at the current moment of each priority type positioning task obtained in S102 into the computational complexity model, and obtain the computational complexity and computational delay of the corresponding positioning task.

4. The edge-cloud collaborative positioning method based on composite sensors according to claim 3, characterized in that: The step of constructing the storage-transmission regulation model in S1 further includes: S107: construct the current input state of the storage-transmission control model according to the acquired positioning task processing priority, positioning acquisition data packet size, current signal strength, unit time data transmission rate and current chip SOC state quantity. Where soct represents the current storage capacity of the positioning chip. represents the computational complexity of the i-th positioning task at the current moment, represents the calculation delay of the i-th positioning task at the current moment, qt represents the network signal evaluation index corresponding to the current moment, represents the positioning calculation accuracy of the jth local positioning task at the current moment, soc0 represents the minimum working power storage capacity threshold corresponding to the positioning chip; S108, according to the current input state, setting the current execution action trigger information, including: When the current moment qt=qk and soct>soc0, the first execution action trigger information is triggered; In the state of triggering the first execution action trigger information, the corresponding subordinate execution action trigger information includes: S1081. When di is greater than or equal to the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is greater than or equal to the positioning calculation accuracy threshold, the second execution action trigger information is triggered; S1082, when di is greater than or equal to the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is less than the positioning calculation accuracy threshold, the third execution action trigger information is triggered.

5. The edge-cloud collaborative positioning method based on composite sensors according to claim 4, characterized in that: The setting of the trigger information for executing the action at the current moment also includes: S1083. When di is greater than or equal to the configured task processing priority threshold, Greater than or equal to the maximum complexity threshold of the locally executable calculation or is greater than the configured computational latency threshold and When it is less than the positioning calculation accuracy threshold, the fourth execution action trigger information is triggered; S1084. When di is less than the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is greater than or equal to the positioning calculation accuracy threshold, the fifth execution action trigger information is triggered; S1085. When di is less than the configured task processing priority threshold, Greater than or equal to the maximum complexity threshold of the locally executable calculation or When it is less than the configured calculation delay threshold, the sixth execution action trigger information is triggered; S1086. Set the maximum throughput threshold corresponding to each network signal evaluation indicator. When the size of the positioning acquisition data packet set to be uploaded for the positioning task at the current moment is equal to the maximum throughput threshold corresponding to the network signal evaluation indicator, upload the corresponding batch task to the cloud, and perform positioning calculation by calling the collaborative positioning sub-model.

6. The edge-cloud collaborative positioning method based on composite sensors according to claim 5, characterized in that: The setting of the trigger information for executing the action at the current moment also includes: When the current moment qt≠qk and soct>soc0, the seventh execution action trigger information is triggered; In the state of triggering the seventh execution action triggering information, the corresponding subordinate execution action triggering information includes: S1087, when di is greater than or equal to the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is greater than or equal to the positioning calculation accuracy threshold, the second execution action trigger information is triggered; S1088, when di is greater than or equal to the configured task processing priority threshold, Less than the maximum computation complexity threshold that can be performed locally or is less than the configured computational latency threshold and When it is less than the positioning calculation accuracy threshold, the eighth execution action trigger information is triggered; S1089, when di is less than or equal to the configured task processing priority threshold, if When it is greater than or equal to the positioning calculation accuracy threshold, the fifth execution action trigger information is triggered. When it is less than the positioning calculation accuracy threshold, the eighth execution action trigger information is triggered; S1090: When soct≤soc0 at the current moment, the second execution action trigger information is triggered, and a low battery warning is issued.

7. The edge-cloud collaborative positioning method based on composite sensors according to claim 1, characterized in that: The calculation step of performing edge positioning calculation or edge-cloud collaborative allocation positioning calculation on the positioning task according to the determination result in S4 includes: S401, real-time monitoring and collection of positioning data packets corresponding to the positioning task, performing initial positioning calculations through the deployed edge positioning sub-model, and directly feeding back the corresponding results to the corresponding user if the corresponding positioning accuracy meets the positioning calculation accuracy threshold; S402. When the corresponding positioning accuracy does not meet the positioning calculation accuracy threshold and the corresponding data is uploaded to the cloud, the collaborative positioning sub-model is called to perform secondary calculation based on the positioning acquisition data packet of the corresponding positioning task and the initial positioning calculation result of the edge positioning sub-model to obtain the secondary positioning task calculation result.

8. An edge-cloud collaborative positioning system based on a composite sensor, which is used to implement the edge-cloud collaborative positioning method based on a composite sensor as described in any one of claims 1 to 7, characterized in that: include: Transmission discrimination module, complexity module and discrimination calculation module; The transmission identification module is used for data acquisition and data transmission identification; The transmission determination module includes a data acquisition unit and a transmission determination unit; The data acquisition unit is used to acquire the positioning task status data, power storage status and signal strength data of the positioning chip and perform pre-processing and local chip storage; The transmission discrimination unit is used to perform local storage batch data transmission discrimination through the configured storage-transmission control model, and perform edge-to-cloud data transmission through the configured full-network module; The complexity module is used to obtain the current positioning task information and obtain the computational complexity of the corresponding positioning task through the configured computational complexity model.

9. The edge-cloud collaborative positioning system based on composite sensors as claimed in claim 8, characterized in that: The discrimination computing module includes a deployment unit, a computation discrimination unit, and an edge-to-cloud computing unit; The deployment unit is used to build the edge positioning sub-model in the configured edge-cloud collaborative positioning model into the chip, and deploy the collaborative positioning sub-model to the cloud; The calculation and determination unit is used to perform edge positioning calculation and edge-cloud collaborative positioning calculation and determination according to the calculation complexity of the corresponding positioning task, the chip power storage state and the signal strength, and obtain the determination result; The edge-cloud computing unit is used to perform edge positioning calculation or edge-cloud collaborative positioning calculation according to the judgment result of the calculation judgment unit and the corresponding collected positioning task status data, through the deployed edge positioning sub-model or collaborative positioning sub-model, to obtain the positioning result of the corresponding positioning task.

10. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the computer instructions are executed, the edge-cloud collaborative positioning method based on composite sensors as described in any one of claims 1 to 7 is executed.

Images