LBS (Location Based Service)-driven industrial chain precise intelligent docking method

Abstract

The invention relates to the field of industrial chain docking, and discloses an LBS-driven industrial chain precise intelligent docking method, which comprises the following steps: acquisition and alignment: acquiring corresponding data through a multi-modal sensor network deployed at industrial chain nodes, and performing space-time alignment to generate multi-source fusion data under a unified coordinate system; error compensation: based on the multi-source fusion data, performing anti-interference processing on a signal through spectrum sensing and a deep learning model, dynamically compensating a positioning error, and outputting preliminary positioning data; and calculation optimization: inputting the preliminary positioning data into an edge calculation node for real-time calculation, and aggregating multi-node positioning error characteristics and model parameters through a federated learning mechanism. The method comprises the following steps: acquiring multi-source data, dynamically switching ultra-wideband frequency bands by using a spectrum sensing technology to resist interference, compensating positioning errors, resolving in real time by using edge computing nodes, optimizing a global positioning model and dynamically adjusting space-time constraint weights in combination with a federated learning mechanism, and analyzing a supply-demand relationship to trigger a docking instruction.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial chain docking, and specifically to a precise intelligent docking method for the industrial chain driven by LBS. Background Art

[0002] In today's digital age, the efficient operation of industrial chains in various industries is crucial for enterprise competitiveness and industry development. As a key link in the efficient operation of the industrial chain, industrial chain docking involves the coordinated cooperation of multiple links, from raw material supply, product manufacturing, to logistics transportation, product sales, etc. With the rapid development of technologies such as the Internet of Things, big data, and artificial intelligence, industrial chain docking is moving towards the direction of intelligence and precision. By collecting, analyzing, and applying various types of data, realizing information sharing and collaborative decision-making among all links of the industrial chain can effectively improve production efficiency, reduce costs, and improve product quality, thereby enhancing the market competitiveness of the entire industrial chain. Against this background, the application of location-based service (LBS) technology in industrial chain docking has gradually received wide attention, which can provide key location-related information for all links of the industrial chain and provide strong support for precise docking.

[0003] When traditional docking methods are applied, in complex industrial and urban environments, there are a large number of electromagnetic interference sources. For example, large motors, welders and other equipment in the production workshop will generate strong electromagnetic interference, affecting the accurate collection of position signals by sensors. At the same time, obstacles such as buildings and mountains will block the signals, further reducing the accuracy and stability of positioning, resulting in a significant discount in the reliability of location information. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a precise intelligent docking method for the industrial chain driven by LBS, which solves the problem that traditional docking methods reduce the accuracy and stability of positioning in the face of electromagnetic interference and multiple obstructions, resulting in a significant discount in the reliability of location information.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: A precise intelligent docking method for the industrial chain driven by LBS, including the following steps: S1. Acquisition and Alignment: Collect corresponding data through a multi-modal sensor network deployed at industrial chain nodes, and perform spatio-temporal alignment to generate multi-source fusion data in a unified coordinate system; S2. Error Compensation: Based on the multi-source fusion data, perform anti-interference processing on the signal through spectrum sensing and deep learning models, dynamically compensate for positioning errors, and output preliminary positioning data; S3. Computational Optimization: Input the preliminary positioning data into the edge computing node for real-time solution to generate real-time positioning parameters. Meanwhile, aggregate the positioning error features and model parameters of multiple nodes through the federated learning mechanism, optimize the global positioning model, and output the positioning parameters including spatio-temporal constraint weights. S4. Mapping and Docking: Drive the digital twin model based on the positioning parameters, map the physical industrial chain resource status in real time, analyze the supply-demand relationship through the intelligent decision-making engine, and trigger dynamic docking instructions. S5. Verification and Construction: Verify the positioning accuracy and docking efficiency according to the dynamic docking instructions, construct a standardized data interface and cross-industry LBS data trading ecosystem based on the verification results, and feedback the evaluation results.

[0006] Preferably, the corresponding data in S1 includes the real-time coordinates, movement trajectories, and load status of logistics transportation carriers, the operating angles, vibration frequencies, and temperature and humidity environments of production equipment. The multi-modal sensor network includes satellite navigation and positioning devices, ultra-wideband transceivers, acoustic sensing devices, computer vision, and autonomous navigation devices. The spatio-temporal alignment is to map multi-source data to a unified coordinate system through an aerial view model, , where is the aerial view transformation matrix, , are the sensor calibration parameters, is the ultra-wideband ranging value, are the visual pixel coordinates.

[0007] Preferably, the positioning errors in S2 include errors caused by multipath effects, electromagnetic interference, or occlusion. The anti-interference processing includes dynamically switching the ultra-wideband frequency band, and its switching logic is: , where is the interference power.

[0008] Preferably, the deep learning model in S2 is a convolutional neural network-long short-term memory network, which inputs the spectral features of the input signal and the environmental point cloud data and outputs the coordinate correction amount.

[0009] Preferably, the dynamic compensation in S2 includes multipath suppression error correction, and its formula is: , where = [frequency, signal strength, signal-to-noise ratio], is the lidar point cloud. During occlusion, the trajectory is predicted through inertial-visual fusion, where is the initial velocity, is the acceleration, is the visual optical flow velocity, is the adaptive weight.

[0010] Preferably, the global model update formula of the federated learning mechanism in S3 is , where is the amount of node data, is the local model parameter, and the spatio-temporal constraint weight is dynamically adjusted according to the regional environment.

[0011] Preferably, in S4, the intelligent decision-making engine optimizes the logistics path through reinforcement learning, and its action-value function update formula is , where is the state, is the next state the maximum of all possible actions under value, is the reward, is the discount factor.

[0012] Preferably, in S4, the intelligent decision-making engine integrates a long short-term memory network to construct a supply-demand prediction model. The supply-demand prediction model inputs historical regional demand data and outputs predicted values for future time periods, triggering dynamic docking instructions.

[0013] Preferably, S5 specifically includes the following steps: S501: According to the dynamic docking instruction, obtain the actual position data of the positioning object and the time data of the docking process. At the same time, record the time node from the instruction issuance to the docking completion; S502: Compare the actual position data with the position determined by the positioning parameters, calculate the deviation between the two, and statistically calculate the mean and variance of the positioning deviation to evaluate the positioning accuracy; S503: Calculate the total duration of the docking operation according to the time node, and establish an efficiency evaluation model in combination with different business scenarios and constraint conditions to comprehensively calculate the docking efficiency index.

[0014] S504: Based on the evaluation results of the positioning accuracy and docking efficiency index, construct a standardized data interface. At the same time, build a cross-industry LBS data trading ecosystem; S505: Feed back the evaluation results to the positioning error and federated learning mechanism, adjust the error compensation parameters according to the positioning accuracy problem, and optimize the federated learning strategy for the docking efficiency problem.

[0015] Preferably, in the evaluation of the positioning accuracy in S502, when the deviation exceeds the predetermined threshold, the areas and links that need to be optimized with key points are marked. The considerations in the efficiency evaluation model in S503 include the docking duration and resource utilization rate.

[0016] The present invention provides a method for precise intelligent docking of an LBS-driven industrial chain. It has the following beneficial effects: 1. The present invention collects multi-source data, uses spectrum sensing technology to dynamically switch the ultra-wideband frequency band for anti-interference, compensates for positioning errors, uses edge computing nodes to perform real-time calculations, combines the federated learning mechanism to optimize the global positioning model and dynamically adjusts the spatio-temporal constraint weights, analyzes the supply-demand relationship to trigger docking instructions, verifies the positioning accuracy and docking efficiency according to the docking instructions, and feeds back the results to optimize the relevant models and mechanisms, realizing high-precision and stable positioning and precise docking of the industrial chain, and solving the problems of poor positioning accuracy and stability of traditional docking methods in the electromagnetic interference and multi-occlusion environment, resulting in low reliability of position information.

[0017] 2. The present invention monitors the interference situation by using spectrum sensing technology, dynamically switches the ultra-wideband frequency band according to the interference power to avoid interference signals. At the same time, a deep learning model composed of a convolutional neural network and a long short-term memory network is adopted. The spectral features of the input signal and the environmental point cloud data are input, and the coordinate correction amount is output to dynamically compensate for the positioning error, improving the accuracy and stability of positioning.

[0018] 3. The present invention inputs the preliminary positioning data into the edge computing node for real-time calculation to generate real-time positioning parameters, reduces data transmission delay, aggregates the positioning error features and model parameters of multiple nodes by means of the federated learning mechanism, optimizes the global positioning model, and dynamically adjusts the spatio-temporal constraint weights according to the regional environment, enabling the positioning model to adapt to different environments, further improving the positioning accuracy, and enhancing the stability and reliability of positioning in complex environments.

[0019] 4. The present invention optimizes the logistics path through reinforcement learning by the intelligent decision-making engine, constructs a supply-demand prediction model by using the long short-term memory network, analyzes the supply-demand relationship and triggers dynamic docking instructions. The precise positioning provides a guarantee for the precise docking of the industrial chain, improves the collaborative efficiency of the industrial chain, and avoids docking mistakes caused by inaccurate positioning.

[0020] 5. The present invention feeds back the evaluation results to the error compensation model and the federated learning mechanism, adjusts the error compensation parameters according to the positioning accuracy problem, and optimizes the federated learning strategy for the docking efficiency problem, thereby continuously improving the positioning and docking effects, and ensuring that the positioning accuracy and stability can continuously maintain a high level in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a flowchart of the method for precise intelligent docking of the industrial chain driven by LBS proposed by the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0022] The technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0023] Please refer to the attached Figure 1 , the embodiment of the present invention provides an LBS-driven industrial chain precise intelligent docking method, including the following steps: S1. Acquisition and alignment: Collect corresponding data through a multi-modal sensor network deployed at industrial chain nodes, and perform spatio-temporal alignment to generate multi-source fusion data in a unified coordinate system; the corresponding data in S1 includes the real-time coordinates, movement trajectories, and load states of logistics transportation carriers, the operating angles, vibration frequencies, and temperature and humidity environments of production equipment, and the multi-modal sensor network includes satellite navigation and positioning devices, ultra-wideband transceivers, acoustic sensing devices, computer vision and autonomous navigation devices. The spatio-temporal alignment is to map multi-source data to a unified coordinate system through an aerial view model, , where is the aerial view transformation matrix, , are sensor calibration parameters, is the ultra-wideband ranging value, is the visual pixel coordinate.

[0024] Specifically, collect corresponding data through a multi-modal sensor network deployed at industrial chain nodes, and perform spatio-temporal alignment to generate multi-source fusion data in a unified coordinate system. Thus, at each key node of the industrial chain, deploy a multi-modal sensor network composed of satellite navigation and positioning devices, ultra-wideband transceivers, acoustic sensing devices, computer vision and autonomous navigation devices. The satellite navigation and positioning device accurately obtains the real-time coordinates and movement trajectories of logistics transportation carriers, and simultaneously monitors their load states; the acoustic sensing device and the temperature and humidity sensor are responsible for collecting the vibration frequency and temperature and humidity environment data around the production equipment; the ultra-wideband transceiver is used to obtain accurate distance information; the computer vision and autonomous navigation device provides rich visual pixel coordinate data. Then, using the aerial view model, map the data from different sensors that are different in time and space to a unified coordinate system according to specific conversion rules (involving the aerial view transformation matrix, sensor calibration parameters, etc.), and finally generate multi-source fusion data in a unified coordinate system. Thus, it integrates key data of different links and different types in the industrial chain, provides a comprehensive and unified data basis for subsequent precise positioning and docking, realizes the transformation of data from decentralized collection to centralized unity, and improves the availability and accuracy of data.

[0025] S2. Error Compensation: Based on multi-source fusion data, anti-interference processing is performed on the signal through spectrum sensing and a deep learning model, and the positioning error is dynamically compensated to output preliminary positioning data; the positioning error in S2 includes errors caused by multipath effects, electromagnetic interference, or occlusion, and the anti-interference processing includes dynamically switching the ultra-wideband frequency band, and its switching logic is: , where is the interference power.

[0026] In S2, the deep learning model is a convolutional neural network-long short-term memory network, which inputs the signal spectrum features and environmental point cloud data and outputs the coordinate correction amount.

[0027] The dynamic compensation in S2 includes multipath suppression error correction, and its formula is: , where = [frequency, signal strength, signal-to-noise ratio], is the lidar point cloud, and the trajectory is predicted through inertial-visual fusion during occlusion, where is the initial velocity, is the acceleration, is the visual optical flow velocity, is the adaptive weight.

[0028] Specifically, based on multi-source fusion data, anti-interference processing is performed on the signal through spectrum sensing and a deep learning model, and the positioning error is dynamically compensated to output preliminary positioning data, so as to use spectrum sensing technology to monitor the signal environment in real time during the transmission process of multi-source fusion data. When the detected interference power is greater than or equal to -90 dBm, the ultra-wideband frequency band is automatically switched to 3.5 - 4.5 GHz; in other interference cases, it is switched to 5.5 - 6.5 GHz to avoid interference signals. At the same time, the signal spectrum features and environmental point cloud data are input into a deep learning model composed of a convolutional neural network-long short-term memory network. After complex calculations, the signal deviation is analyzed, and the coordinate correction amount is output to dynamically compensate for the positioning error caused by factors such as multipath effects, electromagnetic interference, or occlusion, and then preliminary positioning data is obtained, effectively reducing the influence of external interference on the positioning signal, improving the positioning accuracy, and providing a more accurate basis for subsequent calculation optimization and docking operations.

[0029] S3. Calculation Optimization: The preliminary positioning data is input into the edge computing node for real-time calculation to generate real-time positioning parameters; at the same time, the multi-node positioning error features and model parameters are aggregated through the federated learning mechanism to optimize the global positioning model and output the positioning parameters including spatio-temporal constraint weights; the global model update formula of the federated learning mechanism in S3 is , where is the node data volume, is the local model parameter, and the spatio-temporal constraint weight is dynamically adjusted according to the regional environment.

[0030] Specifically, the preliminary positioning data is input into the edge computing node for real-time calculation to generate real-time positioning parameters; at the same time, the positioning error features and model parameters of multiple nodes are aggregated through the federated learning mechanism to optimize the global positioning model, and the positioning parameters including spatio-temporal constraint weights are output. Thus, the preliminary positioning data is input into the edge computing node, and the powerful real-time computing ability of the edge computing node is utilized to quickly calculate and generate real-time positioning parameters. At the same time, through the federated learning mechanism, the positioning error features and model parameters of multiple nodes are collected, just like gathering the experiences of all parties. Then, based on this information, the global positioning model is optimized, and the spatio-temporal constraint weights are dynamically adjusted according to the characteristics of different regional environments. Finally, the positioning parameters including spatio-temporal constraint weights are output. On the one hand, edge computing reduces the delay caused by data transmission and quickly generates real-time positioning parameters, meeting the real-time requirements of practical applications; on the other hand, the global positioning model optimized by the federated learning mechanism and the dynamically adjusted spatio-temporal constraint weights further improve the accuracy and adaptability of positioning, realizing the precise optimization of positioning parameters and making the positioning more in line with the actual scenario requirements.

[0031] S4. Mapping and docking: Drive the digital twin model based on the positioning parameters, map the resource status of the physical industrial chain in real time, and analyze the supply-demand relationship through the intelligent decision-making engine to trigger dynamic docking instructions; in S4, the intelligent decision-making engine optimizes the logistics path through reinforcement learning, and its action value function update formula is , where is the state, is the next state the maximum of all possible actions under value, is the reward, is the discount factor.

[0032] In S4, the intelligent decision-making engine integrates long short-term memory networks to construct a supply-demand prediction model. The supply-demand prediction model inputs historical regional demand data and outputs predicted values for future periods to trigger dynamic docking instructions.

[0033] Specifically, by driving the digital twin model based on positioning parameters, the status of physical industrial chain resources is mapped in real time, and the supply-demand relationship is analyzed through an intelligent decision-making engine to trigger dynamic docking instructions. Thus, by driving the digital twin model based on positioning parameters, the status of physical industrial chain resources can be presented in real time and intuitively, enabling relevant personnel to clearly understand the operation of the industrial chain. At the same time, the intelligent decision-making engine starts to work, continuously optimizing the logistics path through reinforcement learning, making optimal decisions based on different states and expected rewards. In addition, the intelligent decision-making engine uses a long short-term memory network to construct a supply-demand prediction model, inputs historical regional demand data, and predicts the demand situation in the future period. When it is predicted that the supply and demand are unbalanced, dynamic docking instructions are triggered, realizing the real-time visualization of the status of physical industrial chain resources, facilitating monitoring and management. The intelligent decision-making engine optimizes the logistics path, improving the logistics efficiency; the triggering of the supply-demand prediction model and dynamic docking instructions realizes the precise matching of the supply and demand of the industrial chain, improving the collaborative operation efficiency of the industrial chain, and enabling all links of the industrial chain to cooperate more closely.

[0034] S5. Verification and Construction: Verify the positioning accuracy and docking efficiency according to the dynamic docking instructions, construct a standardized data interface and a cross-industry LBS data trading ecosystem based on the verification results, and feedback the evaluation results.

[0035] S5 specifically includes the following steps: S501. According to the dynamic docking instructions, obtain the actual position data of the positioning object and the time data of the docking process. At the same time, record the time node from the instruction issuance to the docking completion; S502. Compare the actual position data with the position determined by the positioning parameters, calculate the deviation between the two, and statistically calculate the mean and variance of the positioning deviation to evaluate the positioning accuracy; S503: Calculate the total duration of the docking operation according to the time node, establish an efficiency evaluation model in combination with different business scenarios and constraints, and comprehensively calculate the docking efficiency index.

[0036] S504: Based on the evaluation results of the positioning accuracy and docking efficiency indicators, construct a standardized data interface, and at the same time, build a cross-industry LBS data trading ecosystem; S505: Feedback the evaluation results to the positioning error and federated learning mechanism, adjust the error compensation parameters according to the positioning accuracy problem, and optimize the federated learning strategy for the docking efficiency problem.

[0037] In the evaluation of the positioning accuracy in S502, when the deviation exceeds the predetermined threshold, mark the areas and links that need to be optimized with priority. The considerations in the efficiency evaluation model in S503 include the docking duration and resource utilization rate.

[0038] Specifically, by verifying the positioning accuracy and docking efficiency according to the dynamic docking instruction, constructing a standardized data interface and a cross-industry LBS data trading ecosystem based on the verification results, and feeding back the evaluation results, the actual position data of the positioning object and the time data of the docking process can be obtained according to the dynamic docking instruction, and each key time node from the instruction issuance to the docking completion is detailedly recorded. The actual position data is compared with the position determined by the positioning parameters, the deviation between the two is calculated, and the positioning accuracy is evaluated by statistically calculating the mean and variance of the positioning deviation. The total duration of the docking operation is calculated according to the recorded time nodes, and combined with different business scenarios and constraints (such as the actual usage of resources, production requirements of different links, etc.), an efficiency evaluation model is established, and the docking efficiency index is comprehensively calculated. Based on the evaluation results of the positioning accuracy and docking efficiency indexes, a unified and standardized data interface is formulated, and a cross-industry LBS data trading ecosystem is built. Finally, the evaluation results are fed back to the model and federated learning mechanism responsible for positioning error compensation, the error compensation parameters are adjusted according to the positioning accuracy problem, and the federated learning strategy is optimized for the docking efficiency problem, so as to comprehensively evaluate and feedback-optimize the entire docking process. The construction of the standardized data interface and the cross-industry LBS data trading ecosystem promotes data sharing and interaction among all links of the industrial chain and different industries, improves the overall collaborative efficiency, and feeding back the evaluation results to the relevant models and mechanisms realizes the self-optimization and continuous improvement of the entire docking method, continuously improves the positioning accuracy and docking efficiency, and makes the precise and intelligent docking of the industrial chain more perfect.

[0039] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made therein without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The LBS-driven precise intelligent docking method for the industrial chain is characterized by: The following steps are involved: S1. Collection and alignment: Collect corresponding data through the multimodal sensor network deployed at the nodes of the industrial chain, and perform spatiotemporal alignment to generate multi-source fusion data in a unified coordinate system; S2. Error compensation: Based on the multi-source fusion data, the signal is processed for anti-interference through spectrum sensing and deep learning models, the positioning error is dynamically compensated, and preliminary positioning data is output; S3, Computation optimization: Input the preliminary positioning data into the edge computing node for real-time solution to generate real-time positioning parameters; at the same time, aggregate the multi-node positioning error characteristics and model parameters through the federated learning mechanism, optimize the global positioning model, and output the positioning parameters containing the spatiotemporal constraint weights; S4, Mapping and docking: Drive the digital twin model based on positioning parameters, map the resource status of the physical industrial chain in real time, analyze the supply and demand relationship through the intelligent decision-making engine, and trigger dynamic docking instructions; S5. Verification and construction: Verify the positioning accuracy and docking efficiency according to dynamic docking instructions, build a standardized data interface and cross-industry LBS data transaction ecosystem based on the verification results, and feedback the evaluation results.

2. The LBS-driven precise intelligent docking method for the industrial chain according to claim 1, characterized in that: The corresponding data in S1 include the real-time coordinates, motion trajectory and load status of the logistics transport carrier, the operating angle, vibration frequency and temperature and humidity environment of the production equipment. The multimodal sensor network includes satellite navigation and positioning equipment, ultra-wideband transceiver, acoustic sensor equipment, computer vision and autonomous navigation equipment. The spatiotemporal alignment is to map multi-source data to a unified coordinate system through a bird's-eye view model. ,in is the bird's eye view transformation matrix, , Calibrate the sensor parameters. is the ultra-wideband ranging value, is the visual pixel coordinate.

3. The LBS-driven precise intelligent docking method for the industrial chain according to claim 1, characterized in that: The positioning error in S2 includes errors caused by multipath effects, electromagnetic interference or shielding. The anti-interference processing includes dynamically switching the ultra-wideband frequency band, and the switching logic is: ,in is the interference power.

4. The LBS-driven precise intelligent docking method for the industrial chain according to claim 1, characterized in that: The deep learning model in S2 is a convolutional neural network-long short-term memory network, which inputs signal spectrum characteristics and environmental point cloud data and outputs coordinate corrections.

5. The LBS-driven industry chain precise intelligent docking method according to claim 1, characterized in that: The dynamic compensation in S2 includes multipath suppression error correction, and its formula is: ,in =[frequency, signal strength, signal-to-noise ratio], is the lidar point cloud, and the trajectory is predicted by inertial-visual fusion during occlusion, where is the initial velocity, is the acceleration, is the visual optical flow speed, is the adaptive weight.

6. The LBS-driven industry chain precise intelligent docking method according to claim 1, characterized in that: The global model update formula of the federated learning mechanism in S3 is: ,in For the Node data volume, is a local model parameter, and the spatiotemporal constraint weight is dynamically adjusted according to the regional environment.

7. The LBS-driven industry chain precise intelligent docking method according to claim 1, characterized in that: The intelligent decision engine in S4 optimizes the logistics path through reinforcement learning, and its action value function update formula is: ,in For status, For the next state The maximum of all possible moves value, For reward, is the discount factor.

8. The LBS-driven precise intelligent docking method for the industrial chain according to claim 1, characterized in that: The intelligent decision engine in S4 integrates a long short-term memory network to build a supply and demand forecasting model. The supply and demand forecasting model inputs historical regional demand data, outputs future time period forecast values, and triggers dynamic docking instructions.

9. The LBS-driven industry chain precise intelligent docking method according to claim 1, characterized in that: The S5 specifically includes the following steps: S501, according to the dynamic docking instruction, obtain the actual position data of the positioning object and the time data of the docking process, and at the same time, record the time node from the issuance of the instruction to the completion of the docking; S502: Compare the actual position data with the position determined by the positioning parameters, calculate the deviation between the two, and evaluate the positioning accuracy by statistically analyzing the mean and variance of the positioning deviation; S503: Calculate the total duration of the docking operation according to the time node, establish an efficiency evaluation model based on different business scenarios and constraints, and comprehensively calculate the docking efficiency index; S504: Based on the evaluation results of positioning accuracy and docking efficiency indicators, a standardized data interface is constructed, and at the same time, a cross-industry LBS data transaction ecosystem is established; S505: Feedback the evaluation results to the positioning error and federated learning mechanism, adjust the error compensation parameters according to the positioning accuracy problem, and optimize the federated learning strategy for the docking efficiency problem.

10. The LBS-driven industry chain precise intelligent docking method according to claim 9, characterized in that: When the deviation in the positioning accuracy evaluation in S502 exceeds a predetermined threshold, the areas and links that need to be optimized are marked. The efficiency evaluation model in S503 considers the connection time and resource utilization.

Images