Positioning model adaptive adjustment method, system and device and storage medium

Abstract

The invention discloses a positioning model adaptive adjustment method, system and device and a storage medium, and relates to the technical field of artificial intelligence, and the method comprises the steps: selecting a corresponding positioning model in a model library based on initial environment parameters and scene demands; positioning based on the positioning model to obtain positioning data; based on the environment data and the positioning data collected in real time, monitoring indexes are determined, and the monitoring indexes at least comprise a performance index and an environment change index; when the monitoring index meets the adjustment triggering condition, determining a self-adaptive adjustment strategy; and adjusting the positioning model based on a self-adaptive adjustment strategy. Through the mode, the performance of the positioning model and the current environment are monitored in real time, when the performance of the model is reduced or the environment is remarkably changed, the positioning model is adjusted in time, self-adaptive adjustment based on the monitoring result is achieved, the method is suitable for various complex scenes, the positioning precision in different scenes can be improved, meanwhile, the calculation complexity is reduced, and the user experience is improved. And user experience is improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and particularly to a method, system, device, and storage medium for adaptively adjusting a positioning model. Background Art

[0002] With the rapid development of machine learning (ML) and artificial intelligence (AI) technologies, direct AI / ML positioning and AI / ML-assisted positioning have begun to be used in NR (New Radio). Since different AI model frameworks are applicable to different scenarios and there are significant differences in performance, in order to enable AI / ML models to exhibit high stability and accuracy in practical applications, performance monitoring of the models and selection management of the models have become increasingly important. However, existing model monitoring systems usually rely on static monitoring strategies and cannot dynamically adapt to the performance changes of models in different locations and environments, making it difficult to make timely adjustments according to dynamic changes, which affects the positioning accuracy.

[0003] The above content is only used to assist in understanding the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main purpose of this application is to provide a method, system, device, and storage medium for adaptively adjusting a positioning model, aiming to solve the technical problem that static monitoring strategies in the prior art are difficult to dynamically adjust the positioning model in a timely manner according to scene changes, which affects the positioning accuracy.

[0005] To achieve the above purpose, this application provides a method for adaptively adjusting a positioning model, and the method includes:

[0006] Select a corresponding positioning model from the model library based on the initial environmental parameters and scene requirements;

[0007] Perform positioning based on the positioning model to obtain positioning data;

[0008] Determine monitoring metrics based on the real-time collected environmental data and the positioning data, where the monitoring metrics at least include performance metrics and environmental change metrics;

[0009] When the monitoring metrics meet the adjustment trigger condition, determine an adaptive adjustment strategy based on the monitoring metrics;

[0010] Adjust the positioning model based on the adaptive adjustment strategy.

[0011] In one embodiment, the environmental data at least includes temperature, humidity, and air pressure, and the positioning data at least includes positioning coordinates and positioning time delay. The steps of determining the monitoring metrics based on the real-time collected environmental data and the positioning data include:

[0012] Based on the environmental data and the reference state data, calculate the dynamic cosine similarity, and use the dynamic cosine similarity as the environmental change metric;

[0013] Based on the positioning coordinates and the received signal strength corresponding to the environmental data, calculate the positioning error confidence level, and use the positioning error confidence level and the positioning time delay as the performance metrics.

[0014] In one embodiment, the steps of calculating the dynamic cosine similarity based on the environmental data and the reference state data include:

[0015] Perform normalization processing on the environmental data to obtain environmental state data;

[0016] Based on historical steady-state data, determine the reference state data;

[0017] Obtain the first correspondence relationship between the reference state data, the environmental state data, and the dynamic cosine similarity;

[0018] Based on the reference state data, the environmental state data, and the first correspondence relationship, obtain the dynamic cosine similarity.

[0019] In one embodiment, the received signal strength includes the maximum received signal strength and the minimum received signal strength. The steps of calculating the positioning error confidence level based on the positioning coordinates and the received signal strength corresponding to the environmental data include:

[0020] Obtain the reference coordinates collected by the measuring device;

[0021] Obtain the second correspondence relationship between the positioning coordinates, the reference coordinates, the maximum received signal strength, the minimum received signal strength, and the positioning error confidence level;

[0022] Based on the positioning coordinates, the reference coordinates, the maximum received signal strength, the minimum received signal strength, and the second correspondence relationship, obtain the positioning error confidence level.

[0023] In one embodiment, the steps of determining the adaptive adjustment strategy based on the monitoring metrics when the monitoring metrics meet the adjustment trigger condition include:

[0024] Based on historical similarity data, determine the similarity mean and the similarity standard deviation;

[0025] Obtain the third corresponding relationship among the similarity mean, the similarity standard deviation, and the similarity threshold;

[0026] Based on the similarity mean, the similarity standard deviation, and the third corresponding relationship, obtain the similarity threshold;

[0027] When the dynamic cosine similarity is less than the similarity threshold, determine that the monitoring index meets the adjustment trigger condition, and determine that the adaptive adjustment strategy is model switching.

[0028] In one embodiment, the step of determining the adaptive adjustment strategy based on the monitoring index when the monitoring index meets the adjustment trigger condition includes:

[0029] When the positioning error confidence is less than the first confidence threshold and greater than or equal to the second confidence threshold, determine that the monitoring index meets the adjustment trigger condition, and determine that the adaptive adjustment strategy is model optimization, where the second confidence threshold is less than the first confidence threshold;

[0030] When the positioning error confidence is less than the second confidence threshold, determine that the monitoring index meets the adjustment trigger condition, and determine that the adaptive adjustment strategy is any one of model switching and model fallback.

[0031] In one embodiment, the step of adjusting the positioning model based on the adaptive adjustment strategy includes:

[0032] When the adaptive adjustment strategy is model optimization, optimize the positioning model, where the optimization is any one of retraining and updating parameters;

[0033] When the adaptive adjustment strategy is model switching, select a target positioning model from the model library based on the environmental data and the scenario requirements, and update the positioning model to the target positioning model;

[0034] When the adaptive adjustment strategy is model fallback, perform positioning based on a preset positioning strategy.

[0035] In addition, to achieve the above object, the present application also proposes a positioning model adaptive adjustment system, and the positioning model adaptive adjustment system includes:

[0036] A model decision module, configured to select a corresponding positioning model from the model library based on initial environmental parameters and scenario requirements;

[0037] A model monitoring module, configured to perform positioning based on the positioning model to obtain positioning data;

[0038] The model monitoring module is further configured to determine monitoring metrics based on the real-time collected environmental data and the positioning data, where the monitoring metrics at least include performance metrics and environmental change metrics;

[0039] The model decision-making module is further configured to determine an adaptive adjustment strategy based on the monitoring metrics when the monitoring metrics meet the adjustment trigger condition;

[0040] The model decision-making module is further configured to adjust the positioning model based on the adaptive adjustment strategy.

[0041] In addition, to achieve the above object, the present application further provides a positioning model adaptive adjustment device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the computer program is configured to implement the steps of the positioning model adaptive adjustment method as described above.

[0042] In addition, to achieve the above object, the present invention further provides a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium, and when the computer program is executed by a processor, it implements the steps of the positioning model adaptive adjustment method as described above.

[0043] In addition, to achieve the above object, the present application further provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the positioning model adaptive adjustment method as described above.

[0044] The present application provides a positioning model adaptive adjustment method. Based on initial environmental parameters and scenario requirements, a corresponding positioning model is selected from a model library; positioning is performed based on the positioning model to obtain positioning data; monitoring metrics are determined based on the real-time collected environmental data and the positioning data, where the monitoring metrics at least include performance metrics and environmental change metrics; when the monitoring metrics meet the adjustment trigger condition, an adaptive adjustment strategy is determined based on the monitoring metrics; and the positioning model is adjusted based on the adaptive adjustment strategy. The present application monitors the performance of the positioning model and the current environment in real time. When the model performance deteriorates or the environment changes significantly, the positioning model is adjusted in a timely manner, and the most suitable model is used for positioning, realizing adaptive adjustment based on the monitoring results, being applicable to various complex scenarios, improving the positioning accuracy in different scenarios, reducing the computational complexity at the same time, and enhancing the user experience, solving the technical problem that a static monitoring strategy is difficult to dynamically adjust the positioning model in a timely manner according to scenario changes, affecting the positioning accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings here are incorporated into the specification and form a part of this specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application.

[0046] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0047] Figure 1 It is a schematic flowchart of the first embodiment of the positioning model adaptive adjustment method of the present application;

[0048] Figure 2 It is a schematic overall architecture diagram of the positioning model adaptive adjustment method provided by the first embodiment of the present application;

[0049] Figure 3 It is a schematic diagram of the indirect positioning monitoring decision signaling process of the positioning model adaptive adjustment method provided by the first embodiment of the present application;

[0050] Figure 4 It is a schematic diagram of the direct positioning monitoring decision signaling process of the positioning model adaptive adjustment method provided by the first embodiment of the present application;

[0051] Figure 5 It is a schematic flowchart of the second embodiment of the positioning model adaptive adjustment method of the present application;

[0052] Figure 6 It is a schematic brief flowchart of the positioning model adaptive adjustment method provided by the second embodiment of the present application;

[0053] Figure 7 It is a schematic diagram of the module structure of the positioning model adaptive adjustment system of the embodiments of the present application;

[0054] Figure 8 It is a schematic diagram of the device structure of the hardware operating environment involved in the positioning model adaptive adjustment method in the embodiments of the present application.

[0055] The realization of the objectives of the present application, the functional features and advantages will be further described with reference to the embodiments and the drawings. Specific Embodiments

[0056] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0057] To better understand the technical solutions of the present application, the following will be described in detail in combination with the drawings of the specification and the specific embodiments.

[0058] The main solution of the embodiment of the present application is as follows: Based on the initial environmental parameters and scene requirements, select the corresponding positioning model in the model library; perform positioning based on the positioning model to obtain positioning data; determine monitoring indicators based on the real-time collected environmental data and the positioning data, and the monitoring indicators at least include performance indicators and environmental change indicators; when the monitoring indicators meet the adjustment trigger condition, determine an adaptive adjustment strategy based on the monitoring indicators; and adjust the positioning model based on the adaptive adjustment strategy.

[0059] Currently, existing model monitoring systems usually rely on static monitoring strategies, and cannot dynamically adapt to the performance changes of the model in different locations and environments. It is difficult to make timely adjustments according to dynamic changes, which affects the positioning accuracy.

[0060] The present application provides a solution to monitor the performance of the positioning model and the current environment in real time. When the model performance deteriorates or the environment changes significantly, the positioning model is adjusted in a timely manner, and the most suitable model is used for positioning, realizing adaptive adjustment based on the monitoring results. It is applicable to various complex scenarios, can improve the positioning accuracy in different scenarios, reduce the computational complexity at the same time, and enhance the user experience. It solves the technical problem that it is difficult for static monitoring strategies to dynamically adjust the positioning model in a timely manner according to scene changes, which affects the positioning accuracy.

[0061] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication, and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device, a positioning model adaptive adjustment device, etc. that can implement the above functions. This embodiment does not make specific limitations in this regard. Hereinafter, taking the positioning model adaptive adjustment device as an example, this embodiment and the following embodiments will be described.

[0062] The embodiment of the present application provides a method for adaptive adjustment of a positioning model. Refer to Figure 1 , Figure 1 which is a schematic flowchart of the first embodiment of the method for adaptive adjustment of the positioning model of the present application.

[0063] In this embodiment, the method for adaptive adjustment of the positioning model includes steps S10 to S50:

[0064] Step S10, based on the initial environmental parameters and scene requirements, select the corresponding positioning model in the model library;

[0065] It should be noted that in this embodiment, a positioning model is used for positioning. The positioning model can be of types such as AI / ML positioning model, machine learning model, statistical model, etc., and no specific limitation is made thereto. Since different models are applicable to different scenarios, a model library is set up in this embodiment. The model library stores positioning models for different application scenarios, such as positioning models for indoor high-density population scenarios, indoor industrial scenarios, outdoor urban scenarios, and outdoor rural positioning scenarios, and no specific limitation is made thereto. Each model has a unique identifier, and the information of the identifier includes at least parameters such as model name, model version, application scenario, model accuracy, and model latency.

[0066] In addition, it should be noted that the initial environmental parameters are the initial environmental conditions, such as the initial temperature, initial humidity, initial air pressure, etc., which are used to judge the current scenario. The initial environmental parameters can be obtained through devices such as sensors. The scenario requirements are the positioning requirements of the current scenario, which are usually provided by the user through the user interface. The scenario requirements include at least accuracy requirements and latency requirements. The accuracy requirement is the accuracy requirement during positioning, usually setting an accuracy threshold as the minimum accuracy. The latency requirement is the latency requirement during positioning, usually setting a latency threshold as the maximum latency. Generally speaking, meeting the accuracy requirement means being greater than or equal to the accuracy threshold, and meeting the latency requirement means being less than or equal to the latency threshold. That is to say, to meet the scenario requirements, the positioning accuracy needs to be greater than or equal to the accuracy threshold and the positioning latency needs to be less than or equal to the latency threshold.

[0067] It can be understood that different scenarios usually correspond to different scenario requirements. Exemplarily, Scenario A requires high accuracy and low latency. At this time, a relatively large value can be set for the accuracy threshold, and a relatively small value can be set for the latency threshold; Scenario B requires high accuracy but has no requirement for latency. At this time, a relatively large value can be set for the accuracy threshold, and a relatively large value can be set for the latency threshold; Scenario C requires low latency but has no requirement for accuracy. At this time, a relatively small value can be set for the accuracy threshold, and a relatively small value can be set for the latency threshold.

[0068] It should be understood that according to the current scenario and the accuracy requirements and latency requirements, the most suitable positioning model is found in the preset model library. Exemplarily, if Scenario A requires high accuracy and low latency, and there are multiple models in the model library that meet the conditions, when the positioning accuracies are the same, the model with a lower positioning latency is selected; when the positioning latencies are the same, the model with a higher positioning accuracy is selected.

[0069] Step S20, perform positioning based on the positioning model to obtain positioning data;

[0070] It should be noted that when using the selected positioning model for positioning, the generated relevant data is the positioning data. The positioning data includes at least the positioning coordinates and the positioning time delay. The positioning coordinates are the coordinates of the located position, and the positioning time delay is the time taken to obtain the positioning coordinates.

[0071] Step S30: Based on the real-time collected environmental data and the positioning data, determine the monitoring metrics, where the monitoring metrics include at least the performance metrics and the environmental change metrics.

[0072] It should be noted that the environmental data is the surrounding environmental conditions collected in real time, including at least temperature, humidity, and air pressure, and no specific limitations are imposed thereon. The environmental data is usually obtained in real time using devices such as sensors. In addition to the environmental data and the positioning data, device data is usually also collected, such as: device model, sensor accuracy, etc., and no specific limitations are imposed in this embodiment. In specific implementation, a data collection module can be set up to collect the environmental data and the device data in real time.

[0073] It can be understood that based on the real-time collected data, the performance of the model and the environmental change situation are monitored. The monitoring metrics are the relevant metrics used during monitoring, including at least the performance metrics and the environmental change metrics. The performance metrics are used to monitor the performance of the model, and the environmental change metrics are used to monitor the environmental change situation.

[0074] It should be understood that according to the performance metrics, it can be determined whether the performance of the model meets the scenario requirements. For example: whether the positioning accuracy reaches the accuracy requirement, and whether the positioning time delay reaches the time delay requirement. According to the environmental change metrics, it can be determined whether the current environment has changed significantly. In specific implementation, data such as positioning accuracy, positioning time delay, positioning error, and model stability can be used as the performance metrics, and the similarity between environments can be used as the environmental change metrics, and no specific limitations are imposed in this embodiment.

[0075] Step S40: When the monitoring metrics meet the adjustment trigger conditions, determine the adaptive adjustment strategy based on the monitoring metrics.

[0076] It should be noted that the adjustment trigger conditions are the conditions for triggering model adjustment. For example: the performance does not meet the standard (the performance does not meet the scenario requirements), and the environment has changed significantly. Among them, the performance not meeting the standard includes the accuracy not meeting the standard (the positioning accuracy does not meet the accuracy requirement) and the time delay not meeting the standard (the positioning time delay does not meet the time delay requirement). Generally speaking, as long as one item is met, it can be considered that the adjustment trigger conditions are met. Exemplarily, if the current accuracy does not meet the standard, the time delay meets the standard, and the environment has not changed significantly, then the adjustment trigger conditions are met. If the current accuracy meets the standard, the time delay meets the standard, and the environment has changed significantly, then the adjustment trigger conditions are met.

[0077] It can be understood that if the adjustment trigger condition is met, the adaptive adjustment of the positioning model is triggered at this time, and a corresponding optimal adjustment strategy, that is, the adaptive adjustment strategy, will be generated to ensure the accuracy and stability of positioning. The adaptive adjustment strategy includes at least model optimization, model switching, and model fallback. Model optimization means optimizing the currently selected positioning model, and there are usually two ways of optimization: updating parameters and retraining. Model switching means switching the currently selected positioning model to other more suitable models. Model fallback means falling back to the traditional positioning method and no longer using the positioning model for positioning. If the adjustment trigger condition is not met, it means that the model can continue to be used, and the positioning model remains unchanged at this time.

[0078] It should be understood that when the environment changes significantly, it means that the scene has changed, and the current model is no longer suitable for continued use. Usually, the model needs to be replaced, and the adaptive adjustment strategy at this time is usually set to model switching. When the accuracy does not meet the standard, it means that the performance of the current model is poor. Usually, it is necessary to further judge the degree of error. If the positioning accuracy is not much different from the accuracy threshold, the currently selected positioning model is fine-tuned and upgraded. The adaptive adjustment strategy at this time can be set to model optimization. If the positioning accuracy is quite different from the accuracy threshold, the adaptive adjustment strategy at this time can be set to model switching. If there is no other model that can be replaced, the adaptive adjustment strategy at this time can be set to model fallback.

[0079] Step S50: Adjust the positioning model based on the adaptive adjustment strategy.

[0080] In a feasible implementation manner, step S50 may include: when the adaptive adjustment strategy is model optimization, optimizing the positioning model, and the optimization is any one of retraining and updating parameters; when the adaptive adjustment strategy is model switching, selecting a target positioning model from the model library based on the environmental data and the scene requirements, and updating the positioning model to the target positioning model; when the adaptive adjustment strategy is model fallback, performing positioning based on a preset positioning strategy.

[0081] It can be understood that if the adaptive adjustment strategy is model optimization, the positioning model can be retrained or the parameters of the positioning model can be updated. If the adaptive adjustment strategy is model switching, according to the current environmental conditions, accuracy requirements, and latency requirements, a more suitable model, that is, the target positioning model, is searched in the model library, and the current positioning model is replaced with the target positioning model. If the adaptive adjustment strategy is model fallback, positioning is performed according to the traditional positioning method (preset positioning strategy).

[0082] In specific implementation, refer to Figure 2, a data collection module, a positioning model library, a model monitoring module, a model decision-making module, and a user interface can be set up. The data collection module is responsible for collecting environmental data, device data, and positioning data; the positioning model library stores multiple positioning models applicable to different scenarios; the model monitoring module is responsible for monitoring metric calculations and measurements; the model decision-making module selects an appropriate model for positioning based on the environmental data and positioning data, monitors the model performance, and performs model optimization or model switching when necessary; the user interface provides a user interaction interface to display the positioning results and decision-making suggestions. The user can view information such as real-time positioning data, model performance metrics, and adjustment records through the graphical interface, and at the same time, a user feedback mechanism is provided, allowing the user to make manual adjustments and settings so that the system can continuously optimize and improve. Among them, the data collection module can be set on the UE (User Equipment) side or the gNB (gNodeB, next-generation base station) side, the model decision-making module can be deployed on the UE side or the LMF (location management function) side, and the model monitoring module can be deployed on the UE / gNB / LMF side.

[0083] It should be understood that by adaptively selecting and optimizing the positioning model, high-precision positioning services can be provided in various complex environments. At the same time, real-time monitoring of the model performance and making decisions can improve the response speed and stability of the positioning service. In addition, by displaying the positioning results and decision-making suggestions through the user interface and providing a user feedback mechanism, users can use the positioning service more conveniently and put forward improvement suggestions.

[0084] Exemplarily, refer to Figure 3, assuming model assistance on the gNB side and indirect positioning using the positioning model, the UE sends an uplink reference signal to the gNB. The gNB estimates the channel information for model training and inference, and the inferred result is sent to the LMF. The LMF compares the historical information with the current information, and calculates and forms a set of monitoring metrics for different scenarios offline in advance. Among them, the metric information for each scenario is a multi-dimensional vector, including RSRP (Reference Signal Received Power), location, speed, RSSI (Received Signal Strength Indicator), etc. Combining the UE positioning scenario, the performance metrics of the current positioning model are compared with the preset threshold to determine whether the error is within the acceptable range. If the condition is met, the model remains unchanged; if the condition is not met and fine-tuning is possible, then fine-tuning is performed. If fine-tuning is not possible, a more suitable model is switched. The LMF feeds back the decision result to the gNB, and the gNB continues positioning according to the model set by the decision result. Among them, the decision results include activation, deactivation, handover, upgrade (if the positioning error exceeds the threshold by a small amount, the current positioning model is fine-tuned and upgraded), fallback (if the positioning error exceeds the threshold by a large amount, fallback to the traditional positioning method), etc. The model identification signaling includes version, identification, and application scenario.

[0085] Exemplarily, referring to Figure 4 , assuming gNB assistance and direct positioning using the positioning model, the LMF requests the positioning accuracy and latency requirements from the UE. The UE reports the positioning accuracy and latency requirements. The UE sends SRS (Sounding Reference Signal) to the gNB. The gNB receives the SRS and measures RSRP, CIR (Channel Impulse Response), etc. The gNB sends the channel or signal measurement values to the LMF for positioning model training and inference. The LMF directly outputs the positioning result. At the same time, the LMF performs monitoring metric calculation and model decision. Assuming the monitoring metrics are latency and model stability, it is judged whether the latency meets the UE's positioning latency requirements. If not, a model update decision is executed. At the same time, it is necessary to judge whether the model stability is normal. The model stability can be judged according to the following method: the positioning error of the positioning model for a stationary UE is within the acceptable range within a certain period of time, or whether the positioning trajectory for a moving UE is smooth within a certain period of time. The LMF can select a suitable model or dynamically update the model according to the positioning accuracy and latency requirements of the UE scenario, improving the UE's positioning accuracy and the adaptability of the positioning model to the scenario.

[0086] This embodiment provides a method for adaptively adjusting a positioning model. Based on initial environmental parameters and scenario requirements, a corresponding positioning model is selected from a model library; positioning is performed based on the positioning model to obtain positioning data; based on the real-time collected environmental data and the positioning data, monitoring metrics are determined, and the monitoring metrics at least include performance metrics and environmental change metrics; when the monitoring metrics meet the adjustment trigger condition, an adaptive adjustment strategy is determined based on the monitoring metrics; and the positioning model is adjusted based on the adaptive adjustment strategy. This embodiment monitors the performance of the positioning model and the current environment in real time. When the model performance deteriorates or the environment changes significantly, the positioning model is adjusted in a timely manner, and the most suitable model is used for positioning, realizing adaptive adjustment based on the monitoring results. It is applicable to various complex scenarios, can improve the positioning accuracy in different scenarios, reduce the computational complexity, and enhance the user experience.

[0087] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar content as that in the above-mentioned first embodiment can be referred to the above introduction and will not be elaborated hereinafter. On this basis, please refer to Figure 5 , step S30 may include steps S301 to S302:

[0088] Step S301, calculating a dynamic cosine similarity based on the environmental data and the reference state data, and using the dynamic cosine similarity as the environmental change metric;

[0089] It should be noted that when the environment changes significantly, it is usually necessary to automatically switch to a model more suitable for the current environment. Exemplarily, in the indoor-outdoor transition scenario at the tunnel entrance and exit, the signal propagation environments inside the tunnel and the open space outside the tunnel are completely different. The signal is blocked by the mountain body, tunnel structure, etc., and the signal weakens or is lost. The positioning model suitable for outdoor needs to be switched to the positioning model suitable for indoor. In addition, the environment also changes with time and seasons, such as road construction, vegetation growth, etc. Therefore, it is necessary to use the environmental change metric to accurately measure the degree of environmental change. In this embodiment, the dynamic cosine similarity is used as the environmental change metric.

[0090] In a feasible implementation manner, the step of calculating the dynamic cosine similarity based on the environmental data and the reference state data may include: performing standardization processing on the environmental data to obtain environmental state data; determining the reference state data based on historical steady-state data; obtaining a first correspondence relationship between the reference state data, the environmental state data, and the dynamic cosine similarity; and obtaining the dynamic cosine similarity based on the reference state data, the environmental state data, and the first correspondence relationship.

[0091] It should be noted that the reference state data, i.e., the reference distribution of environmental parameters under normal conditions, is usually generated by modeling based on historical steady-state data and is used to compare the degree of deviation of the real-time environmental state. Since environmental data has multi-dimensional data, these data (such as temperature, humidity, air pressure, etc.) need to be standardized for subsequent operations. The data after standardization is the environmental state data. The dynamic cosine similarity can characterize the similarity between the environmental state data and the reference state data. The first correspondence relationship among the reference state data, the environmental state data, and the dynamic cosine similarity is the calculation formula of the dynamic cosine similarity, as follows:

[0092]

[0093] In the formula, s t represents the dynamic cosine similarity, v base represents the reference state data, represents the environmental state data. Substituting the relevant data into the above first correspondence relationship, the dynamic cosine similarity can be calculated.

[0094] Furthermore, in a feasible implementation manner, step S40 may include: determining the similarity mean and the similarity standard deviation based on historical similarity data; obtaining the third correspondence relationship among the similarity mean, the similarity standard deviation, and the similarity threshold; obtaining the similarity threshold based on the similarity mean, the similarity standard deviation, and the third correspondence relationship; when the dynamic cosine similarity is less than the similarity threshold, determining that the monitoring index meets the adjustment trigger condition and determining the adaptive adjustment strategy as model switching.

[0095] It should be noted that the similarity threshold is the set threshold of the dynamic cosine similarity and is related to the historical similarity data. The historical similarity data is the historical data of the dynamic cosine similarity within 24 hours. The mean value is calculated based on the historical similarity data to obtain the similarity mean, and the standard deviation is calculated based on the historical similarity data to obtain the similarity standard deviation. The third correspondence relationship among the similarity mean, the similarity standard deviation, and the similarity threshold is the calculation formula of the similarity threshold, as follows:

[0096]

[0097] In the formula, θ s represents the similarity threshold, represents the similarity mean, represents the similarity standard deviation.

[0098] It can be understood that if the dynamic cosine similarity is less than the similarity threshold, it indicates that the degree of environmental change is large, that is, the current environment has changed significantly. At this time, the adjustment trigger condition is met and model switching is required.

[0099] It should be understood that through the dynamic cosine similarity and the 24-hour dynamic similarity threshold, the degree of environmental change can be accurately measured, and it can adapt to periodic environmental changes such as day-night alternation and human flow tides, realizing 24-hour dynamic regulation and enhancing environmental adaptability.

[0100] Step S302: Calculate the positioning error confidence based on the positioning coordinates and the received signal strength corresponding to the environmental data, and use the positioning error confidence and the positioning delay as the performance indicators.

[0101] It should be noted that when the model performance deteriorates, it is usually necessary to optimize the model or switch the model. Therefore, performance indicators are needed to accurately measure the performance status of the model. In this embodiment, the positioning error confidence is used as the performance indicator.

[0102] In a feasible implementation manner, the step of calculating the positioning error confidence based on the positioning coordinates and the received signal strength corresponding to the environmental data may include: obtaining the reference coordinates collected by the measuring device; obtaining the second corresponding relationship between the positioning coordinates, the reference coordinates, the maximum received signal strength, the minimum received signal strength, and the positioning error confidence; and obtaining the positioning error confidence based on the positioning coordinates, the reference coordinates, the maximum received signal strength, the minimum received signal strength, and the second corresponding relationship.

[0103] It should be noted that the received signal strength corresponding to the environmental data is the received signal strength indication value in the current environment. Generally speaking, the received signal strength includes the maximum received signal strength and the minimum received signal strength. The maximum received signal strength is the maximum received signal strength indication value in the current environment, and the minimum received signal strength is the minimum received signal strength indication value in the current environment. The reference coordinate is the true coordinate reference value of the positioning, usually the sub-meter reference position obtained by high-precision measuring devices such as differential base stations and laser trackers. The positioning error confidence is a combined index of the accuracy of the positioning system and the quality of the signal environment. The second corresponding relationship between the positioning coordinates, the reference coordinates, the maximum received signal strength, the minimum received signal strength, and the positioning error confidence is the calculation formula of the positioning error confidence, as shown below:

[0104]

[0105] In the formula, c e represents the positioning error confidence, represents the reference coordinate, represents the positioning coordinate, RSSI max represents the maximum received signal strength, RSSI minIt represents the minimum received signal strength, and n represents the number of coordinates. Substituting the relevant data into the above second corresponding relationship, the positioning error confidence level can be calculated.

[0106] Further, in a feasible implementation manner, step S40 may include: when the positioning error confidence level is less than the first confidence threshold and greater than or equal to the second confidence threshold, determining that the monitoring index meets the adjustment trigger condition, and determining that the adaptive adjustment strategy is model optimization, where the second confidence threshold is less than the first confidence threshold; when the positioning error confidence level is less than the second confidence threshold, determining that the monitoring index meets the adjustment trigger condition, and determining that the adaptive adjustment strategy is any one of model switching and model fallback.

[0107] It should be noted that both the first confidence threshold and the second confidence threshold are set thresholds for the positioning error confidence level, and the second confidence threshold is less than the first confidence threshold. The first confidence threshold and the second confidence threshold can be set according to the accuracy requirements. Exemplarily, the first confidence threshold is set to 0.6, corresponding to a 40% environmental tolerance. When the positioning error confidence level is less than 0.6, that is, the error exceeds the 40% environmental tolerance.

[0108] It can be understood that if the positioning error confidence level is less than the first confidence threshold and greater than or equal to the second confidence threshold, it indicates that the positioning error has exceeded the error threshold corresponding to the accuracy requirement, and the accuracy is not up to standard. At this time, it meets the adjustment trigger condition. Since the positioning error confidence level is not less than the second confidence threshold, it indicates that the positioning error is not much different from the error threshold. At this time, adaptive adjustment can be achieved through model optimization. If the positioning error confidence level is less than the second confidence threshold, it indicates that the positioning error has exceeded the error threshold corresponding to the accuracy requirement, and the accuracy is not up to standard. At this time, it meets the adjustment trigger condition. Since the positioning error confidence level is less than the second confidence threshold, it indicates that the positioning error is much larger than the error threshold. At this time, adaptive adjustment can be achieved through model switching. If a more suitable model cannot be found in the model library, adaptive adjustment can be achieved through model fallback.

[0109] It should be understood that by fusing geometric error and environmental characteristics, a multi-dimensional quantitative evaluation of model performance degradation is achieved. Compared with a single error index, it can reduce the false alarm rate in complex scenarios and effectively improve the monitoring performance of the positioning model.

[0110] This embodiment provides a method for adaptively adjusting a positioning model. Based on environmental data and reference state data, the dynamic cosine similarity is calculated, and the dynamic cosine similarity is used as an environmental change index. Based on the positioning coordinates and the received signal strength corresponding to the environmental data, the positioning error confidence is calculated, and the positioning error confidence and the positioning delay are used as the performance indicators. This embodiment monitors the performance of the positioning model and the current environment in real time. When the model performance deteriorates or the environment changes significantly, the positioning model is adjusted in a timely manner, and the most suitable model is used for positioning, realizing adaptive adjustment based on the monitoring results. It is applicable to various complex scenarios, can improve the positioning accuracy in different scenarios, reduce the computational complexity, and enhance the user experience.

[0111] Exemplarily, to facilitate understanding of the implementation process of the positioning model adaptive adjustment method obtained by combining this embodiment with the above-mentioned Embodiment 2, please refer to Figure 6 , Figure 6 A brief flow schematic diagram of a method for adaptively adjusting a positioning model is provided. Specifically:

[0112] The model generation module is responsible for generating a positioning model according to the positioning scenario requirements or updating the positioning model according to the model update instruction information of the model management module; the positioning model library is responsible for storing all positioning models; the model management module is the bridge between the model decision module and the positioning model library, responsible for managing and receiving the transmitted model decision results. If a model switch is required, a model switch request is sent to the positioning model library. If a model update is required, a model update instruction information is sent to the model generation module, and the updated model is received, and the model update result is sent to the local model monitoring module; the local model monitoring module monitors the performance of the model and the environmental state according to the environmental parameters and positioning data, and feeds back the monitoring indicators to the model decision module; the user interface can feed back user suggestions to the model decision module; the model decision module comprehensively gives the model decision result based on the monitoring indicators of the local model monitoring module, user suggestions or monitoring indicators of other positioning modes.

[0113] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the positioning model adaptive adjustment method of this application. Based on this technical concept, more forms of simple transformations are within the protection scope of this application.

[0114] This application also provides a positioning model adaptive adjustment system. Please refer to Figure 7 , the positioning model adaptive adjustment system includes:

[0115] The model decision module 10 is used to select a corresponding positioning model in the model library based on the initial environmental parameters and scenario requirements;

[0116] The model monitoring module 20 is used to perform positioning based on the positioning model to obtain positioning data;

[0117] The model monitoring module 20 is further configured to determine monitoring metrics based on the real-time collected environmental data and the positioning data, where the monitoring metrics at least include performance metrics and environmental change metrics;

[0118] The model decision-making module 10 is further configured to determine an adaptive adjustment strategy based on the monitoring metrics when the monitoring metrics meet the adjustment trigger condition;

[0119] The model decision-making module 10 is further configured to adjust the positioning model based on the adaptive adjustment strategy.

[0120] In a feasible implementation manner, the environmental data at least includes temperature, humidity, and air pressure, the positioning data at least includes positioning coordinates and positioning delay, and the model monitoring module 20 is further configured to calculate a dynamic cosine similarity based on the environmental data and reference state data, and use the dynamic cosine similarity as the environmental change metric;

[0121] Calculate a positioning error confidence level based on the positioning coordinates and the received signal strength corresponding to the environmental data, and use the positioning error confidence level and the positioning delay as the performance metrics.

[0122] In a feasible implementation manner, the model monitoring module 20 is further configured to perform normalization processing on the environmental data to obtain environmental state data;

[0123] Determine the reference state data based on historical steady-state data;

[0124] Obtain a first correspondence relationship among the reference state data, the environmental state data, and the dynamic cosine similarity;

[0125] Obtain the dynamic cosine similarity based on the reference state data, the environmental state data, and the first correspondence relationship.

[0126] In a feasible implementation manner, the received signal strength includes the maximum received signal strength and the minimum received signal strength, and the model monitoring module 20 is further configured to obtain the reference coordinates collected by the measurement device;

[0127] Obtain a second correspondence relationship among the positioning coordinates, the reference coordinates, the maximum received signal strength, the minimum received signal strength, and the positioning error confidence level;

[0128] Obtain the positioning error confidence level based on the positioning coordinates, the reference coordinates, the maximum received signal strength, the minimum received signal strength, and the second correspondence relationship.

[0129] In a feasible implementation manner, the decision-making module 10 is further configured to determine a similarity mean value and a similarity standard deviation based on historical similarity data;

[0130] Obtain a third correspondence relationship among the similarity mean value, the similarity standard deviation, and a similarity threshold;

[0131] Based on the similarity mean value, the similarity standard deviation, and the third correspondence relationship, obtain the similarity threshold;

[0132] When the dynamic cosine similarity is less than the similarity threshold, determine that the monitoring index meets the adjustment trigger condition, and determine that the adaptive adjustment strategy is model switching.

[0133] In a feasible implementation manner, the decision-making module 10 is further configured to, when the positioning error confidence is less than a first confidence threshold and greater than or equal to a second confidence threshold, determine that the monitoring index meets the adjustment trigger condition, and determine that the adaptive adjustment strategy is model optimization, where the second confidence threshold is less than the first confidence threshold;

[0134] When the positioning error confidence is less than the second confidence threshold, determine that the monitoring index meets the adjustment trigger condition, and determine that the adaptive adjustment strategy is any one of model switching and model fallback.

[0135] In a feasible implementation manner, the decision-making module 10 is further configured to, when the adaptive adjustment strategy is model optimization, optimize the positioning model, where the optimization is any one of retraining and updating parameters;

[0136] When the adaptive adjustment strategy is model switching, select a target positioning model from the model library based on the environmental data and the scenario requirements, and update the positioning model to the target positioning model;

[0137] When the adaptive adjustment strategy is model fallback, perform positioning based on a preset positioning strategy.

[0138] The positioning model adaptive adjustment system provided by this application adopts the positioning model adaptive adjustment method in the above embodiment, and can solve the technical problem that a static monitoring strategy is difficult to dynamically adjust the positioning model in a timely manner according to scenario changes, affecting the positioning accuracy. Compared with the prior art, the beneficial effects of the positioning model adaptive adjustment system provided by this application are the same as those of the positioning model adaptive adjustment method provided by the above embodiment, and other technical features in the positioning model adaptive adjustment system are the same as the features disclosed in the method of the above embodiment, and will not be elaborated here.

[0139] The present application provides a positioning model adaptive adjustment device, and the positioning model adaptive adjustment device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the positioning model adaptive adjustment method in the first embodiment above.

[0140] Reference is made below to Figure 8 , which shows a schematic structural diagram of a positioning model adaptive adjustment device suitable for implementing the embodiments of the present application. The positioning model adaptive adjustment device in the embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions: tablet computers), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 8 The shown positioning model adaptive adjustment device is merely an example and should not impose any limitation on the functions and usage scope of the embodiments of the present application.

[0141] As Figure 8As shown in the figure, the positioning model adaptive adjustment device may include a processing system 1001 (such as a central processing unit, a graphics processing unit, etc.), which may perform various appropriate actions and processes according to the program stored in the ROM (Read Only Memory) 1002 or the program loaded from the storage system 1003 into the RAM (Random Access Memory) 1004. In the RAM 1004, various programs and data required for the operation of the positioning model adaptive adjustment device are also stored. The processing system 1001, the ROM 1002, and the RAM 1004 are connected to each other through a bus 1005. The input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems may be connected to the I / O interface 1006: an input system 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output system 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage system 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication system 1009. The communication system 1009 may allow the positioning model adaptive adjustment device to communicate with other devices wirelessly or wiredly to exchange data. Although the figure shows a positioning model adaptive adjustment device with various systems, it should be understood that it is not required to implement or have all the shown systems. More or fewer systems may be implemented or had alternatively.

[0142] Specifically, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart may be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes program codes for performing the method shown in the flowchart. In such an embodiment, the computer program may be downloaded and installed from the network through the communication system, or installed from the storage system 1003, or installed from the ROM 1002. When the computer program is executed by the processing system 1001, the above-mentioned functions defined in the method of the embodiments disclosed in the present application are executed.

[0143] The positioning model adaptive adjustment device provided in the present application adopts the positioning model adaptive adjustment method in the above embodiments, and can solve the technical problem that the static monitoring strategy is difficult to dynamically adjust the positioning model in a timely manner according to the scene change, which affects the positioning accuracy. Compared with the prior art, the beneficial effects of the positioning model adaptive adjustment device provided in the present application are the same as those of the positioning model adaptive adjustment method provided in the above embodiments, and other technical features in the positioning model adaptive adjustment device are the same as those disclosed in the method of the previous embodiment, and will not be elaborated here.

[0144] It should be understood that each part disclosed in this application can be implemented by hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in a suitable manner in any one or more embodiments or examples.

[0145] The above are only specific embodiments of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0146] This application provides a computer-readable storage medium with computer-readable program instructions (i.e., computer programs) stored thereon, and the computer-readable program instructions are used to execute the positioning model adaptive adjustment method in the above embodiments.

[0147] The computer-readable storage medium provided by this application can be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) or flash memory, optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In this embodiment, the computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or combined with an instruction execution system, device, or device. The program code contained on the computer-readable storage medium can be transmitted by any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0148] The above computer-readable storage medium can be included in the positioning model adaptive adjustment device; or it can exist alone without being assembled into the positioning model adaptive adjustment device.

[0149] The above computer-readable storage medium carries one or more programs, which, when executed by the positioning model adaptive adjustment device, cause the positioning model adaptive adjustment device to: select a corresponding positioning model from the model library based on the initial environmental parameters and the scenario requirements; perform positioning based on the positioning model to obtain positioning data; determine monitoring metrics based on the real-time collected environmental data and the positioning data, where the monitoring metrics at least include performance metrics and environmental change metrics; determine an adaptive adjustment strategy based on the monitoring metrics when the monitoring metrics meet the adjustment trigger condition; and adjust the positioning model based on the adaptive adjustment strategy.

[0150] Computer program code for performing the operations of this application can be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0151] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment, or a part of code that contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the boxes can occur in a different order than that marked in the accompanying drawings. For example, two consecutive boxes shown can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0152] The modules involved in the embodiments of the present application can be implemented in software or in hardware. Among them, the name of the module does not constitute a limitation to the unit itself in some cases.

[0153] The readable storage medium provided by the present application is a computer-readable storage medium, and the computer-readable storage medium stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned positioning model adaptive adjustment method, which can solve the technical problem that the static monitoring strategy is difficult to dynamically adjust the positioning model in a timely manner according to the scene change, affecting the positioning accuracy. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as those of the positioning model adaptive adjustment method provided by the above embodiments, and will not be elaborated here.

[0154] The present application also provides a computer program product, including a computer program, and when the computer program is executed by a processor, it realizes the steps of the positioning model adaptive adjustment method as described above.

[0155] The computer program product provided by the present application can solve the technical problem that the static monitoring strategy is difficult to dynamically adjust the positioning model in a timely manner according to the scene change, affecting the positioning accuracy. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as those of the positioning model adaptive adjustment method provided by the above embodiments, and will not be elaborated here.

[0156] The above are only some embodiments of the present application, and do not limit the patent scope of the present application. Any equivalent structural transformation made by using the content of the specification and drawings of the present application under the technical concept of the present application, or any direct / indirect application in other related technical fields, is included in the patent protection scope of the present application.

Claims

1. A method for adaptively adjusting a positioning model, characterized in that The method described above includes: Based on the initial environmental parameters and scene requirements, select the corresponding positioning model in the model library; Perform positioning based on the positioning model to obtain positioning data; Based on the real-time collected environmental data and the positioning data, determine the monitoring indicators, where the monitoring indicators at least include performance indicators and environmental change indicators; When the monitoring indicators meet the adjustment trigger condition, determine the adaptive adjustment strategy based on the monitoring indicators; Based on the adaptive adjustment strategy, adjust the positioning model.

2. The method according to claim 1, characterized in that The environmental data at least includes temperature, humidity, and air pressure, and the positioning data at least includes positioning coordinates and positioning delay. The step of determining the monitoring indicators based on the real-time collected environmental data and the positioning data includes: Calculate the dynamic cosine similarity based on the environmental data and the reference state data, and use the dynamic cosine similarity as the environmental change indicator; Calculate the positioning error confidence based on the positioning coordinates and the received signal strength corresponding to the environmental data, and use the positioning error confidence and the positioning delay as the performance indicators.

3. The method according to claim 2, wherein The step of calculating the dynamic cosine similarity based on the environmental data and the reference state data includes: Perform standardization processing on the environmental data to obtain environmental state data; Based on the historical steady-state data, determine the reference state data; Obtain the first correspondence between the reference state data, the environmental state data, and the dynamic cosine similarity; Based on the reference state data, the environmental state data, and the first correspondence, obtain the dynamic cosine similarity.

4. The method according to claim 2, wherein The received signal strength includes the maximum received signal strength and the minimum received signal strength. The step of calculating the positioning error confidence based on the positioning coordinates and the received signal strength corresponding to the environmental data includes: Obtain the reference coordinates collected by the measuring device; Obtain the second correspondence between the positioning coordinates, the reference coordinates, the maximum received signal strength, the minimum received signal strength, and the positioning error confidence; Based on the positioning coordinates, the reference coordinates, the maximum received signal strength, the minimum received signal strength, and the second correspondence, obtain the positioning error confidence.

5. The method according to claim 2, wherein The step of determining the adaptive adjustment strategy based on the monitoring indicators when the monitoring indicators meet the adjustment trigger condition includes: Based on the historical similarity data, determine the similarity mean and the similarity standard deviation; Obtain the third correspondence between the similarity mean, the similarity standard deviation, and the similarity threshold; Based on the similarity mean, the similarity standard deviation, and the third correspondence, obtain the similarity threshold; When the dynamic cosine similarity is less than the similarity threshold, determine that the monitoring indicators meet the adjustment trigger condition, and determine the adaptive adjustment strategy as model switching.

6. The method according to claim 2, wherein The step of determining the adaptive adjustment strategy based on the monitoring indicators when the monitoring indicators meet the adjustment trigger condition includes: When the confidence level of the positioning error is less than the first confidence threshold and greater than or equal to the second confidence threshold, it is determined that the monitoring metric meets the adjustment trigger condition, and the adaptive adjustment strategy is determined to be model optimization, where the second confidence threshold is less than the first confidence threshold; When the confidence level of the positioning error is less than the second confidence threshold, it is determined that the monitoring metric meets the adjustment trigger condition, and the adaptive adjustment strategy is determined to be either model switching or model fallback.

7. The method according to any one of claims 1 to 6, characterized in that, The step of adjusting the positioning model based on the adaptive adjustment strategy includes: When the adaptive adjustment strategy is model optimization, the positioning model is optimized, and the optimization is either retraining or updating parameters; When the adaptive adjustment strategy is model switching, based on the environmental data and the scenario requirements, a target positioning model is selected from the model library, and the positioning model is updated to the target positioning model; When the adaptive adjustment strategy is model fallback, positioning is performed based on a preset positioning strategy.

8. An adaptive adjustment system for a positioning model, characterized in that The system includes: A model decision module for selecting a corresponding positioning model from the model library based on initial environmental parameters and scenario requirements; A model monitoring module for performing positioning based on the positioning model to obtain positioning data; The model monitoring module is further configured to determine a monitoring metric based on the real-time collected environmental data and the positioning data, where the monitoring metric at least includes a performance metric and an environmental change metric; The model decision module is further configured to determine an adaptive adjustment strategy based on the monitoring metric when the monitoring metric meets the adjustment trigger condition; The model decision module is further configured to adjust the positioning model based on the adaptive adjustment strategy.

9. A positioning model adaptive adjustment device, characterized in that, The device includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the computer program is configured to implement the steps of the positioning model adaptive adjustment method according to any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium, and when the computer program is executed by the processor, it implements the steps of the positioning model adaptive adjustment method according to any one of claims 1 to 7.

Images