Bluetooth+uwb+rtk fusion positioning method and system

Abstract

The application discloses a Bluetooth+UWB+RTK fusion positioning method and system, relates to the technical field of multi-source fusion positioning and intelligent safety management and control, and comprises the following steps: collecting characteristic data of an environment in which a positioning object is located to obtain environment characteristic data; performing pretreatment to obtain pretreated environment characteristic data; judging a space scene in which the positioning object is located to obtain a space scene type; setting a fusion strategy parameter; collecting Bluetooth ranging data, UWB positioning data, RTK positioning data and inertial navigation data according to the fusion strategy parameter; performing fusion calculation to obtain real-time three-dimensional high-precision positioning data; and monitoring the position and trajectory of power grid workers in real time by using the real-time three-dimensional high-precision positioning data to detect the occurrence of an emergency event. The application improves the positioning precision, trajectory continuity and anti-interference capability in a complex scene, significantly improves the abnormal condition detection efficiency and alarm response speed of power grid workers, and provides technical support for power grid safety production and personnel protection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of multi-source fusion positioning and intelligent safety management and control, in particular to a Bluetooth+UWB+RTK fusion positioning method and system. BACKGROUND

[0002] In recent years, with the rapid development of Internet of Things, wireless communication and high-precision positioning technology, multi-source fusion-based positioning methods have gradually become a research hotspot in the fields of intelligent power, industrial Internet and emergency management. Bluetooth, Ultra-Wideband (UWB) and Real-Time Kinematic (RTK) of Global Navigation Satellite System (GNSS) and other technologies have been widely used in indoor and outdoor human and object positioning tasks. The above technologies each have advantages in specific scenarios, such as Bluetooth being suitable for low-power large-scale control, UWB providing centimeter-level high-precision positioning, and RTK being able to achieve sub-meter to centimeter-level positioning in a wide area. However, in actual applications, a single technology often cannot meet the higher requirements for positioning accuracy, continuity and real-time in large-scale complex scenarios. To improve the intelligence and universality of the positioning system, the academic and industrial communities continue to promote the fusion and dynamic optimization of multi-source heterogeneous positioning data.

[0003] However, the existing positioning technology system still has many shortcomings. First, existing multi-mode fusion methods are often based on static weights or simple scene discrimination, which makes it difficult to achieve intelligent adaptive switching of positioning modes according to changing environments such as power plant areas and industrial sites, resulting in unstable positioning continuity and accuracy. Second, existing technologies are mostly limited to planar (two-dimensional) positioning, and lack high-precision three-dimensional fusion compensation capabilities based on multi-mode features for three-dimensional complex spaces such as tall buildings and cross-floor spaces, which can easily cause floor misjudgment or height drift. In addition, traditional fusion algorithms respond slowly to environmental changes, signal blockage and abnormal motion, which can easily cause false positives and false negatives in positioning and other safety hazards. In scenarios such as power grids that require high reliability and real-time warning, simply relying on a certain positioning source or a fusion method lacking dynamic adaptive mechanisms cannot achieve accurate control and efficient response to the location, trajectory and sudden events of workers.

[0004] Therefore, there is an urgent need for a personnel safety positioning and control method that can fully perceive the environmental state, intelligently identify the spatial scene, and dynamically optimize the participation weights and algorithm strategies of various positioning technologies for different scenarios, thereby achieving high-precision and intelligent personnel safety positioning and control in all spaces and time periods. SUMMARY

[0005] In view of the above problems, the present application is proposed.

[0006] Therefore, the technical problem solved by the present application is that the existing multi-source fusion positioning method has the problems of being difficult to realize intelligent scene recognition and dynamic weight adjustment in a complex environment, lacking a high-precision fusion compensation mechanism for three-dimensional space, being insufficient in positioning continuity and safety warning response capability, and how to realize global high-precision intelligent positioning of personnel and real-time safety response to emergencies in a complex power grid scene.

[0007] To solve the above technical problems, the present application provides the following technical solutions.

[0008] In a first aspect, the present application provides a Bluetooth+UWB+RTK fusion positioning method, comprising: collecting feature data of an environment in which a positioning object is located to obtain environment feature data;

[0009] Preprocessing the environment feature data to obtain preprocessed environment feature data;

[0010] Determining a spatial scene in which the positioning object is located based on the preprocessed environment feature data to obtain a spatial scene type;

[0011] Setting fusion strategy parameters of Bluetooth, UWB, RTK and inertial navigation according to the spatial scene type;

[0012] Collecting Bluetooth ranging data, UWB positioning data, RTK positioning data and inertial navigation data according to the fusion strategy parameters;

[0013] Fusion calculating the Bluetooth ranging data, UWB positioning data, RTK positioning data and inertial navigation data to obtain real-time three-dimensional high-precision positioning data;

[0014] Using the real-time three-dimensional high-precision positioning data to monitor the position and trajectory of power grid personnel in real time, and realizing alarm response when a sudden event is detected.

[0015] As a preferred scheme of the Bluetooth+UWB+RTK fusion positioning method, the collecting of the feature data of the environment in which the positioning object is located comprises collecting acceleration data, angular velocity data, air pressure data, Bluetooth signal strength data, UWB signal quality data, GNSS satellite signal strength data and illumination data of the current position of the positioning object, and combining to obtain the environment feature data.

[0016] As a preferred scheme of the Bluetooth+UWB+RTK fusion positioning method, the preprocessing of the environment feature data comprises denoising processing, normalization processing, time alignment and missing value completion processing of the environment feature data to obtain the preprocessed environment feature data.

[0017] As a preferred scheme of the Bluetooth+UWB+RTK fusion positioning method, the method comprises the following steps: inputting the preprocessed environment feature data into a preset scene discrimination model to discriminate the space scene where the positioning object is located, and outputting the corresponding space scene type.

[0018] As a preferred scheme of the Bluetooth+UWB+RTK fusion positioning method, the method comprises the following steps: inputting the preprocessed environment feature data into a preset scene discrimination model to discriminate the space scene where the positioning object is located, and outputting the corresponding space scene type.

[0019] As a preferred scheme of the Bluetooth+UWB+RTK fusion positioning method, the method comprises the following steps: inputting the preprocessed environment feature data into a preset scene discrimination model to discriminate the space scene where the positioning object is located, and outputting the corresponding space scene type.

[0020] As a preferred scheme of the Bluetooth+UWB+RTK fusion positioning method, the method comprises the following steps: inputting the preprocessed environment feature data into a preset scene discrimination model to discriminate the space scene where the positioning object is located, and outputting the corresponding space scene type.

[0021] As a preferred scheme of the Bluetooth+UWB+RTK fusion positioning method, the method comprises the following steps: inputting the preprocessed environment feature data into a preset scene discrimination model to discriminate the space scene where the positioning object is located, and outputting the corresponding space scene type.

[0022] As a preferred scheme of the Bluetooth+UWB+RTK fusion positioning method, the method comprises the following steps: inputting the preprocessed environment feature data into a preset scene discrimination model to discriminate the space scene where the positioning object is located, and outputting the corresponding space scene type.

[0023] As a preferred scheme of the Bluetooth+UWB+RTK fusion positioning method, the method comprises the following steps: inputting the preprocessed environment feature data into a preset scene discrimination model to discriminate the space scene where the positioning object is located, and outputting the corresponding space scene type.

[0024] As a preferred scheme of the Bluetooth+UWB+RTK fusion positioning method, the real-time monitoring of the position and trajectory of the power grid worker comprises continuously comparing the current position of the power grid worker with the preset safety region boundary, historical trajectory and operation specification requirement based on the real-time three-dimensional high-precision positioning data, and judging whether there is an abnormal situation such as boundary crossing, long-time static, personnel gathering, abnormal motion trajectory and entering a dangerous region.

[0025] When any abnormal situation is detected, an alarm response is automatically performed and pushed to the monitoring terminal.

[0026] In a second aspect, the embodiments of the present application provide a Bluetooth+UWB+RTK fusion positioning system, comprising:

[0027] An environment feature data acquisition module acquires feature data of an environment where a positioning object is located to obtain environment feature data.

[0028] A preprocessing module pre-processes the environment feature data to obtain pre-processed environment feature data.

[0029] A space scene discrimination module discriminates a space scene where the positioning object is located based on the pre-processed environment feature data to obtain a space scene type.

[0030] A fusion strategy parameter setting module sets fusion strategy parameters of Bluetooth, UWB, RTK and inertial navigation according to the space scene type.

[0031] A multi-source data acquisition module acquires Bluetooth ranging data, UWB positioning data, RTK positioning data and inertial navigation data according to the fusion strategy parameters.

[0032] A multi-source fusion calculation module performs fusion calculation on the Bluetooth ranging data, UWB positioning data, RTK positioning data and inertial navigation data to obtain real-time three-dimensional high-precision positioning data.

[0033] A trajectory monitoring and alarm module uses the real-time three-dimensional high-precision positioning data to perform real-time monitoring of the position and trajectory of the power grid worker, and realizes an alarm response when a sudden event is detected.

[0034] The application has the advantages that the application realizes high-precision three-dimensional positioning in full space and full scene under complex power grid environment by fusing multiple positioning data sources of Bluetooth, UWB, RTK and inertial navigation, combining intelligent discrimination of environmental characteristics and dynamic weight adaptive strategy. Compared with the prior art, the application can not only significantly improve the positioning continuity and accuracy of indoor and outdoor cross-regional and cross-floor positioning, but also can intelligently switch the optimal positioning strategy according to the actual scene changes, effectively reducing the errors caused by signal shielding, multipath interference and the like. Through real-time monitoring and trajectory analysis, sudden events can be quickly found and responded, and the work safety and emergency disposal ability of power grid workers are improved. The method has the advantages of flexible deployment, strong expansibility, high adaptability and stable positioning accuracy, and can be widely applied to the fields of intelligent power, industrial safety and intelligent positioning management of large-scale complex scenes. BRIEF DESCRIPTION OF DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can be obtained without creative labor based on these drawings.

[0036] Figure 1 The application provides a Bluetooth+UWB+RTK fusion positioning method. DETAILED DESCRIPTION

[0037] In order to make the above-mentioned purposes, features and advantages of the application more apparent and easy to understand, the specific embodiments of the application will be described in detail below with reference to the drawings. Obviously, the described embodiments are only some embodiments of the application, not all embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor should be within the protection scope of the application.

[0038] Embodiment 1, refer to Figure 1 For an embodiment of the application, a Bluetooth+UWB+RTK fusion positioning method is provided, which comprises:

[0039] S1: collecting characteristic data of an environment where a positioning object is located to obtain environmental characteristic data.

[0040] Acceleration data, angular velocity data, air pressure data, Bluetooth signal strength data, UWB signal quality data, GNSS satellite signal strength data and illumination data of the current position of the positioning object are collected and combined to obtain the environmental characteristic data.

[0041] In step S1, the process involves multiple types of original environment perception data. The "acceleration data" and "angular velocity data" respectively specify the linear acceleration and rotational angular velocity of the positioning object at each moment in three-dimensional space, which can be obtained by real-time measurement of the accelerometer and gyroscope in the inertial measurement element (IMU), and can efficiently reflect the instantaneous motion state and dynamic change trend of the positioning object. The "air pressure data" is used to perceive the height change of the space where the object is located, which can realize accurate identification of scenes such as up and down floors, stairs, elevators, etc., and is one of the core auxiliary information for three-dimensional space positioning. The "Bluetooth signal strength data" reflects the distance change between the positioning object and the surrounding Bluetooth beacons, providing effective support for coarse-grained or auxiliary positioning. The "UWB signal quality data" evaluates the smoothness of the wireless environment and the multipath effect by analyzing the received strength and arrival time of the ultra-wideband signal, which helps to distinguish between occluded scenes and high-precision visible scenes, and improves the adaptability and anti-interference ability of positioning fusion. The "GNSS satellite signal strength data" can reflect the object's ability to receive Beidou / GPS satellite signals in outdoor space, which helps to judge whether the object is in an open environment or is structurally occluded, and provides a basis for dynamically switching positioning modes. The "lighting data" can be collected by light-sensitive elements and used to assist in determining whether the object is in a bright or dark environment or whether it has entered a closed or semi-closed space.

[0042] The above seven types of data are combined to form a multi-modal environment feature data set, achieving comprehensive perception of the space environment. Compared with relying on only a single sensing source (such as relying only on inertia or only on wireless signals), the present scheme achieves multi-dimensional collaborative perception, which not only effectively identifies the complex spatial scenes (such as indoor / outdoor, different floors, occluded / non-occluded) in which the positioning object is located, but also captures dynamic environmental changes (such as entering special areas such as elevators, corridors, tunnels) in real time. This all-factor environment perception mechanism lays a foundation for subsequent intelligent discrimination of spatial scenes, adaptive adjustment of positioning source weights, and reliability of multi-source fusion algorithms. Specifically, relying only on IMU inertial information is prone to cumulative drift errors, and relying only on Bluetooth or UWB signals is severely affected by occlusion. However, the fusion of acceleration, angular velocity, air pressure, and multiple wireless signal quality and lighting level data can dynamically improve the environmental adaptability and fault robustness of the positioning system in actual engineering scenarios, effectively ensuring the sustainability and consistency of high-precision positioning results in large-scale and variable environments.

[0043] S2: Preprocess the environment feature data to obtain preprocessed environment feature data.

[0044] The environment feature data is denoised, normalized, time-aligned, and missing value completed to obtain preprocessed environment feature data.

[0045] In step S2, the "preprocessing of the environmental feature data to obtain preprocessed environmental feature data" includes a series of signal processing operations on the collected acceleration data, angular velocity data, air pressure data, Bluetooth signal strength data, UWB signal quality data, GNSS satellite signal strength data and illumination data, specifically including denoising processing, normalization processing, time alignment and missing value completion.

[0046] Among them, "denoising processing" refers to filtering out high-frequency noise and outliers in the original environmental feature data due to electromagnetic interference, device jitter, occasional failure and the like, and methods such as moving average, low-pass filtering, Kalman filtering can be used to effectively improve the stability and reliability of the data. "Normalization processing" refers to mapping data of different sources and different dimensions to the same numerical interval (such as 0~1), eliminating the weight bias of each feature in the subsequent discrimination and fusion algorithm, so that acceleration, signal strength and other features have comparability and synergy in the same fusion framework. "Time alignment" refers to the orderly alignment of multi-source data with different collection timestamps or inconsistent sampling frequencies at the same time through interpolation, synchronous sampling and other methods, to ensure the time sequence consistency of subsequent multi-modal analysis and the integrity of scene description. "Missing value completion processing" is used to correct data gaps caused by short-time signal loss, sensor anomalies and the like in the actual collection process, using linear interpolation, nearest neighbor interpolation or model prediction and other means to ensure the completeness of the feature data in each time window and improve the robustness of the algorithm.

[0047] The preprocessing step plays a key role in actual system deployment. On the one hand, through denoising and missing value completion, the realistic pain points of signal distortion and interruption in complex power / industrial environments are effectively overcome, ensuring the quality of the basic data for subsequent spatial scene discrimination and multi-source fusion; on the other hand, normalization and time alignment ensure that environmental features from different sensor sources have comparability and spatio-temporal synergy, greatly reducing the risk of fusion errors caused by sensor precision differences or collection delays. Compared with the common raw data direct input or single preprocessing method of the prior art, the multi-link, multi-method comprehensive preprocessing of the present application greatly improves the reliability and utilization efficiency of the environmental feature data, providing a solid foundation for subsequent intelligent discrimination of complex scenes and adaptive adjustment of fusion strategies, thereby supporting the entire system to achieve the goal of full-scene high-precision personnel positioning and safety control.

[0048] S3: discriminating the spatial scene where the positioning object is located based on the preprocessed environmental feature data, to obtain a spatial scene type.

[0049] The preprocessed environmental feature data is input into a preset scene discrimination model to discriminate that the positioning object is currently in an indoor ordinary area, an indoor high-precision area, an outdoor open area or a floor transition area, and output the corresponding spatial scene type.

[0050] In this step, the so-called "space scene type" refers to the subdivision of the current environment of the positioning object, for example: indoor general area (such as large factory general operation area), indoor high-precision area (such as precision equipment area that needs accurate operation or high risk), outdoor open area (such as factory road, open equipment area), floor change area (such as stairs, elevator room, slope, vertical motion scene such as grade separation).

[0051] The preprocessed environment feature data obtained in step S2 is input into a preset scene discrimination model, and the model automatically completes environment classification according to the spatio-temporal distribution and statistical characteristics of various features such as acceleration, air pressure, and signal strength. For example, air pressure mutation and vertical direction change of acceleration often indicate floor change area, strong GNSS signal and weak UWB / Bluetooth can be determined as outdoor, and vice versa as indoor. This automatic discrimination realizes instant recognition of the environment type in complex scenes, avoids the problems of traditional methods such as manual threshold setting and scene switching delay, and provides scientific basic data for the downstream adaptive fusion strategy.

[0052] The preset scene discrimination model includes using the labeled historical preprocessed environment feature data and the actual space scene type to train through a machine learning algorithm to establish a classification model that can output the corresponding space scene type based on the input preprocessed environment feature data.

[0053] Further, the "space scene type classification model" uses a "rule screening + supervised learning" two-level judgment process to discriminate the space scene type. With a two-second length and a zero-point-five-second step sliding time window, the preprocessed multi-source features are extracted as model input, including: the mean and variance of three-axis acceleration and angular velocity in the time window, the mean of vertical acceleration, the discrimination score of static or zero speed; the mean of air pressure, the air pressure difference within three seconds and the air pressure change rate; the stability of ultra-wideband time of arrival estimation, the line of sight confidence obtained by the ratio of first path to multipath energy, the variance of ranging residual; the number of visible satellites, the median value of carrier-to-noise ratio, horizontal dilution of precision and position dilution of precision, whether the differential correction is available and the stability of baseline solution; the number of visible Bluetooth beacons, the mean and variance of signal strength, and the adjacent beacon density index; the mean and fluctuation amplitude of illumination. The first-level judgment is a rule-based fast screening, for example, when the number of visible satellites is not less than eight and the horizontal dilution of precision is not greater than one point five, it is preferentially marked as an outdoor candidate; when the three-second air pressure difference is not less than zero point six hundred pascals and the vertical acceleration variance significantly increases, it is marked as a floor change candidate. The second-level judgment uses a supervised classifier (such as random forest, about two hundred trees, maximum depth six, and class weight balance) trained by labeled data to output the probability distribution of outdoor open, indoor general, indoor high-precision, and floor change, and set a hysteresis condition that requires three consecutive time windows to be consistent to allow switching, to avoid category jitter.

[0054] Further, the corresponding relationship between the features and the scenarios is that: outdoor open is usually represented by not less than eight visible satellites, not greater than 1.5 horizontal dilution of precision, not less than 35 decibel-hertz median value of carrier-to-noise ratio, and less number of visible Bluetooth beacons and less than indoor typical value of ultra-wideband line-of-sight confidence; indoor ordinary is usually represented by not less than five visible Bluetooth beacons, available ultra-wideband and moderate line-of-sight confidence, and unstable satellite positioning (for example, not more than four visible satellites or greater than 3 horizontal dilution of precision); indoor high precision is usually represented by not less than 0.7 ultra-wideband line-of-sight confidence and low ranging residual variance, not less than eight visible Bluetooth beacons, and basically unavailable satellite positioning; floor transition is usually represented by not less than 0.6 hundred pascal of air pressure difference within three seconds, and variance of vertical acceleration increasing by about 50% compared with the previous time window and accompanied by monotonous change of height or short-time decrease of horizontal speed. The above rules are used as the basis for the first-level rapid screening, and also provide interpretable prior constraints for the supervised learning model.

[0055] It should also be noted that the "preset scene discrimination model" refers to using a large amount of historical collected labeled data (i.e., each set of feature data corresponds to its real scene type) as a training set, training through a machine learning algorithm, and finally forming an automatic classifier with input of multi-dimensional environmental features and output of spatial scene type. In specific implementation, methods such as decision tree, support vector machine (SVM), random forest, and neural network can be selected; among them, the decision tree can realize rule interpretation based on feature segmentation, and the neural network is suitable for nonlinear feature extraction in complex environments with large samples. The training process optimizes the classification accuracy by repeatedly adjusting the model parameters, and finally obtains a discrimination model with strong generalization ability and good adaptability to new environments. Unlike traditional rule judgment or single feature threshold method, this method can learn the discrimination logic of complex scenes from the global multi-modal features, effectively avoid misjudgment when a single feature is unstable or abnormal, and greatly improve the accuracy of system environmental adaptation and spatial type discrimination.

[0056] S4: setting the fusion strategy parameters of Bluetooth, UWB, RTK, and inertial navigation according to the spatial scene type;

[0057] According to the spatial scene type obtained by discrimination, the weight parameters of the positioning data of Bluetooth, UWB, RTK, and inertial navigation in fusion calculation are dynamically allocated, the data source type participating in fusion calculation and its priority are determined, and the acquisition frequency and activation state of each data source are set.

[0058] In step S4, "setting the fusion strategy parameters of Bluetooth, UWB, RTK, and inertial navigation according to the spatial scene type", it involves multiple key technical links such as allocation of fusion weight parameters of multi-source positioning data, determination of data source participation type, setting of acquisition frequency, and switching of activation state.

[0059] Among them, the "fusion weight parameter" refers to assigning different numerical weights to the four types of positioning data of Bluetooth, UWB, RTK and inertial navigation in subsequent fusion calculation according to the currently judged spatial scene type (such as indoor ordinary, indoor high precision, outdoor open, floor transformation, etc.). For example, in the outdoor open area, since the GNSS / RTK signal is strong, the RTK data weight is significantly improved, and the UWB and Bluetooth data can be automatically reduced in weight or even ignored; while in the indoor high precision area, the UWB data weight is improved, and inertial navigation and Bluetooth assistance are appropriately introduced. For floor transformation or signal shielding scenes, the participation of inertial navigation and barometric data is strengthened to compensate for the failure of traditional wireless positioning in vulnerable areas.

[0060] Specifically, the weight parameters of each positioning data in the fusion calculation are dynamically allocated, and the scene base weight is set for Bluetooth, ultra-wideband, real-time differential, and inertial navigation according to the judged scene type, and is corrected online according to the source quality. Scene base weight examples are: real-time differential in outdoor open area accounts for about 55%, inertial accounts for about 25%, ultra-wideband accounts for about 15%, and Bluetooth accounts for about 5%; in indoor ordinary, ultra-wideband accounts for about 45%, Bluetooth accounts for about 25%, inertial accounts for about 25%, and real-time differential accounts for about 5%; in indoor high precision, ultra-wideband accounts for about 60%, inertial accounts for about 30%, and Bluetooth accounts for about 10%; in floor transformation, inertial accounts for about 45%, ultra-wideband accounts for about 30%, Bluetooth accounts for about 15%, and real-time differential accounts for about 10%. The source quality is obtained in the following way: real-time differential is mapped according to the median value of carrier-to-noise ratio, horizontal or position accuracy factor, and the number of visible satellites; ultra-wideband is mapped according to the line-of-sight confidence and ranging residual variance; Bluetooth is mapped according to the number of visible beacons and signal strength variance; and inertial is mapped according to the short window stability and inverse index of zero offset estimation. The temporary weight is obtained by normalizing the multiplication of the scene base weight and the quality fraction of each source, and is limited in the range of 5% to 75% of a single source; the exponential smoothing is adopted in time, and the update period is 0.5 seconds to suppress transient fluctuations.

[0061] The data source type participating in fusion calculation and its priority refers to automatically determining which data source will be included in fusion calculation and its processing order according to a real-time space scene. For example, when a UWB signal quality decrease or GNSS signal loss is detected, the system can automatically increase the priority of inertial navigation and Bluetooth to ensure the continuity and availability of positioning. Specifically, the data source to be included or bypassed is automatically determined through source selection gating and hysteresis strategy. When the number of visible satellites is less than five or the position dilution of precision is greater than four, the real-time kinematic differential is bypassed and only used as constraint information without directly participating in calculation; when the number of visible satellites is restored to not less than eight and the median value of carrier-to-noise ratio is not less than 35 dBHz, the main solution is included. When the line-of-sight confidence of UWB is less than 0.35 or the ranging residual variance is out of limit, it is temporarily not included, and when it is restored to 0.5 and above for three consecutive time windows, the weight is gradually increased. When the number of visible beacons is less than three or the signal strength fluctuation exceeds the threshold, Bluetooth is only used as auxiliary information. In the floor transition scene, the participation of inertial and barometric pressure is forced to increase, so that the combined weight of the two is not less than 40%, and the growth rate of the weight of other sources is limited to avoid high jitter. Regardless of the scene, the minimum participation of the inertial source is not less than 10% to ensure trajectory continuity.

[0062] The "acquisition frequency" and "activation state" settings further reflect the dynamic trade-off between power consumption and accuracy of the system. In the ordinary operating area, some data sources can reduce the sampling frequency or sleep to save energy, while in high-risk or critical locations, the acquisition frequency and working state of UWB / RTK and inertial navigation are automatically increased to achieve fine control on demand.

[0063] S5: Collect Bluetooth ranging data, UWB positioning data, RTK positioning data and inertial navigation data according to the fusion strategy parameters.

[0064] Bluetooth ranging data is obtained by measuring the signal strength of Bluetooth and calculating the distance, UWB positioning data is obtained by measuring the two-way time of flight of UWB signals, RTK positioning data is obtained by analyzing the positioning differential signals of satellites through RTK, and inertial navigation data is obtained by measuring the acceleration and angular velocity of the positioning object.

[0065] In step S5, "Bluetooth ranging data" is obtained by measuring the signal strength (RSSI) between the positioning object and the surrounding Bluetooth beacons, estimating the distance based on the propagation loss model, and then obtaining the distance information of the positioning object relative to each beacon. This method can achieve low-power auxiliary positioning in an environment with wide coverage and flexible deployment, especially in large-scale ordinary operating areas as an alternative data source when UWB or RTK fails.

[0066] The "UWB positioning data" refers to the high time resolution characteristics of UWB (Ultra-Wide Band) communication, using a two-way time of flight (TOF, Time of Flight) ranging method to accurately measure the propagation delay between the positioning object and multiple UWB base stations, and combining the spatial coordinates of the base stations to achieve high-precision two-dimensional or three-dimensional positioning at the centimeter level. UWB ranging has strong anti-multipath interference capability and performs well in complex factory buildings, metal-intensive areas and other scenarios, and is the core data source for high-precision positioning in key scenarios.

[0067] The "RTK positioning data" is obtained by receiving satellite navigation signals and using a real-time dynamic difference correction (RTK, Real-Time Kinematic) algorithm to analyze high-precision three-dimensional coordinates. This data is suitable for open outdoor areas and cross-regional scenarios that require absolute spatial coordinates, and can provide accurate positioning within an error range of sub-meters or even centimeters, and provide a reference for the global consistency of multiple source positioning systems.

[0068] The "inertial navigation data" measures the three-axis acceleration and angular velocity of the positioning object to calculate its motion state, trend and short-time displacement compensation in real time. Inertial data is particularly suitable for periods of temporary signal loss or environmental changes, and can achieve seamless continuous trajectory compensation, effectively suppressing error accumulation caused by system interruption.

[0069] According to the fusion strategy parameters, the collection frequency, participation, and priority of each type of data can be dynamically adjusted according to the spatial scene and real-time system state. For example, during floor changes or signal blocking periods, the system can increase the inertial navigation data collection frequency to ensure trajectory continuity; in areas with good UWB and RTK signals, the workload of inertial and Bluetooth can be reduced to achieve a synergistic optimum of precision and energy consumption.

[0070] S6: Fuse the Bluetooth ranging data, UWB positioning data, RTK positioning data, and inertial navigation data to obtain real-time three-dimensional high-precision positioning data.

[0071] Based on the fusion strategy parameters, calculate the initial three-dimensional position information of the positioning object using Bluetooth ranging data, UWB positioning data, and RTK positioning data;

[0072] Use inertial navigation data to calculate the real-time motion trend of the positioning object and dynamically correct the initial three-dimensional position information;

[0073] Fuse the corrected initial three-dimensional position information and the motion trend calculated by the inertial navigation data through adaptive weighting to obtain real-time updated real-time three-dimensional high-precision positioning data.

[0074] In step S6, "Bluetooth ranging data, UWB positioning data, RTK positioning data and inertial navigation data are fused to obtain real-time three-dimensional high-precision positioning data", which is the core technical link of the multi-modal positioning system of the application to realize high precision, continuity and anti-interference capability improvement.

[0075] The "fusion strategy parameters" are dynamically set by the previous step according to the intelligent discrimination of the space scene, and are used to control the weight distribution and participation priority of various data sources in the fusion algorithm. For example, giving UWB higher weight in indoor high-precision area, and improving the participation of inertial navigation in floor change area. This dynamic adjustment of weight greatly enhances the adaptability of the system.

[0076] "Using Bluetooth ranging data, UWB positioning data and RTK positioning data to calculate the initial three-dimensional position information of the positioning object", that is, through multi-point positioning solution (such as least squares method, Kalman filter, extended Kalman filter, etc.) to the distance / coordinate information collected by each data source, the preliminary three-dimensional coordinates of the current position are obtained. The spatial coverage and precision complement of different data sources in different scenes can ensure the stability and global consistency of the initial positioning in the whole space.

[0077] "Using inertial navigation data to calculate the real-time motion trend of the positioning object, and dynamically correcting the initial three-dimensional position information", refers to the short-time motion estimation calculated according to the acceleration and angular velocity, which is used to predict and compensate the initial three-dimensional coordinates. In this way, when the wireless signal is attenuated or lost temporarily, the trajectory continuity can be realized through the short-term high-frequency advantage of inertial navigation, and the positioning jump or interruption caused by the abnormality of a single data source can be avoided.

[0078] "Adaptive weighted fusion" refers to the system using weighted Kalman filter, Bayesian inference and other fusion algorithms to weighted average or probability fusion on the initial positioning coordinates and inertial navigation motion trend results according to the real-time fusion strategy, making full use of the advantages of each data source, eliminating the errors caused by abnormality or drift, and finally outputting the real-time updated and error minimized three-dimensional high-precision positioning results. Here, the weight not only depends on the environmental discrimination, but also dynamically perceives the actual error level and confidence of each source, realizing accurate and robust all-scene positioning.

[0079] S7: Using real-time three-dimensional high-precision positioning data to monitor the position and trajectory of power grid workers in real time, and realizing alarm response when detecting a sudden event.

[0080] Based on the real-time three-dimensional high-precision positioning data, the current position of the power grid worker is continuously compared with the preset safety area boundary, historical trajectory and operation specification requirements to judge whether there are abnormal situations such as boundary crossing, long-time static, personnel gathering, abnormal motion trajectory and entering dangerous area.

[0081] When any abnormal situation is detected, an alarm response is automatically performed and pushed to the monitoring terminal.

[0082] In step S7, "real-time monitoring of the position and trajectory of the power grid worker using real-time three-dimensional high-precision positioning data, and alarm response when a sudden event is detected", this step is the core landing step of the safety scenario intelligent management of the present application.

[0083] "Real-time three-dimensional high-precision positioning data" refers to the current three-dimensional coordinate point and time sequence trajectory of the power grid worker dynamically updated by the foregoing multi-source fusion algorithm, which includes absolute spatial position, height information and motion direction, etc., and provides high-precision, full-period position support for safety control.

[0084] "Continuously comparing the current position with the preset safety area boundary, historical trajectory and operation specification requirements", that is, the system is built-in with various types of spatial boundaries (such as operation area, prohibited area, dangerous area, etc.), historical trajectory data (used to determine the normal or abnormal moving path of personnel) and operation specification (such as single operation requirement, maximum stay time, etc.). Each time a set of real-time positioning data is obtained, the system will automatically complete rule comparison and space-time analysis.

[0085] "Judging whether there is an abnormal situation of crossing the boundary, long-time static, personnel gathering, abnormal motion trajectory and entering the dangerous area", specifically includes:

[0086] Boundary crossing: the positioning point crosses the boundary of the operation permitted area, indicating a violation or misentry into a dangerous area;

[0087] Long-time static: the three-dimensional coordinates of the same personnel have no significant change for a long time, which may cause an accident or a violation of stay;

[0088] Personnel gathering: multiple workers gather in the same space within a short time, which violates the on-site control regulations or an emergency situation occurs;

[0089] Abnormal motion trajectory: such as abnormal motion speed, sudden breakpoint of trajectory, large amplitude Z-axis jump, etc., which may be a fall, a slip or an abnormal device;

[0090] Entering a dangerous area: the positioning point enters a high-voltage, flammable, restricted space or other dangerous partition.

[0091] "Automatically performing an alarm response and pushing to the monitoring terminal" means that as soon as the system detects any of the above abnormal events, the alarm mechanism is automatically triggered. It can include sound and light alarm, SMS / APP push, monitoring large screen pop-up window and other multi-channel synchronous early warning, which pushes the event position, time, personnel information to the monitoring center and related persons in real time, realizes rapid intervention and disposal.

[0092] Embodiment 2, which is a second embodiment of the present application, differs from the previous embodiment in that:

[0093] The functions described can be implemented in software, firmware, hardware, or any combination thereof. If implemented in software and as an independent application, it can be stored in a computer-readable storage medium, such as a read-only memory (ROM), a flash memory, a random access memory (RAM), a magnetic disk storage, or an optical disk storage, etc. The computer-readable storage medium can be distributed among computer systems over a network, and it can be stored for use as a distribution medium.

[0094] The logic and / or steps represented in the flowcharts and / or described herein, for example, can be considered as a sequence of instructions to cause a processor to perform a function, and can be stored in a computer-readable medium, which can be any medium that can be accessed by a computer system, such as a server, a personal computer, a network appliance, or other computing device. The computer-readable medium can be a computer program product, which can be distributed over a network, for example, and can be stored in a computer storage or memory, which can be any medium (medium) that can store data that can be accessed by a computer system, such as a server, a personal computer, a network appliance, or other computing device.

[0095] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection (electronic device) having one or more wires; a portable computer diskette (magnetic device); a random access memory (RAM); a read-only memory (ROM); an erasable programmable read-only memory (EPROM or flash memory); an optical fiber device; and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium can even be paper or other suitable medium on which the program can be printed, as the program can be electronically obtained, for example, by optical scanning of the paper or other medium, followed by electronic conversion, interpretation, or processing, if necessary, in other suitable ways, and then stored in a computer memory.

[0096] Embodiment 3 provides a Bluetooth+UWB+RTK fusion positioning system as an embodiment of the application, comprising an environmental feature data acquisition module, a preprocessing module, a spatial scene discrimination module, a fusion strategy parameter setting module, a multi-source data acquisition module, a multi-source fusion calculation module and a trajectory monitoring and alarm module.

[0097] The environmental feature data acquisition module acquires feature data of the environment in which the positioning object is located to obtain environmental feature data.

[0098] The preprocessing module pre-processes the environmental feature data to obtain pre-processed environmental feature data.

[0099] The spatial scene discrimination module discriminates the spatial scene in which the positioning object is located based on the pre-processed environmental feature data to obtain a spatial scene type.

[0100] The fusion strategy parameter setting module sets fusion strategy parameters of Bluetooth, UWB, RTK and inertial navigation according to the spatial scene type.

[0101] The multi-source data acquisition module acquires Bluetooth ranging data, UWB positioning data, RTK positioning data and inertial navigation data according to the fusion strategy parameters.

[0102] The multi-source fusion calculation module performs fusion calculation on the Bluetooth ranging data, UWB positioning data, RTK positioning data and inertial navigation data to obtain real-time three-dimensional high-precision positioning data.

[0103] The trajectory monitoring and alarm module monitors the position and trajectory of the power grid worker in real time using the real-time three-dimensional high-precision positioning data, and realizes alarm response when a sudden event is detected.

[0104] Embodiment 4 provides a Bluetooth+UWB+RTK fusion positioning method as an embodiment of the application. In order to verify the beneficial effects of the application, economic benefit calculation and simulation / contrast experiments are used for scientific demonstration.

[0105] In this experiment, a certain power facility is selected as a typical test scene, and different areas in the power grid facility (including indoor ordinary areas, indoor high-precision areas, outdoor open areas and floor change areas) are selected as positioning test sites. Six power grid field workers A to F are selected for positioning test. Before the test, a positioning system is constructed by laying Bluetooth beacons, UWB base stations and RTK ground differential reference stations, and each test object is equipped with a fusion positioning terminal to ensure that the test environment is close to the actual power grid working scene.

[0106] During the experiment, each test subject carries a fusion positioning terminal and moves in different areas. First, the positioning terminal collects real-time environmental feature data, including acceleration data, angular velocity data, air pressure data, Bluetooth signal strength, UWB signal quality, GNSS satellite signal strength, and illumination data. These data are stored in the local server after synchronous collection. Then, the server preprocesses the collected environmental feature data, including denoising, normalization, time alignment, and missing value completion, to ensure data integrity and accuracy. Next, the preprocessed environmental feature data is input into the pre-trained scene discrimination model to quickly and accurately determine the current space scene of the positioning object. According to the scene type, the strategy parameters of Bluetooth, UWB, RTK, and inertial navigation fusion positioning data are automatically set, and Bluetooth ranging data, UWB positioning data, RTK positioning data, and inertial navigation data are collected. These data are sent to the fusion positioning calculation module in real time to obtain real-time updated high-precision three-dimensional positions. Finally, through the positioning system background monitoring platform, the position and motion trajectory of each test object are monitored in real time. Once abnormal events such as boundary crossing, long-term static personnel, or entering dangerous areas are detected, the system triggers an alarm mechanism and quickly pushes alarm information to the on-site monitoring terminal to achieve rapid response. The entire experiment process strictly follows the pre-set experiment plan to objectively record the actual effect of the method. For specific data, refer to Table 1.

[0107] Table 1 Data Reference Table

[0108]

[0109] From the above test data table, it can be seen that the fusion positioning method of the present application is significantly better than single technology positioning method in various indicators. First, in terms of positioning accuracy, the Bluetooth single technology positioning error is generally more than 2 meters, which cannot meet the precise positioning requirement; the UWB and RTK single technology positioning error is between 0.33 meters and 0.45 meters, although it has higher accuracy, but there is a certain stability and accuracy fluctuation risk in complex environment or dynamic scene. After the fusion positioning method proposed by the present application, the positioning error of each test object is significantly reduced, and the fusion positioning error is stably maintained between 0.15 meters and 0.20 meters, showing obvious accuracy advantage. Especially in the case of test object D, the fusion positioning error is 0.15 meters, which is more than one time higher than the positioning accuracy of single UWB and RTK.

[0110] In addition, from the abnormal event detection and alarm response speed index, the method of the present application also performs outstandingly. The fusion positioning technology can not only quickly detect the abnormal position behavior of the test object, but also the detection time is generally in the range of 13.5 seconds to 16.7 seconds, and the alarm response time is stable between 2.1 seconds to 2.6 seconds, showing high real-time early warning and rapid response capability. Compared with the alarm response delay or false alarm and missed alarm problems commonly existing in the prior art, the present application can effectively improve the safety management efficiency and emergency handling speed of power grid workers in the actual production operation process, and has obvious on-site practical value and significant technical advantages.

[0111] In summary, the innovation and novelty of the present application lies in the intelligent dynamic fusion of multiple positioning technologies, which significantly improves the system positioning accuracy, continuity, anti-interference ability and abnormal event response speed, effectively solving the problems of poor adaptability to complex environment and insufficient safety protection capability in the prior art.

[0112] To further prove the mechanism and technical effect of the key technology in real working conditions, four types of representative working conditions are selected for link recording (input elements → scene discrimination → weight adjustment and source inclusion / bypass → positioning error and continuity → comparative example);

[0113] Outdoor open working condition (verify the effect of "scene discrimination + dynamic weight"): continuous walking on the road outside the factory, the number of visible satellites is 11, the horizontal precision factor is 1.2, and the median value of carrier-to-noise ratio is 36; the number of visible Bluetooth beacons is 2, and the ultra-wideband line-of-sight confidence is low. The probability of two-stage discrimination output is 0.91, and it is stable as the scene after triggering the hysteresis condition. According to the quality, the system will increase the fusion weight of real-time dynamic difference to about 0.62, about 0.23 for inertia, about 0.10 for ultra-wideband, and about 0.05 for Bluetooth. The plane error of the fusion result is 0.18 m, which is reduced by about 42% compared with the comparative example of fixed weight (0.31 m); on the same trajectory, the scene jitter times are reduced from 5 times of the comparative example to 0 times;

[0114] Indoor high-precision working condition (verify the effect of "indoor UWB dominant + Z-axis suppression drift"): walking around the high-precision equipment area, the ultra-wideband line-of-sight confidence is 0.78, the ranging residual variance is low, the number of visible Bluetooth beacons is 9, and the satellite is unavailable. The probability of two-stage discrimination output is 0.88, and it is stable after three-window hysteresis confirmation. The system will increase the fusion weight of ultra-wideband to about 0.66, about 0.27 for inertia, about 0.07 for Bluetooth, and implement time smoothing for height measurement. The three-dimensional fusion error is 0.14 m, which is reduced by about 55% compared with the comparative example of using only ultra-wideband, and the Z-axis drift peak is less than 0.25 m; the comparative example appears instantaneous height jitter at the edge of turning and shielding, with a peak of 0.41 m;

[0115] Floor transition condition (verify the effect of "gate / bypass + minimum participation + transition continuous chain"): After going up the stairs, entering the elevator and going down, the pressure difference within three seconds is 0.7 hPa, and the vertical acceleration variance increases by about 60% compared to the previous time window. The two-level discrimination output probability of "floor transition" is 0.86 and remains stable; the system forces the total participation of inertia and air pressure to be no less than 40%, while limiting the sudden increase rate of other source weights to prevent height jitter; the satellite signal is once weak, and the real-time dynamic difference is automatically set to bypass, providing only constraint information; after the elevator exits, the signal recovers, and the three-window slow rise strategy is automatically included in the main solution. The positioning is continuously uninterrupted within 10 s of transition, and the final height error is 0.28 m; the comparative example (without gate and hysteresis) has a coordinate jump of 0.63 m in the elevator section, and a second jump occurs when it recovers;

[0116] Indoor ordinary and short-time non-line-of-sight obstruction condition (verify the effect of "quality-driven weight reduction and jump suppression"): In a large metal equipment channel, the ultra-wideband line-of-sight confidence decreases from 0.65 to 0.28 and lasts for 2 s, and there are 6 visible Bluetooth beacons. The system automatically reduces the ultra-wideband weight to about 0.12 within 2 time windows, while increasing the combined weight of Bluetooth and inertia to about 0.63, and maintaining the minimum participation of inertia to ensure trajectory continuity; after the obstruction is removed, the three-window slow rise strategy is used to gradually restore the dominance of ultra-wideband. The maximum position jump in this section is 0.42 m, and the comparative example (fixed weight) is 0.95 m, and the time required to recover to steady state is 3 time windows longer than the proposed scheme;

[0117] End-to-end performance of the alarm link (verify the closed loop of "continuity → detection → response"): In the mixed task of the above four types of conditions, the event detection time remains in the range of 13.5-16.7 s, and the alarm response time remains in the range of 2.1-2.6 s; in three safety drills, the false alarm rate caused by jitter switching and jump is reduced from 6.8% of the comparative example to 2.3%, and the average disposal time is shortened by about 27%. The results show that: the two-level scene discrimination, quality-driven dynamic weight, hysteresis and time smoothing, data source gate / bypass and minimum participation mechanism form a synergistic effect in key conditions such as cross-scene, transition and obstruction, which can significantly improve the positioning accuracy, continuity and robustness without sacrificing response speed, and reduce the risk of false alarms.

[0118] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, which should be covered by the claims of the present application.

Claims

1. A Bluetooth+UWB+RTK fusion positioning method, characterized in that, The application relates to a positioning method and device. The method comprises the following steps: Collecting characteristic data of an environment where a positioning object is located to obtain environment characteristic data; Preprocessing the environment characteristic data to obtain preprocessed environment characteristic data; Judging a spatial scene where the positioning object is located based on the preprocessed environment characteristic data to obtain a spatial scene type; Setting fusion strategy parameters of Bluetooth, UWB, RTK and inertial navigation according to the spatial scene type; Collecting Bluetooth ranging data, UWB positioning data, RTK positioning data and inertial navigation data according to the fusion strategy parameters; Fusing and calculating the Bluetooth ranging data, the UWB positioning data, the RTK positioning data and the inertial navigation data to obtain real-time three-dimensional high-precision positioning data; Monitoring the position and trajectory of a power grid worker in real time by using the real-time three-dimensional high-precision positioning data, and realizing alarm response when a sudden event is detected; The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps The classification model adopts a two-stage judgment process of rule screening and supervised learning: in the first-stage rule screening, candidate scene labels are marked based on threshold values and relational expressions of preprocessed environmental features; in the second-stage supervised learning, probability distributions of four categories of outdoor open, indoor ordinary, indoor high-precision and floor transformation are output, and a hysteresis condition that allows category switching only when three consecutive time windows are consistent is set; the preprocessed environmental features are organized into model inputs in a sliding time window with a length of two seconds and a step of 0.5 seconds.

2. The Bluetooth+UWB+RTK fusion positioning method of claim 1, wherein, The preprocessing of the environmental feature data includes denoising, normalization, time alignment and missing value completion of the environmental feature data to obtain preprocessed environmental feature data. 3.The Bluetooth+UWB+RTK fusion positioning method of claim 2, wherein, The collection of the Bluetooth ranging data, the UWB positioning data, the RTK positioning data and the inertial navigation data respectively includes: the Bluetooth ranging data is obtained by measuring the Bluetooth signal strength and calculating, the UWB positioning data is obtained by two-way time flight measurement of the UWB signal, the RTK positioning data is obtained by analyzing the positioning differential signal of the satellite in the RTK mode, and the inertial navigation data is collected by measuring the acceleration and angular velocity of the positioning object.

4. The Bluetooth+UWB+RTK fusion positioning method of claim 3, wherein, The real-time monitoring of the position and trajectory of the power grid workers includes continuously comparing the current position of the power grid workers with the preset safety area boundary, historical trajectory and operation specification requirements based on real-time three-dimensional high-precision positioning data to determine whether there are any abnormal situations such as boundary crossing, long-time stillness, personnel gathering, abnormal motion trajectory and entering a dangerous area. When any abnormal situation is detected, an alarm response is automatically triggered and pushed to the monitoring terminal.

5. A Bluetooth+UWB+RTK integrated positioning system for implementing the Bluetooth+UWB+RTK integrated positioning method according to any one of claims 1 to 4, characterized in that, It comprises: an environmental feature data collection module that collects feature data of an environment in which a positioning object is located to obtain environmental feature data; a preprocessing module that preprocesses the environmental feature data to obtain preprocessed environmental feature data; a spatial scene discrimination module that discriminates a spatial scene in which the positioning object is located based on the preprocessed environmental feature data to obtain a spatial scene type; a fusion strategy parameter setting module that sets fusion strategy parameters of Bluetooth, UWB, RTK and inertial navigation according to the spatial scene type; a multi-source data collection module that collects Bluetooth ranging data, UWB positioning data, RTK positioning data and inertial navigation data according to the fusion strategy parameters; a multi-source fusion calculation module that fuses and calculates the Bluetooth ranging data, the UWB positioning data, the RTK positioning data and the inertial navigation data to obtain real-time three-dimensional high-precision positioning data; a trajectory monitoring and alarm module that uses the real-time three-dimensional high-precision positioning data to monitor the position and trajectory of the power grid workers in real time, and triggers an alarm response when a sudden event is detected.

Images