A seat belt monitoring method of centimeter level hanging point determination

Abstract

The present application relates to a kind of centimeter level hanging point determination safety belt monitoring method, specifically relates to radio positioning technical field, by calculating instantaneous confidence weight based on channel impulse response energy distribution, screening high-quality reference base station and revising ranging value, then using Gauss-Newton iteration method to solve hook three-dimensional coordinates, realize centimeter level positioning accuracy;With pre-binding mapping table accurately determine hook belonging hanging point, avoid false hanging misjudgment;Through spatial distance and the consistency of the double criterion of revised ranging value trigger double-hook same point alarm, prevent safety belt lose double-point protection;Based on barometer and inertial measurement unit's attitude compensation, in combination with the elevation of belonging hanging point accurately calculate actual hanging height difference, effectively detect low hanging high use violation;Finally, belonging identification and violation sign are integrated into alarm information and uploaded to cloud monitoring platform through star flash SLE link, realize real-time remote safety supervision.

Description

Technical Field

[0001] This invention relates to the field of radio positioning technology, and more specifically, to a method for detecting seat belt attachment points at the centimeter level. Background Technology

[0002] High-altitude operations are an indispensable part of work in fields such as power tower maintenance, building construction, and wind power maintenance. Workers typically wear full-body safety harnesses, which are connected to secure anchor points above, such as lifelines, scaffold crossbars, or fall arrest rings, via double hooks on both sides. Taking a 220 kV transmission tower as an example, the tower height can reach over 45 meters, with a horizontal lifeline installed every 5 meters along the climbing path, and the spacing between adjacent lifelines is approximately 50 to 80 centimeters. In building scaffolding scenarios, according to the technical specifications for safety in high-altitude operations in building construction, a horizontal safety steel wire rope is installed as a lifeline at a height of approximately 1.8 meters on each step of the external scaffolding, with a fixed anchor point every 6 meters. In wind power tower scenarios, the tower height can reach 90 meters, and a rest platform that also serves as a fall arrest point is installed every 5 to 8 meters on the inner wall ladder. All of the above scenarios are characterized by dense steel structures, confined spaces, and dynamically changing work surfaces, which places extremely high demands on the accurate monitoring of the safety harness attachment status.

[0003] Current safety belt monitoring technology mainly relies on a combination of ultra-wideband (UWB) positioning and inertial measurement units (IMUs) and barometers. The UWB positioning module is typically integrated into a control box within the safety belt's waistband. It obtains the approximate location of the worker by measuring the time-of-flight distance with independently deployed UWB base stations on-site. However, in practical applications, UWB signals suffer severe multipath interference in steel structure environments such as tower angle steel and scaffolding pipes. Non-line-of-sight errors cause positioning accuracy to plummet from the theoretical 10-30 centimeters to 1-3 meters. This level of accuracy can only determine the approximate area where the worker is located, and cannot accurately determine which specific lifeline or which lifeline each of the left and right hooks is attached to. On the hanging rings; especially when the distance between two adjacent lifelines is only 50 to 80 centimeters, the existing UWB solution is completely unable to distinguish whether the two hooks are attached to the same hanging point; the double hooks on the same point is an extremely dangerous violation in high-altitude operations. Once a single hanging point fails, both hooks will lose their protective function at the same time, directly leading to a fall accident; the existing solution can only rely on a rough threshold estimation of the distance between the two hooks to determine such high-risk violations, and the false alarm rate and false alarm rate are both high, which seriously restricts the actual effectiveness of the safety belt intelligent monitoring system; therefore, how to achieve centimeter-level accuracy in determining the hanging point attribution in complex high-altitude operation environments with dense steel structures and severe multi-path interference has become a core technical problem that urgently needs to be solved. Summary of the Invention

[0004] This invention addresses the technical problems existing in the prior art by providing a seat belt monitoring method for determining the attachment point at the centimeter level, thereby resolving the issues raised in the background section.

[0005] The technical solution of this invention to solve the above-mentioned technical problems is as follows: specifically, it includes the following steps:

[0006] Step S1: In response to the Star Flash SLP tag being triggered and woken up by the event, obtain the original ranging value, received signal strength indication and channel impulse response energy distribution of each Star Flash SLP anchor point. Calculate the instantaneous confidence weight of each Star Flash SLP anchor point based on the direct path signal-to-noise ratio and multipath scattering factor in the channel impulse response energy distribution, and establish a pre-binding mapping table between the Star Flash SLP anchor point identifier and the three-dimensional coordinates and elevation of the hanging point in the BIM model.

[0007] Step S2: Select the first three star flash SLP anchor points as reference base stations in descending order of instantaneous confidence weights. Use the direct path signal-to-noise ratio to correct the original ranging value and correct the ranging value. Construct a weighted least squares objective function with the corrected ranging value and the three-dimensional coordinates of the hanging point in the BIM model of the reference base station.

[0008] Step S3: Solve the weighted least squares objective function using the Gauss-Newton iterative method to obtain the estimated three-dimensional coordinates of the left hook and the right hook; traverse the three-dimensional coordinates of the hanging points in the BIM model in the pre-binding mapping table, calculate the Euclidean distance between the estimated three-dimensional coordinates of the left hook and the three-dimensional coordinates of the hanging points in each BIM model, and the Euclidean distance between the estimated three-dimensional coordinates of the right hook and the three-dimensional coordinates of the hanging points in each BIM model. When the minimum Euclidean distance of the left hook is less than 30 cm, determine the ownership of the left hook and output the star-flash SLP anchor point identifier of the left hook and the elevation of the hanging point in the BIM model to which the left hook belongs. When the minimum Euclidean distance of the right hook is less than 30 cm, determine the ownership of the right hook and output the star-flash SLP anchor point identifier of the right hook and the elevation of the hanging point in the BIM model to which the right hook belongs. Otherwise, determine the state of virtual hanging.

[0009] Step S4: Calculate the spatial distance between the estimated three-dimensional coordinates of the left hook and the estimated three-dimensional coordinates of the right hook. When the spatial distance is less than 50 cm or the absolute value of the difference between the corrected distance measurements of the left hook and the right hook from the same StarShock SLP anchor point is less than 20 cm, trigger a double hook same point alarm. At the same time, read the absolute waist height measured by the barometer in the waist belt host and the pitch angle output by the inertial measurement unit. Combine the smaller of the hanging point elevation in the BIM model to which the left hook belongs and the hanging point elevation in the BIM model to which the right hook belongs to calculate the actual hanging height difference. When the actual hanging height difference exceeds the allowable error zone, it is determined to be a violation of low hanging and high use. The StarShock SLP anchor point identifier of the left hook, the StarShock SLP anchor point identifier of the right hook, the double hook same point violation mark and the low hanging and high use violation mark are integrated into alarm information and uploaded to the cloud monitoring platform through the StarShock SLE link.

[0010] In a preferred embodiment, in step S1, in response to the Star Flash SLP tag being triggered and awakened by an event, the Star Flash SLP tag initiates a ranging request to all visible Star Flash SLP anchor points; each Star Flash SLP anchor point returns the original ranging value, the received signal strength indication, and the channel impulse response energy distribution; the channel impulse response energy distribution is composed of an ordered set of energy amplitudes of several sampling points arranged in chronological order, and each sampling point corresponds to a specific time delay tap and its energy amplitude.

[0011] In a preferred embodiment, the process of calculating the instantaneous confidence weight of each starburst SLP anchor point based on the direct path signal-to-noise ratio and multipath scattering factor in the channel impulse response energy distribution specifically involves:

[0012] The specific calculation process of the direct path signal-to-noise ratio is as follows: the value with the largest energy amplitude in the channel impulse response energy distribution is recorded as the peak energy, the average value of the energy amplitude of all delay taps during the no-signal period is recorded as the noise floor energy, and the quotient obtained by dividing the peak energy by the noise floor energy is the direct path signal-to-noise ratio.

[0013] Then, calculate the square of the sum of the absolute values ​​of the energy amplitudes of all delay taps from the channel impulse response energy distribution, and divide it by the product of the total number of delay taps and the sum of the squares of the energy amplitudes of the delay taps to obtain the multipath dispersion factor.

[0014] Then, the direct path signal-to-noise ratio is divided by the pre-calibrated reference direct path signal-to-noise ratio, multiplied by the multipath dispersion factor, and then multiplied by the square of the ratio of the current received signal strength indication to the maximum received signal strength indication among all visible star flash SLP anchor points. The resulting product is used as the instantaneous confidence weight of the star flash SLP anchor point.

[0015] In a preferred embodiment, the process of establishing a pre-binding mapping table between the StarSpark SLP anchor point identifier and the three-dimensional coordinates and elevation of the hanging point in the BIM model specifically involves:

[0016] During the system initialization phase, through on-site surveys or import from the BIM model, the unique hardware identifier of each StarSLP anchor point is mapped one-to-one with the three-dimensional coordinates and elevation of the corresponding hanging point in the BIM model, forming a five-tuple record consisting of the StarSLP anchor point identifier, the three components of the hanging point's three-dimensional coordinates, and the hanging point's elevation. All five-tuple records are then compiled into a static pre-binding mapping table, which is persistently stored throughout the entire monitoring cycle and updated manually or automatically only when the StarSLP anchor point layout changes.

[0017] In a preferred embodiment, the specific process of selecting the first three star flash SLP anchor points as reference base stations in descending order of instantaneous confidence weight in step S2 is as follows:

[0018] From all the visible star flash SLP anchor points output in step S1, sort the instantaneous confidence weight of each star flash SLP anchor point from largest to smallest, and select the top three star flash SLP anchor points as reference base stations; if the number of visible star flash SLP anchor points is less than three, then all visible star flash SLP anchor points are used as reference base stations.

[0019] After selection, for each reference base station, the original ranging value obtained in step S1 is corrected for multipath error using the direct path signal-to-noise ratio calculated by the reference base station in step S1, to obtain the corrected ranging value. The specific process of multipath error correction is as follows: a maximum possible multipath error correction amount and a signal-to-noise ratio attenuation constant are preset; the negative direct path signal-to-noise ratio of the natural constant is divided by the power of the signal-to-noise ratio attenuation constant to obtain the exponential calculation result; the maximum possible multipath error correction amount is multiplied by the exponential calculation result to obtain the multipath error correction amount; the multipath error correction amount is subtracted from the original ranging value, and the difference is the corrected ranging value.

[0020] In a preferred embodiment, the specific operation of constructing the weighted least squares objective function using the corrected ranging value and the three-dimensional coordinates of the anchor point in the reference base station BIM model is as follows:

[0021] Using the three components of the hook's three-dimensional coordinates to be determined as unknown variables, we first calculate the difference between the corrected ranging value of each reference base station and the Euclidean distance between the hook's three-dimensional coordinates to be determined and the hanging point's three-dimensional coordinates in the BIM model of the reference base station, and obtain the residual corresponding to each reference base station.

[0022] Next, the residuals corresponding to each reference base station are squared, and then the squared residuals are multiplied by the instantaneous confidence weights of the reference base stations calculated in step S1 to obtain the weighted residual squares corresponding to each reference base station.

[0023] Finally, the weighted residual squares corresponding to all reference base stations are summed to obtain a total sum; the weighted least squares objective function is constructed with minimizing the total sum as the optimization objective.

[0024] In a preferred embodiment, in step S3, the Gauss-Newton iterative method is used to solve the weighted least squares objective function to obtain the estimated three-dimensional coordinates of the left hook and the right hook, specifically:

[0025] Using the three components of the three-dimensional coordinates of the left hook to be determined as unknown variables, and taking the corrected ranging values ​​of the three reference base stations selected in step S2, the three-dimensional coordinates of the hook point in the BIM model, and the instantaneous confidence weight as inputs, the weighted least squares objective function constructed in step S2 is applied, and the reference base station data corresponding to the left hook is substituted; an adaptive damping factor is introduced, and the damping factor is multiplied by the identity matrix and added to the Hessian matrix approximation term of the Gauss-Newton method to form the augmented method equation, and then the augmented method equation is solved to obtain the coordinate update; the estimated value of the three-dimensional coordinates of the left hook in the current iteration step is added to the coordinate update to obtain the next iteration. The coordinate estimates are obtained; after each iteration, the damping factor is dynamically adjusted according to the ratio of the actual descent to the predicted descent: when the ratio is greater than 0.75, the damping factor is reduced by half; when the ratio is less than 0.25, the damping factor is doubled; the iteration termination condition is that any one of the following three conditions is met first: the gradient norm is less than one ten-thousandth of a meter, the magnitude of the coordinate update is less than one thousandth of a meter, and the number of iterations exceeds twenty; after convergence, the three-dimensional coordinate estimates of the left hook are obtained; the three-dimensional coordinate estimates of the right hook are solved independently using the same process, using the data from the three reference base stations selected in step S2.

[0026] In a preferred embodiment, the specific operation of traversing the three-dimensional coordinates of the hanging points in the BIM model of the pre-binding mapping table is as follows:

[0027] The estimated 3D coordinates of the left hook are compared with the 3D coordinates of the hook point in the BIM model of each hook point in the pre-binding mapping table to calculate the Euclidean distance, resulting in a set of distance values. The minimum value is found from the set of distance values ​​and recorded as the minimum Euclidean distance of the left hook. The minimum Euclidean distance of the left hook is compared with 30 cm. If the minimum Euclidean distance of the left hook is less than 30 cm, it is determined that the left hook belongs to the hook point corresponding to the minimum distance, and the left hook belongs to the Star Flash SLP anchor point identifier and the hook point elevation in the BIM model to which the left hook belongs is output.

[0028] If the minimum Euclidean distance of the left hook is not less than 30 cm, the left hook is determined to be in a dummy state and no ownership information is output; the right hook is processed in the same way, and the estimated three-dimensional coordinates of the right hook are used to replace the estimated three-dimensional coordinates of the left hook to obtain the ownership determination result of the right hook.

[0029] In a preferred embodiment, in step S4, the Euclidean distance between the estimated three-dimensional coordinates of the left hook and the estimated three-dimensional coordinates of the right hook is calculated to obtain the spatial distance;

[0030] Then, the spatial distance is compared with 50 centimeters. If the spatial distance is less than 50 centimeters, a double-hook same-point alarm is triggered directly. If the spatial distance is not less than 50 centimeters, the absolute value criterion of the difference between the corrected ranging values ​​is further executed: the same star-flash SLP anchor point that exists in both reference base stations is found from the three reference base stations corresponding to the left hook and the three reference base stations corresponding to the right hook. The corrected ranging values ​​of the same star-flash SLP anchor point for the left hook and the right hook are taken respectively. The absolute value of the difference between the two corrected ranging values ​​is calculated and compared with 20 centimeters. If the absolute value is less than 20 centimeters, a double-hook same-point alarm is triggered. If the spatial distance is not less than 50 centimeters and the absolute value of the difference between the corrected ranging values ​​is not less than 20 centimeters, a double-hook same-point alarm is not triggered.

[0031] In a preferred embodiment, in step S4, the absolute height of the waist measured by the barometer in the waist belt host and the pitch angle output by the inertial measurement unit are read, and the absolute height of the waist is multiplied by the cosine value of the pitch angle to obtain the actual hanging height.

[0032] Then, select the smaller of the hanging point elevations in the BIM model belonging to the left hook and the right hook, and record it as the minimum allowable height; subtract the minimum allowable height from the actual hanging height to obtain the actual hanging height difference;

[0033] The actual hanging height difference is compared with a pre-set allowable error band of 10 centimeters. If the actual hanging height difference is less than negative 10 centimeters, it is judged as a violation of low hanging and high use, and the low hanging and high use violation mark is set. If the actual hanging height difference is not less than negative 10 centimeters, it is judged as a normal hanging, and the low hanging and high use violation mark is cleared. Finally, the left hook belonging to the StarShan SLP anchor point, the right hook belonging to the StarShan SLP anchor point, the double hook same point violation mark, and the low hanging and high use violation mark are integrated into a structured alarm message and uploaded to the cloud monitoring platform through the StarShan SLE link.

[0034] The beneficial effects of this invention are as follows: By calculating the instantaneous confidence weight based on the channel impulse response energy distribution, high-quality reference base stations are selected and the ranging value is corrected. Then, the three-dimensional coordinates of the hook are solved using the Gauss-Newton iterative method to achieve centimeter-level positioning accuracy. Combined with the pre-binding mapping table, the hook's assigned attachment point is accurately determined, avoiding false attachment misjudgment. The dual-hook same-point alarm is triggered by the consistency criteria of spatial distance and corrected ranging value to prevent the seat belt from losing dual-point protection. Based on the attitude compensation of the barometer and inertial measurement unit, the actual attachment height difference is accurately calculated by combining the elevation of the assigned attachment point, effectively detecting violations of low-attachment and high-use. Finally, the attribution identifier and violation mark are integrated into alarm information and uploaded to the cloud monitoring platform through the Star Flash SLE link to achieve real-time remote safety supervision. Attached Figure Description

[0035] Figure 1This is a flowchart of the method of the present invention. Detailed Implementation

[0036] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0037] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0038] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0039] Example 1: This example provides the following... Figure 1 The method for detecting seat belt attachment points at the centimeter level, as shown, specifically includes the following steps:

[0040] Step S1: In response to the Star Flash SLP tag being triggered and woken up by the event, obtain the original ranging value, received signal strength indication and channel impulse response energy distribution of each Star Flash SLP anchor point. Calculate the instantaneous confidence weight of each Star Flash SLP anchor point based on the direct path signal-to-noise ratio and multipath scattering factor in the channel impulse response energy distribution, and establish a pre-binding mapping table between the Star Flash SLP anchor point identifier and the three-dimensional coordinates and elevation of the hanging point in the BIM model.

[0041] Step S2: Select the first three star flash SLP anchor points as reference base stations in descending order of instantaneous confidence weights. Use the direct path signal-to-noise ratio to correct the original ranging value and correct the ranging value. Construct a weighted least squares objective function with the corrected ranging value and the three-dimensional coordinates of the hanging point in the BIM model of the reference base station.

[0042] Step S3: Solve the weighted least squares objective function using the Gauss-Newton iterative method to obtain the estimated three-dimensional coordinates of the left hook and the right hook; traverse the three-dimensional coordinates of the hanging points in the BIM model in the pre-binding mapping table, calculate the Euclidean distance between the estimated three-dimensional coordinates of the left hook and the three-dimensional coordinates of the hanging points in each BIM model, and the Euclidean distance between the estimated three-dimensional coordinates of the right hook and the three-dimensional coordinates of the hanging points in each BIM model. When the minimum Euclidean distance of the left hook is less than 30 cm, determine the ownership of the left hook and output the star-flash SLP anchor point identifier of the left hook and the elevation of the hanging point in the BIM model to which the left hook belongs. When the minimum Euclidean distance of the right hook is less than 30 cm, determine the ownership of the right hook and output the star-flash SLP anchor point identifier of the right hook and the elevation of the hanging point in the BIM model to which the right hook belongs. Otherwise, determine the state of virtual hanging.

[0043] Step S4: Calculate the spatial distance between the estimated three-dimensional coordinates of the left hook and the estimated three-dimensional coordinates of the right hook. When the spatial distance is less than 50 cm or the absolute value of the difference between the corrected distance measurements of the left and right hooks from the same StarShock SLP anchor point is less than 20 cm, trigger a double hook same point alarm. At the same time, read the absolute waist height measured by the barometer in the waist belt host and the pitch angle output by the inertial measurement unit. Combine the smaller of the hanging point elevation in the BIM model to which the left hook belongs and the hanging point elevation in the BIM model to which the right hook belongs to calculate the actual hanging height difference. When the actual hanging height difference exceeds the allowable error zone, it is determined to be a violation of low hanging and high use. The StarShock SLP anchor point identifier of the left hook, the StarShock SLP anchor point identifier of the right hook, the double hook same point violation mark, and the low hanging and high use violation mark are integrated into alarm information and uploaded to the cloud monitoring platform through the StarShock SLE link.

[0044] In this embodiment, it is specifically necessary to explain step S1, in response to the StarSLP tag being triggered and woken up by an event, the StarSLP tag initiates a ranging request to all visible StarSLP anchor points; each StarSLP anchor point returns the original ranging value, received signal strength indication, and channel impulse response energy distribution; the channel impulse response energy distribution consists of an ordered set of energy amplitudes from several sampling points arranged in chronological order, and each sampling point corresponds to a specific time delay tap and its energy amplitude; wherein, the conditions for event-triggered wake-up include: the pressure change gradient detected by the pressure strain gauge at the hook exceeds a preset threshold, or the inertial measurement unit in the belt host... The combined acceleration detected by the unit exceeds twice the gravitational acceleration, and the angular velocity exceeds 180 degrees per second; when any of these conditions are met, the Star Flash SLP tag switches from deep sleep to ranging operation; all visible Star Flash SLP anchor points refer to anchor points that can respond within a specified time window after the tag broadcasts a ranging request in the 8GHz band, and whose received signal strength indication is higher than -90dBm; the number of sampling points N for the channel impulse response energy distribution is usually set to 128 or 256, and the time resolution corresponding to the sampling interval is 2 nanoseconds, thereby covering the propagation delay range from 0 to 512 nanoseconds, which is sufficient to distinguish multipath components that are more than 0.6 meters apart;

[0045] The process of calculating the instantaneous confidence weight of each starburst SLP anchor point based on the direct path signal-to-noise ratio and multipath scattering factor in the channel impulse response energy distribution is as follows:

[0046] First, extract the peak energy and noise floor energy corresponding to the direct path from the channel impulse response energy distribution. Divide the peak energy by the noise floor energy to obtain the direct path signal-to-noise ratio (SNR). The specific calculation process for the direct path SNR is as follows: the value with the largest energy amplitude in the channel impulse response energy distribution is recorded as the peak energy, and the average energy amplitude of all delay taps during the no-signal period is recorded as the noise floor energy. The quotient obtained by dividing the peak energy by the noise floor energy is the direct path SNR. Here, the no-signal period refers to the delay interval in the channel impulse response energy distribution before the first valid echo. There is no direct or reflected path signal in this interval, which only contains thermal noise and background interference. The noise floor energy is calculated using a moving average method to eliminate the influence of instantaneous noise spikes. The typical value range of the direct path SNR is 5 to 30 dB. When it is below 5 dB, it indicates that the signal is extremely weak, and the ranging result of this anchor point will be considered unreliable.

[0047] Next, calculate the square of the sum of the absolute values ​​of the energy amplitudes of all delay taps in the channel impulse response energy distribution, and divide it by the product of the total number of delay taps and the sum of the squares of the energy amplitudes of the delay taps to obtain the multipath scattering factor. The specific calculation process of the multipath scattering factor is as follows: add the absolute values ​​of the energy amplitudes of all delay taps in the channel impulse response energy distribution to obtain the first summation result, and square the first summation result to obtain the first squared result; square the energy amplitude of each delay tap in the channel impulse response energy distribution and sum them to obtain the second summation result; multiply the total number of delay taps by the second summation result. The product of the results is obtained by multiplying the first squared result with the product result. The quotient obtained by dividing the first squared result by the product result is the multipath scattering factor. The multipath scattering factor is dimensionless and its value ranges between 0 and 1. When the energy is concentrated in a few time delay taps, the multipath scattering factor approaches 1, indicating that the multipath interference is slight. When the energy is evenly distributed on many time delay taps, the multipath scattering factor approaches 0, indicating that the multipath interference is severe. This factor is based on the ratio of the L1 norm to the L2 norm, and it can more sensitively reflect the degree of concentration of multipath energy than the traditional standard deviation or variance, thus effectively distinguishing direct paths from reflected paths in complex steel structure environments.

[0048] Then, the direct path SNR is divided by the pre-calibrated baseline direct path SNR, multiplied by the multipath scattering factor, and then multiplied by the square of the ratio of the current received signal strength indication to the maximum received signal strength indication among all visible starburst SLP anchor points. The resulting product is used as the instantaneous confidence weight of that starburst SLP anchor point. The baseline direct path SNR is a fixed value calibrated through actual measurement in an open environment. The calibration method for the baseline direct path SNR is as follows: in an unobstructed open area, the starburst SLP tag is placed 10 meters away from the starburst SLP anchor point, and 100 sets of data are continuously collected. Channel impulse response data is used to calculate the direct path signal-to-noise ratio (SNR) for each data set and take its average value as the baseline direct path SNR; this value is usually stable between 25 and 30 dB; the multipath dispersion factor is used to quantitatively characterize the concentration of multipath energy, and the square ratio of the received signal strength indication is used to enhance the influence of strong signal anchors; the square ratio of the received signal strength indication amplifies the signal strength difference, making strong signal anchors account for a larger proportion in the weighting, thereby prioritizing the selection of anchors with the best signal quality in the reference base station selection, and avoiding the introduction of excessive errors by weak signal anchors;

[0049] In addition, the calculation result of the instantaneous confidence weight will be used as the basis for the ranking of reference base stations in step S2. Anchor points with larger weight values ​​are considered to have higher ranging confidence at the current moment, and are therefore preferentially selected into the reference base station set.

[0050] The specific method for determining the no-signal period is as follows: Starting from the first sampling point of the channel impulse response energy distribution, scan backwards to find the first time delay tap where the energy amplitude exceeds three times the noise floor energy; the interval before this tap is the no-signal period. The moving average window length of the noise floor energy is set to 10 sampling points with a step size of 1 sampling point to suppress occasional impulse noise. When the direct path signal-to-noise ratio is less than 5 dB, the corresponding original ranging value is marked as invalid and does not participate in the subsequent weighted least squares solution. In the calculation of the multipath scattering factor, the total number of time delay taps N is equal to the number of sampling points of the channel impulse response energy distribution, usually 128 or 256. When the multipath scattering factor is greater than 0... When the multipath scattering factor is less than 0.8, the anchor point is considered to be in a near-line-of-sight environment with a ranging error of less than 10 cm. When the multipath scattering factor is less than 0.3, the anchor point is considered to be in a severely non-line-of-sight environment with a ranging error that may exceed 50 cm. In this case, the instantaneous confidence weight of the anchor point will be further reduced. The calibration of the reference direct path signal-to-noise ratio must be carried out in a constant temperature and humidity environment of 25 degrees Celsius and 50% humidity to ensure the repeatability of the calibration results. After the instantaneous confidence weight is calculated, all anchor points are sorted from largest to smallest according to their weights. Step S2 will select the top three anchor points as reference base stations. If there are fewer than three visible anchor points, only the actual existing anchor points will be selected, and the number of equations for weighted least squares solution will be adjusted accordingly.

[0051] The process of establishing a pre-binding mapping table between StarSpark SLP anchor point identifiers and the 3D coordinates and elevation of hanging points in the BIM model is as follows:

[0052] During the system initialization phase, through on-site surveys or import from the BIM model, the unique hardware identifier of each StarSLP anchor point is mapped one-to-one with the 3D coordinates and elevation of the corresponding hanging point in the BIM model, forming a quintuple record consisting of the StarSLP anchor point identifier, the three components of the hanging point's 3D coordinates, and the hanging point's elevation. On-site surveys include using a total station or RTK-GPS to measure the actual geographical location of each hanging point, and calibrating the measurement results against the theoretical coordinates in the BIM model to ensure an error of less than 1 cm. Importing from the BIM model involves directly reading the IFC or XML format file exported from the BIM software, extracting the design coordinates and elevation of the hanging points. All quintuple records are compiled into a static pre-binding mapping table, which is persistently stored throughout the monitoring cycle and only accessed manually or automatically when the StarSLP anchor point layout changes. The automatic update method is as follows: When a StarShan SLP anchor point is moved or added, the system receives the self-reported identifier change message from the anchor point through the StarShan SLE link, and then re-executes the on-site survey or BIM import process to refresh the mapping table; the three-dimensional coordinates of the anchor points in the pre-bound mapping table are based on the Earth coordinate system or the engineering coordinate system, and the anchor point elevation is based on the altitude or the relative height of the floor, so that subsequent steps can directly query the corresponding spatial location information based on the StarShan SLP anchor point identifier, providing an accurate spatial reference for determining the anchor point's ownership; the Earth coordinate system uses WGS84 latitude and longitude plus ellipsoidal height, and the engineering coordinate system uses the local independent coordinate system (such as Beijing 54 or Xi'an 80) plus normal height; the elevation is uniformly based on the reference surface consistent with the BIM model, such as the ground floor of the building as the zero point or absolute elevation, to avoid positioning deviations caused by the mixing of coordinate systems;

[0053] The pre-binding mapping table is repeatedly queried in step S3: after step S2 calculates the estimated three-dimensional coordinates of the left hook and the right hook, step S3 traverses the three-dimensional coordinates of all hanging points in the mapping table, calculates the Euclidean distance one by one, and determines the hook ownership based on whether the minimum distance is less than 30 centimeters; at the same time, the hanging point elevation in the mapping table is used in step S4 to calculate the actual hanging height difference in order to determine whether there is a violation of low hanging height use.

[0054] During on-site surveying, the total station's angle measurement accuracy should be no less than 2 seconds, and its distance measurement accuracy no less than 2 mm + 2 ppm; the RTK-GPS's planar positioning accuracy should be better than 1 cm, and its elevation accuracy better than 2 cm; when importing the BIM model, the attribute set of the anchor point object in the IFC file should include the fields "Position.X", "Position.Y", "Position.Z", and "Elevation", with corresponding tags in the XML file; the error calibration method is to use the difference between the measured coordinates on-site and the theoretical coordinates in the BIM as the system deviation, and to perform translation correction on all anchor point coordinates in the mapping table so that the maximum residual error is less than 1 cm; during automatic updates, the identifier change message transmitted via the StarSpark SLE link includes the old anchor point identifier, the new anchor point identifier, and the change type (movement or addition), and the monitoring platform automatically triggers the mapping table refresh process upon receiving it; the pre-bound mapping table uses hash table data. The structure uses the SLP anchor point identifier as the key and the five-tuple record as the value to achieve a query with O(1) time complexity. In step S3, when traversing the mapping table, if a certain hanging point is determined to belong, the hanging point identifier is bound to the hook identifier, and the anchor points near the hanging point are queried first in the subsequent monitoring cycle to improve the calculation efficiency. In step S4, the calculation formula for the actual hanging height difference is: actual hanging height difference = waist absolute height × cos(pitch angle) − belonging hanging point elevation. The allowable error band is set to 10 cm, that is, when the actual hanging height difference is greater than 10 cm, it is judged as a violation of low hanging and high use. The value of the allowable error band takes into account the barometer elevation measurement error (usually ±2 meters), IMU pitch angle measurement error (usually ±5 degrees) and BIM model elevation error (usually ±1 cm). After on-site test verification, the 10 cm threshold can effectively detect the low hanging and high use behavior under the premise that the false alarm rate is less than 1%.

[0055] In this embodiment, it is necessary to specifically explain the process of selecting the first three star flash SLP anchor points as reference base stations in step S2 according to the instantaneous confidence weight in descending order:

[0056] From all visible star-flash SLP anchor points output in step S1, sort the instantaneous confidence weight of each star-flash SLP anchor point from largest to smallest, and select the top three star-flash SLP anchor points as reference base stations; if the number of visible star-flash SLP anchor points is less than three, then all visible star-flash SLP anchor points are used as reference base stations. The selection criteria are: the larger the instantaneous confidence weight of the star-flash SLP anchor point, the higher the reliability of its ranging result in the current multipath environment. Prioritizing the selection can reduce the calculation error. The selection result provides a high-quality subset of anchor points for subsequent correction and calculation.

[0057] After selection, for each reference base station, the original ranging value obtained in step S1 is corrected for multipath error using the direct path signal-to-noise ratio (SNR) calculated by the reference base station in step S1, to obtain the corrected ranging value. The specific process of multipath error correction is as follows: a maximum possible multipath error correction amount and an SNR attenuation constant are preset; the negative direct path SNR of the natural constant is divided by the SNR attenuation constant raised to the power of the exponential operation result; the maximum possible multipath error correction amount is multiplied by the exponential operation result to obtain the multipath error correction amount; the multipath error correction amount is subtracted from the original ranging value, and the difference is the corrected ranging value; the maximum possible multipath error correction amount ranges from 0.3 meters to 0.8 meters, and the SNR attenuation constant ranges from 5 dB to 15 dB. These two parameters are determined by offline calibration based on the typical multipath delay of the on-site steel structure environment.

[0058] The specific calibration method for the maximum possible multipath error correction is as follows: In a typical steel structure working environment (such as the angle steel tower body of a 220 kV transmission tower), the Star Flash SLP tag is placed at a known distance anchor point, and multiple sets of deviations between the original distance measurement values ​​and the actual distance are collected. The maximum value of the deviation is taken as the maximum possible multipath error correction. For example, in the angle steel tower environment, the measured maximum deviation is about 0.6 meters, so this value is set as the maximum possible multipath error correction. The calibration method for the signal-to-noise ratio attenuation constant is as follows: In the same environment, the degree of obstruction between the tag and the anchor point is changed, and the distance measurement deviation under different direct path signal-to-noise ratios is recorded. The attenuation constant that minimizes the fitting error of the exponential model is determined by curve fitting. For example, the signal-to-noise ratio attenuation constant obtained by fitting in the tower environment is about 10 dB.

[0059] The design principle of the exponential decay model is as follows: multipath error increases nonlinearly with the deterioration of signal-to-noise ratio. The error increases rapidly at low signal-to-noise ratio and tends to saturate at high signal-to-noise ratio. The exponential form can accurately characterize this law. Compared with the linear correction model, it provides a larger compensation in the low signal-to-noise ratio region and avoids over-correction in the high signal-to-noise ratio region, thus obtaining better ranging accuracy across the entire signal-to-noise ratio range.

[0060] After the distance measurement value is corrected and calculated, it is packaged together with the corresponding reference base station identifier, the three-dimensional coordinates of the hanging point in the BIM model and the instantaneous confidence weight, and passed to step S3 as the output data of step S2.

[0061] The specific steps for constructing a weighted least squares objective function using corrected ranging values ​​and the 3D coordinates of the anchor points in the BIM model of the reference base station are as follows:

[0062] Using the three components of the hook's three-dimensional coordinates to be determined as unknown variables, we first calculate the difference between the corrected ranging value of each reference base station and the Euclidean distance between the hook's three-dimensional coordinates to be determined and the hanging point's three-dimensional coordinates in the BIM model of the reference base station, and obtain the residual corresponding to each reference base station.

[0063] Next, the residuals corresponding to each reference base station are squared, and then the squared residuals are multiplied by the instantaneous confidence weights of the reference base stations calculated in step S1 to obtain the weighted residual squares corresponding to each reference base station.

[0064] Finally, the weighted residual squares corresponding to all reference base stations are summed to obtain a total sum. Minimizing this total sum is used as the optimization objective, and a weighted least squares objective function is constructed. The instantaneous confidence weight controls the contribution of each reference base station to the optimization objective. Reference base stations with larger instantaneous confidence weights have a larger proportion of their corresponding weighted residual squares in the total sum, thus playing a dominant role in the solution process. This makes the calculated hook 3D coordinates more likely to satisfy the ranging constraints of high-confidence reference base stations. The hook 3D coordinates refer to the estimated values ​​of the left and right hook 3D coordinates in step S3. In step S2, the unknowns to be solved have three components corresponding to the east, north, and elevation coordinates in the Earth coordinate system or engineering coordinate system, respectively. The calculation of the residuals reflects the degree of inconsistency between the corrected ranging value and the geometric distance. The smaller the residual, the closer the currently assumed hook coordinates are to the true position. The role of the instantaneous confidence weight is to allow the reference base station with higher ranging confidence to exert a greater influence on the solution results, thereby suppressing the error caused by low-quality ranging. The weighted least squares objective function is a nonlinear function about the three-dimensional coordinates of the hook, which needs to be solved by an iterative method. This is why the Gauss-Newton iteration method is used in step S3.

[0065] The constructed weighted least squares objective function, together with the corrected ranging value of the reference base station, the three-dimensional coordinates of the hanging point in the BIM model, and the instantaneous confidence weight, are used as the output of step S2 for numerical solution in step S3.

[0066] In this embodiment, it is specifically necessary to explain that in step S3, the Gauss-Newton iteration method is used to solve the weighted least squares objective function to obtain the estimated three-dimensional coordinates of the left hook and the right hook. Specifically:

[0067] Using the three components of the three-dimensional coordinates of the left hook as unknown variables, the corrected ranging values ​​of the three reference base stations selected in step S2, the three-dimensional coordinates of the hook point in the BIM model, and the instantaneous confidence weights are used as inputs. The weighted least squares objective function constructed in step S2 is applied, and the reference base station data corresponding to the left hook is substituted. An adaptive damping factor is introduced. The damping factor is multiplied by the identity matrix and added to the Hessian matrix approximation term of the Gauss-Newton method to form an augmented method equation. Then, the augmented method equation is solved to obtain the coordinate update. The coordinate estimate of the next iteration is obtained by adding the estimated three-dimensional coordinates of the left hook in the current iteration step to the coordinate update. After each iteration, the size of the damping factor is dynamically adjusted according to the ratio of the actual descent to the predicted descent: when the ratio is greater than 0.75, the damping factor is reduced by half; when the ratio is less than 0.25, the damping factor is doubled. The iteration termination condition is that any one of the following three conditions is satisfied first: the gradient norm is less than 0.0 ... The conditions of one meter, coordinate update modulus less than one-thousandth of a meter, more than twenty iterations, and gradient norm threshold of one ten-thousandth of a meter (i.e., 0.0001 meters) correspond to the first-order optimality condition of the objective function. When the gradient is close to zero, the solution has approached the stationary point. The coordinate update threshold of one-thousandth of a meter (i.e., 0.001 meters) ensures that the convergence accuracy of the position solution reaches the millimeter level, which is sufficient to meet the centimeter-level accuracy required for the determination of the attachment point. The maximum number of iterations of twenty takes into account the real-time requirements: on the embedded processor, the time for a single iteration is about 2 milliseconds, and the total time for twenty iterations does not exceed 40 milliseconds, which can meet the low latency requirements of the seat belt monitoring system. If convergence is still not achieved after twenty iterations, the current coordinate estimate is forcibly output, and unreliable results are filtered out in subsequent steps through the dummy attachment determination mechanism. After convergence, the three-dimensional coordinate estimate of the left hook is obtained. The three-dimensional coordinate estimate of the right hook is solved independently using the same process, using the data of the three reference base stations selected in step S2.

[0068] The initial value of the adaptive damping factor is set to 0.01. This value is chosen based on the following considerations: if the initial value is too large, it will lead to a small step size and slow convergence in the early stage of iteration; if the initial value is too small, it may cause the iteration to diverge when it is far from the true value; 0.01 has shown good robustness in a large number of simulations and field tests, and can stably start the iteration in various multipath environments; the adjustment thresholds of 0.75 and 0.25 for the damping factor are the classic recommended values ​​of the Levenberg-Marquardt algorithm, which correspond to the two situations of "higher gain ratio, which can accelerate convergence" and "lower gain ratio, which requires stronger regularization", respectively; in practical applications, when the gain ratio is between 0.25 and 0.75, the damping factor remains unchanged to maintain a stable iteration rhythm.

[0069] The ratio of the actual decrease to the predicted decrease is calculated as follows: First, calculate the difference between the function value of the weighted least squares objective function constructed in step S2 before and after this iteration, then subtract half of the product of the residual vector and the coordinate update to get the numerator; the denominator is the absolute value of half of the difference between the residual vector and the product of the Jacobian matrix and the coordinate update; the ratio is obtained by dividing the numerator by the denominator.

[0070] The weighted least squares objective function constructed in step S2 has a value of half multiplied by the product of the weight diagonal matrix, the residual vector, and the transpose of the residual vector;

[0071] The components of the residual vector are the corrected ranging value of the reference base station minus the Euclidean distance between the three-dimensional coordinates of the left hook to be determined and the three-dimensional coordinates of the hook point in the corresponding reference base station BIM model. Taking the left hook as an example, assuming the corrected ranging values ​​of the three reference base stations are 2.35 meters, 2.42 meters, and 2.38 meters, the estimated three-dimensional coordinates of the left hook in the current iteration step are (1.00, 0.50, 0.30) meters, and the three-dimensional coordinates of the hook point in the three reference base station BIM models are (1.10, 0.60, 0.40), (1.05, 0.55, 0.45), and (0.95, 0.65, 0.35) meters, the calculated Euclidean distances are 0.173 meters, 0.091 meters, and 0.174 meters, and the corresponding residuals are 2.177 meters, 2.329 meters, and 2.206 meters, respectively. The larger the residual, the more serious the deviation between the current coordinate estimate and the ranging constraint.

[0072] The elements of the Jacobian matrix are the partial derivatives of the three-dimensional coordinates of the left hook with respect to the corrected ranging value of the corresponding reference base station.

[0073] The gradient norm is the infinite norm of the product of the residual vector and the weight diagonal matrix and the Jacobian matrix;

[0074] The magnitude of the coordinate update quantity is the square root of the sum of the squares of the coordinate update quantity components;

[0075] The specific operation for traversing the 3D coordinates of the hanging points in the BIM model of the pre-bound mapping table is as follows:

[0076] The estimated 3D coordinates of the left hook are compared with the 3D coordinates of each hook point in the BIM model of the pre-binding mapping table using Euclidean distance calculations to obtain a set of distance values. The minimum value in this set is identified and recorded as the minimum Euclidean distance of the left hook. The minimum Euclidean distance of the left hook is compared with 30 cm. If the minimum Euclidean distance of the left hook is less than 30 cm, the left hook is determined to belong to the hook point corresponding to the minimum distance, and the left hook's SLP anchor point identifier and the hook point elevation in the BIM model are output. The 30 cm threshold is set based on the following engineering considerations: the minimum spacing between adjacent hook points in the BIM model is usually 50 cm to 80 cm, and 30 cm is exactly less than the minimum. The spacing is half of the distance to ensure the uniqueness of the attribution determination—that is, any hook can only be matched with one attachment point at most, and there will be no situation where a hook meets the attribution conditions of multiple attachment points at the same time; at the same time, 30 cm is also greater than the typical error range of the coordinate calculation in step S3 (usually 10 to 20 cm), so even if there is a certain degree of positioning error, the correct attachment point can still be accurately identified; if the minimum Euclidean distance of the left hook is not less than 30 cm, the left hook is determined to be in a virtual attachment state. At this time, the left hook is not attached to any legal attachment point, and may be in an abnormal situation such as being suspended in mid-air, attached to a non-standard component, or having a sensor malfunction; the virtual attachment state will trigger a safety reminder, but it does not affect the independent determination of the right hook;

[0077] If the minimum Euclidean distance of the left hook is not less than 30 cm, the left hook is determined to be in a virtual hanging state, and no ownership information is output. The right hook is processed in the same way, and the estimated three-dimensional coordinates of the right hook are used to replace the estimated three-dimensional coordinates of the left hook to obtain the ownership determination result of the right hook. The ownership determination results of the left hook and the right hook are independent of each other and do not affect each other. The right hook ownership SLP anchor point identifier and the right hook ownership hanging point elevation in the BIM model are output as input data for step S4. The output left hook ownership SLP anchor point identifier is used to uniquely identify the physical anchor point currently attached to the left hook. The identification is consistent with the primary key in the pre-bound mapping table of step S1, which facilitates subsequent querying and tracing; the elevation of the hanging point in the BIM model to which the left hook belongs is in meters, and is used together with the absolute height of the waist measured by the barometer of the waist belt host in step S4 to calculate the actual hanging height difference; the same applies to the right hook; if a hook is in a virtual hanging state, its corresponding ownership identifier and elevation are empty values, and step S4 will skip the relevant violation judgment of the hook during processing, and only perform double hook same point and low hanging high use analysis on hooks with valid ownership; all output data are uploaded to the cloud monitoring platform in real time through the StarShine SLE link to form a complete monitoring record.

[0078] In this embodiment, it is particularly important to explain step S4, where the Euclidean distance between the estimated three-dimensional coordinates of the left hook and the right hook is calculated to obtain the spatial distance. The Euclidean distance is a measure of the straight-line distance between the estimated three-dimensional coordinates of the left hook and the right hook in three-dimensional space, reflecting the actual separation degree of the two hooks in space. The spatial distance is calculated based on the estimated three-dimensional coordinates of the left hook and the right hook output in step S3. Both coordinate values ​​are based on the Earth coordinate system or engineering coordinate system, and the unit is meters. The square root operation of the Euclidean distance can be implemented in the embedded processor through a lookup table or a fast approximation algorithm, with a single calculation time of less than 0.1 milliseconds. The physical meaning of the spatial distance is that when two hooks are mounted on different mounting points, their distance is usually greater than 0.8 meters (the minimum distance between adjacent mounting points in the BIM model). If the distance is less than 0.5 meters, they are very likely mounted on the same mounting point or very close adjacent mounting points, at which point the safety belt loses its dual-point protection function.

[0079] Then, the spatial distance is compared with 50 centimeters. If the spatial distance is less than 50 centimeters, a double-hook same-point alarm is triggered directly. If the spatial distance is not less than 50 centimeters, the absolute value criterion for correcting the difference in ranging values ​​is further executed: the same star-flash SLP anchor point that exists in both reference base stations is searched from the three reference base stations corresponding to the left hook and the three reference base stations corresponding to the right hook. The specific method for finding the common star-flash SLP anchor point is: the intersection operation is performed on the identifier list of the three reference base stations corresponding to the left hook and the identifier list of the three reference base stations corresponding to the right hook to obtain the set of anchor points that serve both the left and right hooks. Since the reference base stations of the left and right hooks are selected independently, the intersection may be an empty set. If the intersection is empty, the ranging consistency criterion cannot be executed, and only the spatial distance criterion is relied upon. If the intersection is not empty, the first one in the intersection is taken. An anchor point (or any one of them) is used as a common anchor point, and the corrected distance measurement values ​​of the anchor point relative to the left hook and the right hook are obtained respectively. The corrected distance measurement values ​​are from the output of step S2 and stored in the temporary variable of step S3. The corrected distance measurement values ​​of the same star flash SLP anchor point relative to the left hook and the right hook are taken respectively, and the absolute value of the difference between the two corrected distance measurement values ​​is calculated. The absolute value is compared with 20 centimeters. If the absolute value is less than 20 centimeters, a double hook same point alarm is triggered. If the spatial distance is not less than 50 centimeters and the absolute value of the difference between the corrected distance measurement values ​​is not less than 20 centimeters, a double hook same point alarm is not triggered. Among them, the spatial distance criterion is easily affected by the coordinate calculation error and will produce missed detection. The absolute value criterion of the difference between the corrected distance measurement values ​​is complementary. The two are logically ORed. As long as either criterion meets the condition, a double hook same point alarm is triggered to improve the sensitivity of double hook same point detection.

[0080] The 20cm threshold is set based on the following analysis: When two hooks are attached to the same hook point, their geometric distances to the same anchor point are theoretically equal. However, due to multipath error and residual correction, the difference in corrected distance values ​​is usually less than 10cm. When two hooks are attached to different hook points, the distance difference is at least equal to the distance difference between the two hook points and the anchor point, which is greater than 20cm in typical scenarios. Therefore, the 20cm threshold can effectively distinguish between hooks attached to the same point and hooks attached to different points, while tolerating residual error in distance correction (usually 10 to 15cm). This threshold complements the 50cm spatial distance threshold, together covering various error scenarios.

[0081] The design intent of the logical OR relation is to trigger an alarm whenever any criterion deems there to be a risk of two hooks being at the same point, thereby maximizing detection sensitivity. For example, when a large coordinate calculation error leads to an overestimation of the spatial distance, the distance consistency criterion may still detect the fact of the same point through a small distance difference. Conversely, when the distance value is subject to sudden interference, the spatial distance criterion can still work independently. This redundant design reduces the false detection rate of two hooks at the same point to below 1% in complex multipath environments.

[0082] The actual hook-up height is obtained by reading the absolute waist height measured by the barometer in the belt unit and the pitch angle output by the inertial measurement unit. The absolute waist height is multiplied by the cosine of the pitch angle. The accuracy of the absolute waist height measurement depends on the barometer's performance; a typical MEMS barometer has an absolute accuracy of ±1 meter at room temperature, but this can be improved to ±0.2 meters through differential barometric pressure measurement (compared to a reference barometer fixed at the BIM model benchmark point). The pitch angle is obtained by fusing the accelerometer and gyroscope in the inertial measurement unit using complementary filtering or Kalman filtering, achieving a static accuracy better than ±2 degrees and a dynamic accuracy better than ±5 degrees. The calculation of the actual hook-up height implicitly assumes that the length of the rigid connecting rod between the hook and the belt unit is a fixed value (typically 0.3 to 0.5 meters), which has already been calculated during system calibration. Since the barometer is offset to zero, no explicit deduction is needed. If the length of the connecting rod is not negligible, the actual formula should be corrected to: actual hook height = absolute waist height × cos(pitch angle) + connecting rod length × sin(pitch angle). However, in this scheme, the length of the connecting rod has been calibrated to zero, so a simplified formula is used. The pitch angle is the tilt angle of the human torso relative to the horizontal plane, and the angle is positive when leaning forward. Since the hook is usually connected above the waist belt, the vertical height of the hook relative to the waist will change when the human body leans forward. Therefore, posture compensation is required. The calculation process assumes that the hook and the waist belt main unit are in the same vertical plane, and the rigid connection length from the hook to the waist is negligible or has been included in the calibration. The actual vertical height of the hook is obtained by projecting the waist height in the tilted state onto the vertical direction through the cosine value.

[0083] Then, the smaller of the hanging point elevations in the BIM models of the left and right hooks is selected as the minimum allowable height. The selection logic for the minimum allowable height is based on safety principles: the protection height of the safety belt system is determined by the lower of the hanging points, because if a hook is hung too low, that hook will bear the main impact when a worker falls, while the protection of the higher hook is limited. For example, if the elevation of the left hook is 5.0 meters and the elevation of the right hook is 4.8 meters, then the minimum allowable height is 4.8 meters. If a hook is in a dummy hanging state (no assigned elevation), then the elevation of the other valid hook is taken as the minimum allowable height. If both hooks are dummy hanging, the lower hanging height is skipped, and only the dummy hanging state is retained as a reminder. The minimum protection height of the safety belt system is determined by the lower hook. If one of the hooks is hung too low, the overall protection fails. The difference between the actual hanging height and the minimum allowable height is obtained by subtracting the minimum allowable height from the actual hanging height.

[0084] The actual hanging height difference was compared with a pre-set allowable error band of ten centimeters. This pre-set allowable error band was determined through on-site testing based on three limit values: the maximum absolute error of the barometer, the equivalent height error caused by the pitch angle measurement error, and the maximum elevation error of the BIM model. The specific values ​​and synthesis methods of the three error limit values ​​are as follows: the maximum absolute error of the barometer is ±2 meters (based on the drift limit of the undifferentiated MEMS barometer under the worst temperature environment); the equivalent height error caused by the pitch angle measurement error is calculated according to the most unfavorable situation: when the waist absolute height is 1.5 meters and the pitch angle error is 5 degrees, the equivalent height error is 1.5 × (1 - cos5°) ≈ 0.00. The maximum error in the BIM model elevation is ±0.01 meters (due to residual error after on-site survey and calibration). If simply superimposed linearly, the error band would be approximately 2.03 meters, but this value is too large and would lead to missed detections of numerous instances of low-mounted, high-use structures. In actual engineering, the following measures significantly reduce the error: differential measurement (compared to a benchmark barometer) is used to reduce the absolute error to ±0.2 meters; zero-velocity correction and magnetometer assistance are used in the inertial measurement unit to reduce the pitch angle error to ±2 degrees; the BIM model elevation is averaged through multiple measurements to reduce the error to ±0.005 meters. After comprehensive analysis, the theoretical lower limit of the error band is approximately 0.205 meters. To further reduce the false alarm rate, on-site testing verified that a value of 1... A 0 cm tolerance band effectively detects low-hook, high-use behavior with a false alarm rate below 1%. The 10 cm value includes a safety margin, ensuring that even unmodeled error sources won't lead to missed detections of dangerous situations. If the actual hook height difference is less than negative 10 cm, it's considered a violation, and the low-hook, high-use violation flag is set. If the actual hook height difference is not less than negative 10 cm, it's considered a normal hook connection, and the low-hook, high-use violation flag is cleared. Finally, the left hook's SLP anchor point identifier, the right hook's SLP anchor point identifier, the double hook same-point violation flag, and the low-hook, high-use violation flag are integrated into a single structured alarm message, uploaded to the cloud monitoring platform via the SLE link. The format uses JSON or Protobuf encoding and includes the following fields: timestamp (milliseconds), device number, left hook ownership identifier, right hook ownership identifier, double hook same point indicator (Boolean), low hook high use indicator (Boolean), spatial distance value (floating-point, meters), and actual hooking height difference (floating-point, meters). Alarm information is sent via the StarShine SLE link in a broadcast or point-to-point manner. StarShine SLE operates in the 2.4GHz ISM band, supports a maximum data rate of 2Mbps, and the size of a single alarm message does not exceed 200 bytes, with a transmission latency of less than 10 milliseconds. After receiving the alarm, the cloud monitoring platform can mark the location of the violation on the map in real time and push SMS or APP notifications to the management personnel.

[0085] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0086] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0087] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0088] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0089] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0090] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0091] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for detecting seat belt attachment points at the centimeter level, characterized in that, Specifically, the steps include the following: Step S1: In response to the Star Flash SLP tag being triggered and woken up by the event, obtain the original ranging value, received signal strength indication and channel impulse response energy distribution of each Star Flash SLP anchor point. Calculate the instantaneous confidence weight of each Star Flash SLP anchor point based on the direct path signal-to-noise ratio and multipath scattering factor in the channel impulse response energy distribution, and establish a pre-binding mapping table between the Star Flash SLP anchor point identifier and the three-dimensional coordinates and elevation of the hanging point in the BIM model. Step S2: Select the first three star flash SLP anchor points as reference base stations in descending order of instantaneous confidence weights. Use the direct path signal-to-noise ratio to correct the original ranging value and correct the ranging value. Construct a weighted least squares objective function with the corrected ranging value and the three-dimensional coordinates of the hanging point in the BIM model of the reference base station. Step S3: Solve the weighted least squares objective function using the Gauss-Newton iterative method to obtain the estimated three-dimensional coordinates of the left hook and the right hook; traverse the three-dimensional coordinates of the hanging points in the BIM model in the pre-binding mapping table, calculate the Euclidean distance between the estimated three-dimensional coordinates of the left hook and the three-dimensional coordinates of the hanging points in each BIM model, and the Euclidean distance between the estimated three-dimensional coordinates of the right hook and the three-dimensional coordinates of the hanging points in each BIM model. When the minimum Euclidean distance of the left hook is less than 30 cm, determine the ownership of the left hook and output the star-flash SLP anchor point identifier of the left hook and the elevation of the hanging point in the BIM model to which the left hook belongs. When the minimum Euclidean distance of the right hook is less than 30 cm, determine the ownership of the right hook and output the star-flash SLP anchor point identifier of the right hook and the elevation of the hanging point in the BIM model to which the right hook belongs. Otherwise, determine the state of virtual hanging. Step S4: Calculate the spatial distance between the estimated three-dimensional coordinates of the left hook and the estimated three-dimensional coordinates of the right hook. When the spatial distance is less than 50 cm or the absolute value of the difference between the corrected distance measurements of the left and right hooks from the same StarShock SLP anchor point is less than 20 cm, trigger a double hook same point alarm. At the same time, read the absolute waist height measured by the barometer in the waist belt host and the pitch angle output by the inertial measurement unit. Combine the smaller of the hanging point elevation in the BIM model to which the left hook belongs and the hanging point elevation in the BIM model to which the right hook belongs to calculate the actual hanging height difference. When the actual hanging height difference exceeds the allowable error zone, it is determined to be a violation of low hanging and high use. The StarShock SLP anchor point identifier of the left hook, the StarShock SLP anchor point identifier of the right hook, the double hook same point violation mark, and the low hanging and high use violation mark are integrated into alarm information and uploaded to the cloud monitoring platform through the StarShock SLE link.

2. The safety belt monitoring method for centimeter-level attachment point determination according to claim 1, characterized in that: In step S1, in response to the Star Flash SLP tag being triggered and awakened by an event, the Star Flash SLP tag initiates a ranging request to all visible Star Flash SLP anchor points; each Star Flash SLP anchor point returns the original ranging value, the received signal strength indication, and the channel impulse response energy distribution; the channel impulse response energy distribution is composed of an ordered set of energy amplitudes of several sampling points arranged in chronological order, and each sampling point corresponds to a specific time delay tap and its energy amplitude.

3. The safety belt monitoring method for centimeter-level attachment point determination according to claim 2, characterized in that: The process of calculating the instantaneous confidence weight of each starburst SLP anchor point based on the direct path signal-to-noise ratio and multipath scattering factor in the channel impulse response energy distribution is as follows: The specific calculation process of the direct path signal-to-noise ratio is as follows: the value with the largest energy amplitude in the channel impulse response energy distribution is recorded as the peak energy, the average value of the energy amplitude of all delay taps during the no-signal period is recorded as the noise floor energy, and the quotient obtained by dividing the peak energy by the noise floor energy is the direct path signal-to-noise ratio. Then, calculate the square of the sum of the absolute values ​​of the energy amplitudes of all delay taps from the channel impulse response energy distribution, and divide it by the product of the total number of delay taps and the sum of the squares of the energy amplitudes of the delay taps to obtain the multipath dispersion factor. Then, the direct path signal-to-noise ratio is divided by the pre-calibrated reference direct path signal-to-noise ratio, multiplied by the multipath dispersion factor, and then multiplied by the square of the ratio of the current received signal strength indication to the maximum received signal strength indication among all visible star flash SLP anchor points. The resulting product is used as the instantaneous confidence weight of the star flash SLP anchor point.

4. The safety belt monitoring method for centimeter-level attachment point determination according to claim 3, characterized in that: The process of establishing a pre-binding mapping table between the StarSpark SLP anchor point identifier and the three-dimensional coordinates and elevation of the hanging point in the BIM model is as follows: During the system initialization phase, through on-site surveys or import from the BIM model, the unique hardware identifier of each StarSLP anchor point is mapped one-to-one with the three-dimensional coordinates and elevation of the corresponding hanging point in the BIM model, forming a five-tuple record consisting of the StarSLP anchor point identifier, the three components of the hanging point's three-dimensional coordinates, and the hanging point's elevation. All five-tuple records are then compiled into a static pre-binding mapping table, which is persistently stored throughout the entire monitoring cycle and updated manually or automatically only when the StarSLP anchor point layout changes.

5. The safety belt monitoring method for centimeter-level attachment point determination according to claim 4, characterized in that: In step S2, the specific process of selecting the first three star flash SLP anchor points as reference base stations in descending order of instantaneous confidence weight is as follows: From all the visible star flash SLP anchor points output in step S1, sort the instantaneous confidence weight of each star flash SLP anchor point from largest to smallest, and select the top three star flash SLP anchor points as reference base stations; if the number of visible star flash SLP anchor points is less than three, then all visible star flash SLP anchor points are used as reference base stations. After selection, for each reference base station, the original ranging value obtained in step S1 is corrected for multipath error using the direct path signal-to-noise ratio calculated by the reference base station in step S1, to obtain the corrected ranging value. The specific process of multipath error correction is as follows: a maximum possible multipath error correction amount and a signal-to-noise ratio attenuation constant are preset; the negative direct path signal-to-noise ratio of the natural constant is divided by the power of the signal-to-noise ratio attenuation constant to obtain the exponential calculation result; the maximum possible multipath error correction amount is multiplied by the exponential calculation result to obtain the multipath error correction amount; the multipath error correction amount is subtracted from the original ranging value, and the difference is the corrected ranging value.

6. The safety belt monitoring method for centimeter-level attachment point determination according to claim 5, characterized in that: The specific steps for constructing a weighted least squares objective function based on the corrected ranging value and the three-dimensional coordinates of the anchor point in the BIM model of the reference base station are as follows: Using the three components of the hook's three-dimensional coordinates to be determined as unknown variables, we first calculate the difference between the corrected ranging value of each reference base station and the Euclidean distance between the hook's three-dimensional coordinates to be determined and the hanging point's three-dimensional coordinates in the BIM model of the reference base station, and obtain the residual corresponding to each reference base station. Next, the residuals corresponding to each reference base station are squared, and then the squared residuals are multiplied by the instantaneous confidence weights of the reference base stations calculated in step S1 to obtain the weighted residual squares corresponding to each reference base station. Finally, the weighted residual squares corresponding to all reference base stations are summed to obtain a total; We construct a weighted least squares objective function with the goal of minimizing the sum.

7. The safety belt monitoring method for centimeter-level attachment point determination according to claim 6, characterized in that: In step S3, the Gauss-Newton iterative method is used to solve the weighted least squares objective function to obtain the estimated three-dimensional coordinates of the left hook and the right hook. Specifically: Using the three components of the three-dimensional coordinates of the left hook to be determined as unknown variables, the corrected ranging values ​​of the three reference base stations selected in step S2, the three-dimensional coordinates of the hook point in the BIM model, and the instantaneous confidence weights are used as inputs. The weighted least squares objective function constructed in step S2 is applied, and the reference base station data corresponding to the left hook is substituted. An adaptive damping factor is introduced. The damping factor is multiplied by the identity matrix and added to the Hessian matrix approximation term of the Gauss-Newton method to form an augmented method equation. Then, the augmented method equation is solved to obtain the coordinate update. The coordinate estimate of the next iteration is obtained by adding the estimated three-dimensional coordinates of the left hook in the current iteration step to the coordinate update. After each iteration, the size of the damping factor is dynamically adjusted according to the ratio of the actual drop to the predicted drop: when the ratio is greater than 0.75, the damping factor is reduced by half; when the ratio is less than 0.25, the damping factor is doubled. The iteration termination condition is that any one of the following three conditions is met first: the gradient norm is less than one ten-thousandth of a meter, the magnitude of the coordinate update is less than one thousandth of a meter, and the number of iterations exceeds twenty. After convergence, the estimated three-dimensional coordinates of the left hook are obtained. The three-dimensional coordinates of the right hook are estimated independently using the same process, using data from the three reference base stations selected in step S2.

8. The safety belt monitoring method for centimeter-level attachment point determination according to claim 7, characterized in that: The specific operation of traversing the three-dimensional coordinates of the hanging points in the BIM model of the pre-bound mapping table is as follows: The estimated 3D coordinates of the left hook are compared with the 3D coordinates of the hook point in the BIM model of each hook point in the pre-binding mapping table to calculate the Euclidean distance, resulting in a set of distance values. The minimum value is found from the set of distance values ​​and recorded as the minimum Euclidean distance of the left hook. The minimum Euclidean distance of the left hook is compared with 30 cm. If the minimum Euclidean distance of the left hook is less than 30 cm, it is determined that the left hook belongs to the hook point corresponding to the minimum distance, and the left hook belongs to the Star Flash SLP anchor point identifier and the hook point elevation in the BIM model to which the left hook belongs is output. If the minimum Euclidean distance of the left hook is not less than 30 cm, the left hook is determined to be in a dummy state and no ownership information is output; the right hook is processed in the same way, and the estimated three-dimensional coordinates of the right hook are used to replace the estimated three-dimensional coordinates of the left hook to obtain the ownership determination result of the right hook.

9. The safety belt monitoring method for centimeter-level attachment point determination according to claim 8, characterized in that: In step S4, the Euclidean distance between the estimated three-dimensional coordinates of the left hook and the estimated three-dimensional coordinates of the right hook is calculated to obtain the spatial distance; Then, the spatial distance is compared with 50 centimeters. If the spatial distance is less than 50 centimeters, a double-hook same-point alarm is triggered directly. If the spatial distance is not less than 50 centimeters, the absolute value criterion of the difference between the corrected ranging values ​​is further executed: the same star-flash SLP anchor point that exists in both reference base stations is found from the three reference base stations corresponding to the left hook and the three reference base stations corresponding to the right hook. The corrected ranging values ​​of the same star-flash SLP anchor point for the left hook and the right hook are taken respectively. The absolute value of the difference between the two corrected ranging values ​​is calculated and compared with 20 centimeters. If the absolute value is less than 20 centimeters, a double-hook same-point alarm is triggered. If the spatial distance is not less than 50 centimeters and the absolute value of the difference between the corrected ranging values ​​is not less than 20 centimeters, a double-hook same-point alarm is not triggered.

10. A method for detecting safety belt attachment points at the centimeter level according to claim 9, characterized in that: In step S4, the absolute height of the waist measured by the barometer in the waist belt host and the pitch angle output by the inertial measurement unit are read. The absolute height of the waist is multiplied by the cosine of the pitch angle to obtain the actual hanging height. Then, select the smaller of the hanging point elevations in the BIM model belonging to the left hook and the right hook, and record it as the minimum allowable height; subtract the minimum allowable height from the actual hanging height to obtain the actual hanging height difference; The actual hanging height difference is compared with the preset allowable error range, which is 10 centimeters. If the actual hanging height difference is less than negative 10 centimeters, it is judged as a violation of low hanging and high use, and the low hanging and high use violation sign is set. If the actual hanging height difference is not less than negative 10 centimeters, it is judged as normal hanging, and the violation mark for low hanging and high use is cleared to zero; finally, the left hook belonging to the StarShan SLP anchor point mark, the right hook belonging to the StarShan SLP anchor point mark, the violation mark for double hooks at the same point and the violation mark for low hanging and high use are integrated into a structured alarm message, which is uploaded to the cloud monitoring platform through the StarShan SLE link.

Images