Double-base-station gantry crane positioning method and system based on event records

Abstract

The invention provides a double-base-station gantry crane positioning method and system based on event records, and relates to the technical field of industrial equipment positioning and navigation. According to the method, the distance measurement values of the two base stations in the same straight line direction are received, a state estimation algorithm including speed entropy decision, longitudinal check and noise self-adaptive adjustment is adopted, more reliable observation values can be intelligently recognized and screened out, motion state changes such as equipment starting and stopping and speed changing can be dynamically adapted, and the method is more reliable. And the direct influence of the non-line-of-sight error and the abnormal jump on the positioning result is obviously inhibited. According to the method, state judgment and track compression based on event recognition are combined, a more accurate real-time position is output, and the movement continuity and reasonability of a positioning result are ensured. According to the method, the positioning track of the gantry crane is smoother and more stable and conforms to the actual motion law, the overall precision and reliability of a positioning system are remarkably improved, and the problem that large hoisting equipment such as the gantry crane is inaccurate in equipment positioning is solved.

Description

Technical Field

[0001] This invention relates to the field of industrial equipment positioning and navigation technology, and in particular to a dual-base station gantry crane positioning method and system based on event logging. Background Technology

[0002] With the rapid development of industrial automation and intelligent manufacturing, achieving high-precision, continuous, and reliable position sensing and trajectory tracking for engineering machinery and mobile equipment, especially large lifting equipment such as gantry cranes and bridge cranes, has become a key requirement for improving operational efficiency, ensuring operational safety, and realizing intelligent management. Among many non-visual positioning technologies, ultra-wideband (UWB) technology has been widely used in the field of industrial indoor positioning in recent years due to its advantages such as centimeter-level high precision, strong resistance to multipath interference, and low power consumption.

[0003] UWB positioning technology primarily determines a tag's location by deploying fixed base stations at known locations and measuring the time of flight (ToF) or time difference of arrival (TDOA) of the radio signals between the base station and the mobile tag. However, in real industrial environments (such as storage yards, workshops, and construction sites), the environment is complex and variable, with numerous metal structures, mobile devices, and workers, which can easily obstruct the UWB signal propagation path, resulting in non-line-of-sight (NLOS) errors. NLOS causes a positive deviation in the ranging value (i.e., the measured distance is greater than the actual distance), and in severe cases, the error can reach the meter level, significantly reducing the reliability and accuracy of the positioning system.

[0004] Currently, large lifting equipment such as gantry cranes suffer from inaccurate equipment positioning. Summary of the Invention

[0005] This invention provides a dual-base station gantry crane positioning method and system based on event logging, which solves the problem of inaccurate positioning of large lifting equipment such as gantry cranes.

[0006] In a first aspect, the present invention provides a dual-base station gantry crane positioning method based on event logging. The method includes: receiving a first ranging value from a first base station to a target device and a second ranging value from a second base station to the target device along the same straight direction of the gantry crane; calculating a current state estimate of the target device based on the first and second ranging values ​​using a state estimation algorithm, wherein the state estimation algorithm includes at least one of the following: reliability decision based on velocity entropy, longitudinal consistency verification, and adaptive adjustment of noise parameters; performing event identification based on the current state estimate of the target device and a sequence of historical state estimates to determine the positioning state of the target device, wherein the positioning state includes a stationary state or a moving state; and outputting the positioning result of the target device based on the positioning state and the current state estimate, and performing event-driven trajectory compression and storage on the positioning result.

[0007] In one possible implementation, based on the first and second ranging values, a state estimation algorithm is used to calculate the current state estimate of the target device, including: analyzing velocity entropy based on the first and second ranging values ​​respectively, making a reliability decision, and determining candidate observations; performing longitudinal consistency verification based on the candidate observations and the state estimate from the previous moment to determine valid observations; determining filtered observations based on the valid observations using a Kalman filter update method that adaptively adjusts noise parameters; and determining the current state estimate based on the candidate observations, the valid observations, and / or the filtered observations.

[0008] In one possible implementation, based on the first and second ranging values, velocity entropy is analyzed to make a reliability decision and determine candidate observations, including: calculating the difference between the first and second ranging values; if the difference is less than or equal to a preset error threshold, then selecting either the first or second ranging value as a candidate observation; if the difference is greater than the preset error threshold, then calculating a first velocity entropy based on the historical sequence of the first and second ranging values; calculating a second velocity entropy based on the historical sequence of the second and second ranging values; if the first velocity entropy is greater than the second velocity entropy, then selecting the first ranging value as a candidate observation; if the first velocity entropy is less than the second velocity entropy, then selecting the second ranging value as a candidate observation. In one possible implementation, the step of performing a longitudinal consistency check based on the candidate observation and the state estimate at the previous moment to determine the valid observation includes: calculating the absolute value of the difference between the candidate observation and the state estimate at the previous moment; if the absolute value is less than or equal to a preset motion threshold, then the candidate observation is taken as a valid observation; if the absolute value is greater than the preset motion threshold, then the state prediction is taken as a valid observation, wherein the state prediction is a prediction determined based on the historical state estimate sequence.

[0009] In one possible implementation, the step of determining the filtered observation value based on the effective observation value through an adaptive adjustment of the noise parameter Kalman filter update step includes: adjusting the observation noise covariance based on the entropy ratio of the first velocity entropy to the second velocity entropy; increasing the noise parameter corresponding to the acceleration state in the process noise covariance in response to the detection of a sudden acceleration change in the target device; and performing a Kalman filter update using the adjusted noise parameter to obtain the filtered observation value.

[0010] In one possible implementation, the step of performing event identification and determining the positioning status of the target device based on the current state estimate and the historical state estimate sequence includes: using the current state estimate as the time endpoint, tracing back a historical state estimate sequence for a preset time length; calculating the position change in the historical state estimate sequence; if the position change is less than a stationary determination threshold and the preset time length is greater than or equal to the shortest duration, then the positioning status of the target device is determined to be stationary; otherwise, the positioning status of the target device is determined to be mobile.

[0011] In one possible implementation, the step of outputting the positioning result of the target device based on the positioning state and the current state estimate includes: determining the current timestamp and the location coordinates of the target device based on the current state estimate; and generating the positioning result based on the current timestamp, the location coordinates of the target device, and the positioning state.

[0012] In one possible implementation, the event-driven trajectory compression and storage of the positioning results includes: generating a position coordinate sequence based on the positioning results of the target device over a continuous period; compressing the position coordinate sequence using the Douglas-Peucker algorithm to extract key feature points; identifying and merging adjacent events that are both stationary or both in motion based on the positioning status and timestamps corresponding to the key feature points; storing the merged events in the form of event metadata, which includes at least the event type, start time, end time, and position coordinates of the key feature points, and persistently storing the event metadata in place of the original continuous positioning results.

[0013] In one possible implementation, the method further includes: determining residual energy based on the first ranging value and the second ranging value, and the current state estimate; the residual energy is used to reflect the correction amplitude of the filter; calculating the standard deviation of the position change at adjacent time points based on the historical state estimate sequence in which the current state estimate is located, as a jitter index, the jitter index reflecting the smoothness of the trajectory; inverting the sum of the rangings of the two base stations based on the current state estimate, calculating the deviation between the sum of the rangings of the two base stations and the known baseline distance, as a geometric closure error, the geometric closure error reflecting the smoothness of the trajectory; and counting the number of times the position change at adjacent time points exceeds the motion rationality threshold in the historical state estimate sequence in which the current state estimate is located, as the number of time consistency violation frames, the number of time consistency violation frames reflecting the temporal continuity of the trajectory.

[0014] Secondly, the present invention provides a dual-base station gantry crane positioning device based on event logging. The device includes a communication module and a processing module. The communication module is used to receive a first ranging value from a first base station to a target device and a second ranging value from a second base station to the target device along the same straight direction of the gantry crane. The processing module is used to calculate a current state estimate of the target device based on the first and second ranging values ​​using a state estimation algorithm. The state estimation algorithm includes at least one of the following: reliability decision based on velocity entropy, longitudinal consistency verification, and adaptive adjustment of noise parameters; based on the current state estimate of the target device and a sequence of historical state estimates, perform event identification to determine the positioning state of the target device, where the positioning state includes a stationary state or a moving state; based on the positioning state and the current state estimate, output the positioning result of the target device, and perform event-driven trajectory compression and storage on the positioning result.

[0015] Thirdly, embodiments of the present invention provide a dual-base station gantry crane positioning system based on event logging. The system includes an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor is used to call and run the computer program stored in the memory to perform the steps of the method as described in the first aspect and any possible implementation thereof.

[0016] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, it implements the steps of the method as described in the first aspect and any possible implementation thereof.

[0017] This invention provides a dual-base station gantry crane positioning method and system based on event logging. By receiving ranging values ​​from two base stations along the same straight line, and employing a state estimation algorithm incorporating velocity entropy decision-making, longitudinal verification, and noise adaptive adjustment, this invention can intelligently identify and filter more reliable observations, dynamically adapting to changes in the equipment's motion state such as start-up, shutdown, and speed changes. This significantly suppresses the direct impact of non-line-of-sight errors and abnormal jumps on the positioning results. Combining event-based state judgment with trajectory compression, this invention outputs a more accurate real-time position, ensuring the continuity and rationality of the positioning results. This invention makes the gantry crane's positioning trajectory smoother, more stable, and more consistent with actual motion patterns, significantly improving the overall accuracy and reliability of the positioning system and solving the problem of inaccurate positioning of large lifting equipment such as gantry cranes. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of a dual-base station structure scheme provided in an embodiment of the present invention; Figure 2 This is a flowchart illustrating a dual-base station gantry crane positioning method based on event logging provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of velocity entropy decision-making provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of a vertical consistency verification provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of original trajectory comparison provided by an embodiment of the present invention; Figure 6 This is another schematic diagram of original trajectory comparison provided by an embodiment of the present invention; Figure 7 This is a schematic diagram illustrating the comparison of filtering effects provided in an embodiment of the present invention; Figure 8 This is another comparative diagram of filtering effects provided by an embodiment of the present invention; Figure 9 This is a comparative diagram of evaluation indicators provided by an embodiment of the present invention; Figure 10 This is a schematic diagram of the structure of a dual-base station gantry crane positioning device based on event logging provided in an embodiment of the present invention; Figure 11This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0020] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0021] In the description of this invention, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The term "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Furthermore, "at least one" and "more than one" refer to two or more. The terms "first," "second," etc., do not limit the quantity or order of execution, and "first," "second," etc., do not necessarily imply differences.

[0022] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner to facilitate understanding.

[0023] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or modules is not limited to the steps or modules listed, but may optionally include other steps or modules not listed, or may optionally include other steps or modules inherent to such process, method, product, or device.

[0024] To make the objectives, technical solutions, and advantages of the present invention clearer, the following description will be provided in conjunction with the accompanying drawings and specific embodiments.

[0025] As described in the background section, research on object localization and motion tracking has expanded into the ever-evolving smart world, including the Internet of Things (IoT), smart infrastructure, smart devices, transportation systems, robotics, military applications, smart healthcare, and more.

[0026] Among numerous non-visual positioning technologies, Ultra-Wideband (UWB) positioning technology has attracted much attention due to its unique advantages. It is a radio innovation, unlike currently widely used traditional radio technologies, that calculates the tag's position through ranging / angle measurement results between multiple base stations (BSs) at known locations and the mobile tag. Positioning devices based on UWB radio sensors, such as Kickstarter's Pozyx and DecaWave's DWM 1000, are already widely used in industry. However, its ranging accuracy is significantly affected by non-line-of-sight (NLOS) errors, specifically, in dense and complex environments, the emitted pulse signal fluctuates due to obstructions from moving people or various objects. To address the impact of NLOS errors, a neural dynamic TDOA positioning algorithm based on the Maximum Correlation Criterion (MCC) exists for robust positioning in NLOS environments. This algorithm dynamically adjusts positioning parameters through a neural network to adapt to complex signal environments. There is also an unsupervised LOS / NLOS recognition method based on a two-stage machine learning framework, which effectively identifies LOS and NLOS signals. Furthermore, there is a convex optimization-based method that uses mathematical modeling of the optimization problem to find the optimal positioning solution to minimize the impact of NLOS errors.

[0027] Furthermore, there are currently methods that integrate other measurement systems to assist in positioning. Inertial Measurement Units (IMUs) are measurement tools that use gyroscopes and accelerometers as their main sensing devices. Because they can autonomously calculate positioning without relying on any external information, and the navigation information they generate has good continuity, low noise, high data update rate, and good stability, they are often used in conjunction with UWB positioning technology to compensate for its shortcomings. IMUs possess the characteristics of short-term high accuracy; through data integration, they can obtain attitude, velocity, and position information, providing relatively accurate navigation parameters in a short time. However, their drawback is that errors accumulate and gradually diverge over time, leading to a decrease in positioning accuracy. Combining UWB with IMUs serves two purposes: firstly, the high-precision position information of UWB is used to correct the accumulated errors of the IMU, preventing the errors from diverging indefinitely, thereby improving the reliability of the IMU in long-term navigation; secondly, the short-term high accuracy of the IMU provides continuous and stable navigation parameters to the system when UWB signals are blocked or multipath and NLOS effects occur, reducing the adverse effects caused by UWB signal blockage.

[0028] Ultra-wideband (UWB) technology uses nanosecond-level non-sinusoidal narrow pulses for data transmission, typically operating in the 3.1–10.6 GHz frequency band with a bandwidth greater than 500 MHz. The core of UWB positioning lies in using signal propagation time for distance measurement.

[0029] Inertia is a fundamental property of all bodies of mass. Based on Newton's laws, inertial navigation systems (INS) do not interact with the external environment through any photoelectric means; they can perform continuous three-dimensional positioning and orientation of moving objects solely through the system itself. An inertial navigation system is an autonomous navigation system based on Newtonian mechanics. It uses orthogonally arranged gyroscopes and accelerometers to form an IMU, and employs a recursive trajectory algorithm to calculate the vehicle's position, velocity, attitude, and other navigation parameters in real time.

[0030] The theoretical basis of this system stems from Newton's second law of motion: in an inertial frame of reference, the acceleration vector of a particle is directly proportional to the net external force vector, with the proportionality constant being the reciprocal of the particle's mass, and the direction of acceleration is consistent with the direction of the external force. Its mathematical model can be expressed as: F = m·a, where F represents the net external force vector, m is the particle's mass, and a is the absolute acceleration vector.

[0031] In response to the above technical problems, such as Figure 1 As shown, this invention provides a dual-base station structure scheme. The positioning system proposed in this invention considers a two-dimensional plane from a top-down perspective, as shown... Figure 1 As shown in the diagram. In this invention, data transmission between the base station and the central control unit is achieved through a wireless data transmission system (WDS). The WDS is responsible for transmitting the signal data collected by the base station to the central control unit for processing.

[0032] The received signal is modeled using the Channel Impulse Response (CIR). The CIR describes the delay and attenuation characteristics of a signal as it travels through a wireless channel. The model of the received signal can be expressed as: ; Where r(t) is the received signal, α k τ is the signal attenuation coefficient of the k-th path. k s(t) is the time delay of the k-th path, s(t) is the pulse shape of the transmitted signal, n(t) is additive white Gaussian noise (AWGN), and K is the number of multipath components.

[0033] Then, the central control unit uses the Time-of-Flight (ToF) algorithm to convert the received signal data into real-time observations of the gantry crane's positioning position. Simultaneously, to address the impact of NLOS on UWB signal propagation in this planar area, two base stations are set up along the x-axis, located at both ends of the gantry crane track, to simultaneously receive the x-tag's position. To ensure the accuracy of each acquired x-coordinate and y-coordinate, the y-axis tag uses a separate base station for ranging. In this way, the two-dimensional motion can be simplified into two independent one-dimensional motions. Therefore, for the one-dimensional trajectory recursion, assuming the tag's initial position is S0 and its initial velocity is V0, the velocity Vt of the tag at time t can be obtained by integrating the acceleration a, i.e.: ; Integrating the velocity Vt, we obtain the car's position St at time t: ; Among them, V t Indicates the time of the gantry crane The instantaneous velocity, S t Indicates the time of the gantry crane The displacement (relative to the initial position), 'a' represents the constant acceleration of the gantry crane, 't' is the time variable, accumulating from the initial moment, and 'V0' represents the displacement of the gantry crane at the initial moment ( The initial velocity of ) is S0, which represents the initial position reference term, and dt represents the time derivative, which represents the infinitesimal time interval.

[0034] based on Figure 1 The dual-base station gantry crane positioning system shown is as follows: Figure 2 As shown, this embodiment of the invention provides a dual-base station gantry crane positioning method based on event logging. The method includes steps S101-S104.

[0035] S101, Receive the first ranging value of the first base station to the target equipment and the second ranging value of the second base station to the target equipment in the same straight direction as the gantry crane.

[0036] It should be noted that UWB / INS fusion positioning technology has become a research hotspot for high-precision position sensing in industrial scenarios. Traditional methods generally employ Extended Kalman Filter (EKF) or its variants for sensor data fusion. However, in real industrial environments (such as gantry crane tracks), these methods face two key challenges: First, the reliability of observations. UWB signals can generate ranging errors up to meters in the range under non-line-of-sight (NLOS) conditions. Existing research mainly employs two types of solutions: robust EKF based on residual testing (such as RANSAC outlier removal); and machine learning-based methods. However, these methods have significant drawbacks: residual testing is sensitive to continuous NLOS errors, and machine learning schemes have high computational complexity, making them difficult to implement in real-time on edge devices. Second, there is insufficient adaptability to motion states. The start-up and shutdown of gantry cranes can cause the fixed noise parameters of traditional EKF to become invalid. Although adaptive UKF can dynamically adjust process noise, it does not consider the correlation between observation noise and motion characteristics, and will still produce hysteresis errors in high-speed, abrupt change scenarios.

[0037] Inspired by the concept of entropy as the optimal measure of uncertainty in information theory, this invention proposes velocity entropy as a novel criterion. The core idea is that the true trajectory of a moving object should possess velocity continuity, while NLOS error disrupts this continuity, manifesting as anomalous perturbations in the velocity probability distribution. Based on this, we design the SE-AKF algorithm.

[0038] S102. Based on the first ranging value and the second ranging value, calculate the current state estimate of the target device using a state estimation algorithm.

[0039] In this embodiment of the application, the state estimation algorithm includes at least one of the following: reliability decision based on velocity entropy, longitudinal consistency verification, and adaptive adjustment of noise parameters; As one possible implementation, step S102 can be specifically implemented as steps S1021-S1024.

[0040] S1021. Based on the first ranging value and the second ranging value, analyze the velocity entropy respectively, make a reliability decision, and determine the candidate observation value.

[0041] For example, step S1021 can be implemented as steps A11-A16.

[0042] A11. Calculate the difference between the first distance measurement value and the second distance measurement value.

[0043] A12. If the difference is less than or equal to a preset error threshold, then the first ranging value or the second ranging value is selected as the candidate observation value.

[0044] A13. If the difference is greater than the preset error threshold, then the first velocity entropy is calculated based on the first ranging value and the historical sequence of the first ranging value.

[0045] A14. Calculate the second velocity entropy based on the historical sequence of the second ranging value and the second ranging value.

[0046] A15. If the first velocity entropy is greater than the second velocity entropy, then the first ranging value is selected as the candidate observation value.

[0047] A16. If the first velocity entropy is less than the second velocity entropy, then the second ranging value is selected as the candidate observation value.

[0048] For example, a decision-making method based on velocity entropy analysis includes the following steps.

[0049] Step 1: Pre-screening of observations. First, calculate the distance difference between the two base stations.

[0050] ; in, The difference between the first and second distance measurements. The first distance measurement value, This is the second distance measurement value.

[0051] like This means that the measurements from both base stations are accurate at this time. Any measurement value, such as zk=d, can be used. k BS2, otherwise proceed to the next step. The preset error threshold represents the allowed error threshold.

[0052] Step 2: Velocity entropy decision. If the ranging values ​​are inconsistent, a sliding window is established at that location to calculate the first-order difference of the ranging values ​​from each base station. ; in, Indicates base station In the The velocity difference at each moment Indicates base station In the The original distance measurement value obtained at each time point This represents a first-order difference operator that calculates the difference between adjacent elements in a sequence.

[0053] Then normalize the velocity probability distribution: ; in, This represents the absolute value of the velocity difference, used to eliminate the influence of positive and negative directions. This represents the sum of the absolute values ​​of all velocity differences within the sliding window.

[0054] Finally, calculate the entropy value: ; If HBS1 ≥ HBS2, candidate observations k=dkBS1, otherwise k=dkBS2.

[0055] in, Indicates base station The velocity entropy value within the sliding window. It is a very small positive number used to prevent mathematical undefined errors when the logarithm is 0.

[0056] It's important to note that entropy is a metric in information theory used to measure uncertainty, and velocity entropy reflects the uniformity or uncertainty of velocity data. For example, in real-world scenarios... Figure 3 As shown, points in region A exhibit good consistency and can be used as reliable points in filter prediction. In region B, the observations from the two base stations show significant discrepancy. Therefore, a window of size 30 is established for jz14 and jz15, centered on the point of ambiguity. The sum of the entropy values ​​of the two observation windows is compared. The window with the higher entropy value indicates a more uniform distribution of velocity data within that window. The point with the larger sum of entropy values ​​is selected as a candidate point. Velocity entropy decision-making aims to select the observation we are more inclined to believe when ambiguity arises. For example, in region C, even if both observations deviate from reality, a point is still chosen to proceed to the next step of judgment.

[0057] S1022. Based on the candidate observations and the state estimate from the previous time step, perform a longitudinal consistency check to determine the valid observations.

[0058] For example, step S1022 can be implemented as steps A21-A23.

[0059] A21. Calculate the absolute value of the difference between the candidate observation and the state estimate at the previous time step; A22. If the absolute value is less than or equal to the preset motion threshold, then the candidate observation value is taken as the valid observation value. A23. If the absolute value is greater than the preset motion threshold, the state prediction value is taken as the valid observation value, and the state prediction value is the prediction value determined based on the historical state estimation value sequence.

[0060] An example of a method for longitudinal consistency verification includes the following steps.

[0061] Step 3: Check candidate observations k and historical estimates The difference of k-1.

[0062] If | k - k-1|≤βth (βth is set according to the ranging frequency and experience), accept zk= k; Otherwise, the backup base station data will be rejected and verified. If the backup base station data also deviates from the empirical threshold βth, the collected observations will be abandoned and the predicted value estimated for that point based on the filter will be selected.

[0063] Vertical consistency verification further assesses the points where velocity entropy is determined, and represents the final decision on points of ambiguity. For example... Figure 4 As shown, it is the same Figure 3 For a given time period of data, region D prioritizes the point for velocity entropy decision-making. If the point matches the actual motion characteristics, it is determined as the correct observation value. If both observation values ​​deviate from the normal motion, the value of the point will be predicted based on the historical trajectory and used as the correct observation value, as shown in region E.

[0064] S1023. Based on the effective observations, the filtered observations are determined by an adaptive Kalman filter update method that adjusts the noise parameters.

[0065] For example, step S1023 can be implemented as steps A31-A33.

[0066] A31. Adjust the observation noise covariance based on the entropy ratio of the first velocity entropy to the second velocity entropy.

[0067] A32. In response to the detection of a sudden acceleration change in the target device, increase the noise parameter corresponding to the acceleration state in the process noise covariance.

[0068] A33. Perform Kalman filtering update using the adjusted noise parameters to obtain the filtered observations.

[0069] An exemplary Kalman filter update method for adaptively adjusting noise parameters includes the following steps.

[0070] Step 4: Scaling the observation noise variance based on the entropy ratio: ; Where R0 is the reference noise, , These represent the velocity entropy values ​​of base station 1 and base station 2, respectively. The higher the entropy ratio, the lower the trust in the low-entropy base station. The acceleration term that increases process noise is detected by detecting sudden acceleration changes.

[0071] ; in, The process noise covariance matrix is... This represents the estimated acceleration value at the current moment. This represents the threshold for acceleration abrupt change, used to normalize the magnitude of the change. .

[0072] Step 5: KF Update The Kalman update step, as shown in the following formula, involves calculating the Kalman gain Kk to balance the predicted and measured values ​​for optimal state estimation. This gain is calculated using the predicted state estimation error covariance matrix Pk and the estimated measurement error covariance matrix Rk. Finally, the measured value zk is compared with the estimated measurement value. The difference between k is used to update the state estimate. k. Using the Kalman gain Kk and the covariance matrix of the predicted state estimation error. Perform calculations to update the state estimate covariance. .

[0073] ; in, This represents the Kalman gain matrix at the current moment, used to balance prediction and observation. H represents the prior state estimation covariance matrix, and H is the observation matrix. This represents the prior state estimation vector. It is the posterior state estimation vector (i.e., the final output). This represents the valid observation value at the current moment.

[0074] Step 6: Median Filtering. Median filtering is a nonlinear numerical processing method. Its principle is to replace the value of any point in a sequence with the median value of all points in its neighborhood. To use median filtering for data processing, a window (of odd number) must be selected beforehand. Then, numbers are successively extracted from the input sequence, sorted by size, and the number with the center index is taken as the filtered output. The formula is as follows: ; Here, Med represents the median operation, which takes the median of the data within the window. Indicates the first The sequence data at time step m, where m represents the size of the sliding window.

[0075] The SE-AKF algorithm proposed in this invention innovatively introduces velocity entropy as a quantitative indicator of motion state reliability. It establishes a dual protection mechanism through hardware-level threshold screening and algorithm-level entropy value decision-making combined with vertical verification, which can effectively avoid the impact of single-point failures. At the same time, it improves the algorithm's environmental adaptability by adaptively adjusting noise parameters based on motion state.

[0076] S1024. Based on the candidate observations, the valid observations, and / or the filtered observations, determine the current state estimate.

[0077] For example, in embodiments of the present invention, reliability decisions based solely on velocity entropy can be made to directly determine candidate observations as current state estimates.

[0078] Another exemplary embodiment of the present invention can perform reliability decision-making and longitudinal consistency verification based on velocity entropy, and determine the effective observation value as the current state estimate.

[0079] Another exemplary embodiment of the present invention can perform reliability decision-making based on velocity entropy and adaptively adjust noise parameters to determine the effective observation value as the current state estimate.

[0080] Another exemplary embodiment of the present invention can perform longitudinal consistency verification and adaptive adjustment of noise parameters to determine the effective observation value as the current state estimate.

[0081] Another exemplary embodiment of the present invention can perform reliability decision-making based on velocity entropy, longitudinal consistency verification, and adaptive adjustment of noise parameters, and determine the filtered observation value as the current state estimate.

[0082] S103. Based on the current state estimate of the target device and the historical state estimate sequence, perform event identification to determine the positioning status of the target device.

[0083] In some embodiments, the positioning state includes a stationary state or a moving state.

[0084] As one possible implementation, step S103 can be specifically implemented as steps S1031-S1034.

[0085] S1031. Using the current state estimate as the time endpoint, backtrack the historical state estimate sequence for a preset time length.

[0086] S1032. Calculate the position change in the historical state estimate sequence.

[0087] S1033. If the change in position is less than the static determination threshold, and the duration of the preset time length is greater than or equal to the shortest duration, then the positioning state of the target device is determined to be static.

[0088] S1034. Otherwise, determine the positioning status of the target device as a moving state.

[0089] S104. Based on the positioning status and the current state estimate, output the positioning result of the target device, and perform event-driven trajectory compression and storage on the positioning result.

[0090] As one possible implementation, embodiments of the present invention can determine the current timestamp and the location coordinates of the target device based on the current state estimate; and generate the positioning result based on the current timestamp, the location coordinates of the target device, and the positioning state.

[0091] As one possible implementation, embodiments of the present invention may perform compression and storage according to steps S1041-S1044.

[0092] S1041. Generate a position coordinate sequence based on the positioning results of the target device over a continuous period of time.

[0093] S1042. Based on the position coordinate sequence, the Douglas-Peucker algorithm is used for compression to extract key feature points.

[0094] S1043. Based on the positioning status and timestamp corresponding to the key feature points, identify and merge adjacent events that are both stationary or both in a moving state.

[0095] S1044. The merged events are stored in the form of event metadata, which includes at least the event type, start time, end time, and location coordinates of key feature points. The event metadata is then used to replace the original continuous positioning results for persistent storage.

[0096] It's important to note that in industrial settings, the movement of equipment like gantry cranes typically exhibits segmented patterns (e.g., uniform motion, stationary waiting). Traditional methods continuously record all sensor data (approximately 12GB / day at a 100Hz sampling rate), leading to wasted storage resources and difficulties in later traceability and analysis. However, by analyzing the motion characteristics of industrial equipment, we found that gantry cranes spend 95% of their working time in two typical states: Precise stillness (within a certain error range): remaining stationary during loading / unloading operations or when there is no movement; Uniform motion (which can be considered uniform motion for traceability): during the crane's movement phase.

[0097] Thus, this invention abstracts continuous motion into a discrete event sequence, with each event containing complete motion semantics. Addressing the data redundancy and semantic gaps inherent in traditional continuous sampling methods, this invention proposes an event-driven, three-level processing architecture: Trajectory compression layer: Key feature points are extracted using the Douglas-Peucker (DP) algorithm.

[0098] Event recognition layer: State classification is achieved based on kinematic features.

[0099] Semantic storage layer: Uses event metadata to replace the original data stream.

[0100] The mathematical model of the architecture can be represented as: ; in, This represents the original continuous positioning trajectory data. This represents a compression operation, using the Douglas-Peucker algorithm. This represents an event recognition operation, based on motion state classification. This represents the event merging operation, which merges adjacent events of the same type.

[0101] The event-driven algorithm implementation process is as follows: First, compression is performed. Since the Douglas-Peucker (DP) algorithm can automatically preserve motion state inflection points, it closely aligns with Newton's kinematics principles and strictly guarantees that the maximum deviation between the compressed trajectory and the original path does not exceed a set threshold, fitting the research of this experiment. Therefore, the DP algorithm is used for compression processing. DP is a classic geometric compression method. Its implementation principle is as follows: For a given discrete trajectory sequence P={p1,p2,…,pN}, where pi=(xi,yi,ti) represents the spatial coordinates and timestamp of the i-th sampling point, the DP algorithm first connects the first and last points of the sequence to form a straight line segment L. Then, it calculates the distance from all intermediate points to the straight line L, selects the point with the largest distance, compares it with a threshold, and retains the point as a key point if the distance is greater than the threshold. The mathematical expression is: ; in, Indicates the index arrive sequence of points Perform Douglas-Peucker compression. The compression result only retains the starting point. and the end point , Indicates if Then take the point with the largest distance. Using the segmentation point as the dividing point, the left and right segments are recursively compressed, and the results are merged. The purpose of the event recognition stage is to identify motion events based on the compressed trajectory point sequence. We define an event discrimination function, classifying states as static or moving based on position change and duration. The event classification rule can be expressed as: ; Among them, among them, , These represent the position coordinates at the current time and the previous time, respectively. γ represents the duration of the current time window. static It is the threshold for determining stillness, T min It is the shortest duration threshold.

[0102] The event merging phase further optimizes storage by merging adjacent static events and filtering short-lived events. The merging rule can be expressed as: ; in, γ represents the absolute difference between two key points of an event. merge It is the event merging distance threshold, T merge It is the threshold of the event merging time window.

[0103] This invention provides a dual-base station gantry crane positioning method based on event logging. By receiving ranging values ​​from two base stations along the same straight line, and employing a state estimation algorithm that includes velocity entropy decision-making, longitudinal verification, and noise adaptive adjustment, it can intelligently identify and filter more reliable observations, dynamically adapting to changes in the equipment's motion state such as start-up, shutdown, and speed changes. This significantly suppresses the direct impact of non-line-of-sight errors and abnormal jumps on the positioning results. This invention combines event-based state judgment with trajectory compression to output a more accurate real-time position, ensuring the continuity and rationality of the positioning results. This invention makes the gantry crane's positioning trajectory smoother, more stable, and more consistent with actual motion patterns, significantly improving the overall accuracy and reliability of the positioning system and solving the problem of inaccurate positioning of large lifting equipment such as gantry cranes.

[0104] For example, the dataset used in this invention comes from the positioning of a gantry crane actually operating at a construction site, and is real-time positioning data collected through an ultra-wideband (UWB) system. The original data collection frequency is 2 times / second, and base station 14 and base station 15 are two ranging base stations located on the same horizontal line, 642 meters apart. After simple logical verification, the data shown in the table below is obtained, representing the distance of the 6005 tag to base station 14 and base station 15 at each moment. See Table 1 for details.

[0105] Table 1

[0106] To fully demonstrate the effectiveness of the algorithm, this invention will use a portion of the data to focus on analyzing problems encountered in industrial data acquisition. For example... Figure 5 , Figure 6 As shown, this diagram compares data collected by base stations over different time periods. JZ15 represents directly collected data, while JZ14 is obtained by subtracting the data collected by base station 14 from the total distance (for consistency comparison). The horizontal axis converts time steps into data frames, with one second constituting one frame.

[0107] Ideally, the two line segments in the diagram should be identical and unambiguous. However, in-depth analysis of the original data revealed three key issues with the gantry crane's UWB positioning data: First, under non-line-of-sight (NLOS) conditions, significant discrepancies occurred between the ranging values ​​of the two base stations (when the base station's ranging value minus the actual distance to the base station exceeded a 70cm threshold), leading to a sharp decline in data reliability. Second, the original trajectory contained a large amount of high-frequency noise, with positional variations of tens of centimeters between adjacent frames, severely impacting the smoothness of the trajectory. Third, abnormal measurement values ​​frequently violated physical constraints (64200cm baseline distance). Based on the actual operation of the equipment on-site, the movement of objects lifted and moved by the equipment was reflected in the data fluctuations as either stable stillness or continuous movement in one direction. Overall, the data reflected by JZ15 was closer to the actual equipment movement, but there were still noticeable fluctuations and abrupt changes in the data that needed to be filtered out.

[0108] Furthermore, a comparison of filtering algorithm performance reveals that traditional filtering algorithms are significantly inadequate in handling real-world problems. The standard KF algorithm, using fixed noise parameters, cannot adapt to changes in motion states such as gantry crane start-up, shutdown, and speed changes, resulting in significant lag or overcorrection during abrupt state changes. While the EKF algorithm addresses nonlinear problems through linearization, it is sensitive to outliers and prone to divergence under NLOS conditions. The UKF algorithm improves nonlinear estimation accuracy through unscented transformation, but its high computational complexity and sensitivity to initial parameters make it difficult to operate stably in real-world industrial environments. The KBF algorithm, based on a continuous-time model, exhibits good smoothing characteristics, but lacks an effective outlier handling mechanism, leading to a significant performance degradation in the presence of NLOS errors. A common drawback of these traditional algorithms is the lack of a targeted reliability assessment mechanism, making them unable to effectively handle bistable issues in UWB positioning, resulting in insufficient adaptability in complex industrial environments.

[0109] To address these issues, the SE-AKF algorithm provided in this invention employs multiple innovative mechanisms: First, a velocity entropy-based observation source selection mechanism quantifies the reliability of each base station's measurements by calculating the information entropy of the velocity sequence within a sliding window, automatically selecting a more reliable observation source in NLOS scenarios; Second, a longitudinal consistency verification mechanism checks the historical trajectory consistency of candidate observations, effectively identifying and rejecting abnormal jumps; Third, an adaptive noise adjustment strategy dynamically adjusts process noise and observation noise parameters according to the motion state, enhancing the algorithm's adaptability to changes in motion state; Finally, the SE-AKF algorithm combined with inertial navigation addresses the brief inflection point lag that occurs when the motion state changes. Median filtering of the processing results at this time eliminates the inflection point lag problem caused by inertia. Figure 7 , Figure 8 This section compares the filtering effects of various algorithms.

[0110] From an overall trajectory perspective, the SE-AKF algorithm outputs the smoothest and most fluid trajectory, effectively identifying valid observations for predictive filtering, suppressing anomalous jumps, and maintaining the true motion trend. Other algorithms, in the face of divergent data, fail to clearly identify reasonable motion trajectories, tending to neutralize reliable and noisy points when discrepancies exist, leading to deviations from actual motion. A magnified view further clarifies the differences in processing capabilities at anomalous jumps. SE-AKF, through inertial motion recognition and median-based analysis, outputs the smoothest curve closest to the true trajectory, while EKF and KBF exhibit significant inflection point lag, and the KF algorithm suffers from overfitting, all deviating from the original trajectory. In conclusion, the SE-AKF algorithm, through innovative velocity entropy weighting and adaptive mechanisms, effectively solves the bistable problem and insufficient motion adaptability issues faced by traditional algorithms in industrial UWB positioning, demonstrating superior performance under various working conditions.

[0111] Optionally, in unsupervised industrial positioning scenarios, selecting appropriate evaluation metrics is crucial for objectively assessing the performance of filtering algorithms. This invention selects four metrics: residual energy, jitter, geometric closure error, and the number of frames violating temporal consistency rules, based on the following considerations: First, these metrics reflect algorithm performance from different dimensions, including smoothness, reliability, physical rationality, and temporal continuity; second, all metrics can be directly calculated from the filtering results, without relying on external truth references, making them suitable for practical industrial applications; third, the metrics have clear physical meaning and intuitive interpretability, facilitating performance evaluation and application selection in engineering practice.

[0112] Optionally, the dual-base station gantry crane positioning method based on event logging provided in this embodiment of the invention can also calculate evaluation indicators and perform performance evaluation, as shown in steps S201-S204.

[0113] S201. Based on the first ranging value and the second ranging value, as well as the current state estimate, determine the residual energy.

[0114] In some embodiments, the residual energy is used to reflect the correction magnitude of the filter.

[0115] For example, residual energy reflects the magnitude of the filter's correction to the original data. A larger value indicates a more aggressive correction of the observations by the filter, while a smaller value indicates a smaller difference from the original observations and a higher good fit. The calculation formula is as follows: ; in, These are the original observations. This is the estimated value after filtering. This represents the squared residual, reflecting the magnitude of the single-point filter correction.

[0116] S202. Based on the historical state estimate sequence in which the current state estimate is located, calculate the standard deviation of the position change at adjacent time points as a jitter index.

[0117] In some embodiments, the jitter metric reflects the smoothness of the trajectory.

[0118] For example, jitter, this metric directly measures the smoothness of the filtered trajectory; the smaller the value, the smoother the trajectory and the better the algorithm suppresses high-frequency noise. The calculation formula is as follows: ; in, Indicates the first The filtered estimate at time t. Indicates the first The filtered estimate at time t.

[0119] S203. Based on the current state estimate, invert the sum of the distances between the two base stations, calculate the deviation between the sum of the distances between the two base stations and the known baseline distance, and use it as the geometric closure error.

[0120] In some embodiments, the geometrical closure difference reflects the smoothness of the trajectory.

[0121] For example, geometric closure error. Based on the physical constraints of a dual-base station system (base station distance 64200cm), this index evaluates the algorithm's ability to maintain the physical rationality of the system; the smaller the value, the more the algorithm output conforms to the physical constraints. The calculation method is as follows: ; in, and The first The distance between base station 1 and base station 2 at time 64200 is the accurate distance between the two base stations, and N is the total number of data frames used to evaluate the total number of sampling points within the time period.

[0122] S204. Count the number of times the position change at adjacent time points exceeds the motion rationality threshold in the historical state estimate sequence where the current state estimate is located, and use this as the number of time consistency violation frames.

[0123] In some embodiments, the number of time consistency violation frames reflects the temporal continuity of the trajectory.

[0124] For example, the number of frames violating time consistency rules. This metric identifies unreasonable positional abrupt changes in the trajectory; the smaller the value, the more continuous the trajectory is in the time dimension, and the stronger the algorithm's ability to handle abnormal jumps. The calculation method is statistical. The number of frames under the 50cm condition is the number of data points under this condition within the total data range.

[0125] Although the observations collected by JZ15 also exhibit fluctuations and noise, they are generally more consistent with actual motion. Therefore, the experiments used JZ15 observations for comparative calculations. The performance evaluation results of each algorithm are shown in Figure 9. The SE-AKF algorithm performs best in several key metrics, particularly significantly outperforming other comparative algorithms in the dimensions of jitter, geometric closure error, and number of temporal consistency violations. The comparison results demonstrate that the SE-AKF algorithm exhibits significant advantages in all metrics. In terms of smoothness, SE-AKF's jitter index (3.11 cm) is reduced by 67.0% compared to the original data (9.43 cm), far exceeding other algorithms. This is due to its velocity entropy weighting mechanism, which can effectively identify and suppress abnormal jitter caused by NLOS errors. The KBF algorithm (7.93 cm) performs second best due to its damping characteristics, while the standard KF algorithm (14.47 cm) has the worst performance due to its fixed noise parameters, which cannot adapt to changes in motion state. In terms of physical consistency, SE-AKF effectively ensures the physical consistency of the output trajectory through adaptive noise adjustment and observation screening; its geometric closure error (0.59 cm) is significantly better than other algorithms, and it can maintain the physical constraints of the dual-base station system well. In terms of temporal continuity, SE-AKF benefits from its longitudinal consistency verification mechanism, which can promptly reject abnormal observations that differ too much from historical trajectories; the number of temporal consistency violation frames (42 frames) is reduced by 77.9% compared to the original data (190 frames), which can effectively identify and correct unreasonable positional abrupt changes.

[0126] It is worth noting that the residual energy metric of SE-AKF reflects the degree of modification to the original data. Given the noise in the original data, and considering that SE-AKF makes the fewest modifications compared to other algorithms while achieving superior performance in other metrics, this precisely demonstrates the effectiveness of SE-AKF in handling NLOS errors.

[0127] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0128] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0129] Figure 10 A schematic diagram of a dual-base station gantry crane positioning device based on event logging, provided in an embodiment of the present invention, is shown. The positioning device 300 includes a communication module 301 and a processing module 302.

[0130] The communication module 301 is used to receive the first ranging value of the first base station to the target device and the second ranging value of the second base station to the target device in the same straight direction as the gantry crane.

[0131] The processing module 302 is configured to calculate the current state estimate of the target device based on the first ranging value and the second ranging value using a state estimation algorithm. The state estimation algorithm includes at least one of the following: reliability decision based on velocity entropy, longitudinal consistency verification, and adaptive adjustment of noise parameters; perform event identification based on the current state estimate of the target device and a sequence of historical state estimates to determine the positioning state of the target device, wherein the positioning state includes a stationary state or a moving state; output the positioning result of the target device based on the positioning state and the current state estimate, and perform event-driven trajectory compression and storage on the positioning result.

[0132] Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The electronic device 400 includes: a processor 401, a memory 402, and a computer program 403 stored in the memory 402 and executable on the processor 401. When the processor 401 executes the computer program 403, it implements the steps in the above-described method embodiments. Alternatively, when the processor 401 executes the computer program 403, it implements the functions of each module / unit in the above-described device embodiments.

[0133] For example, the computer program 403 may be divided into one or more modules / units, which are stored in the memory 402 and executed by the processor 401 to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program 403 in the electronic device 400.

[0134] The processor 401 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0135] The memory 402 can be an internal storage unit of the electronic device 400, such as a hard disk or memory of the electronic device 400. The memory 402 can also be an external storage device of the electronic device 400, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD) card, flash card, etc., equipped on the electronic device 400. Furthermore, the memory 402 can include both internal and external storage units of the electronic device 400. The memory 402 is used to store the computer program and other programs and data required by the terminal. The memory 402 can also be used to temporarily store data that has been output or will be output.

[0136] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A dual-base station gantry crane positioning method based on event logging, characterized in that, include: Receive the first ranging value of the first base station to the target device and the second ranging value of the second base station to the target device in the same straight direction as the gantry crane; Based on the first and second ranging values, the current state estimate of the target device is calculated using a state estimation algorithm, wherein the state estimation algorithm includes at least one of the following: reliability decision based on velocity entropy, longitudinal consistency verification, and adaptive adjustment of noise parameters; Based on the current state estimate of the target device and the historical state estimate sequence, event identification is performed to determine the positioning state of the target device, which includes a stationary state or a moving state. Based on the positioning status and the current state estimate, the positioning result of the target device is output, and the positioning result is subjected to event-driven trajectory compression and storage.

2. The dual-base station gantry crane positioning method based on event logging according to claim 1, characterized in that, The step of calculating the current state estimate of the target device based on the first and second ranging values ​​using a state estimation algorithm includes: Based on the first and second ranging values, the velocity entropy is analyzed to make a reliability decision and determine candidate observation values. Based on the candidate observations and the state estimate from the previous time step, a longitudinal consistency check is performed to determine the valid observations. Based on the effective observations, the filtered observations are determined using a Kalman filter update method that adaptively adjusts the noise parameters. Based on the candidate observations, the valid observations, and / or the filtered observations, the current state estimate is determined.

3. The dual-base station gantry crane positioning method based on event logging according to claim 2, characterized in that, The step of analyzing velocity entropy based on the first and second ranging values, making reliability decisions, and determining candidate observation values ​​includes: Calculate the difference between the first ranging value and the second ranging value; If the difference is less than or equal to a preset error threshold, then the first ranging value or the second ranging value is selected as the candidate observation value. If the difference is greater than the preset error threshold, then the first velocity entropy is calculated based on the first ranging value and the historical sequence of the first ranging value; The second velocity entropy is calculated based on the historical sequence of the second ranging value and the second ranging value. If the first velocity entropy is greater than the second velocity entropy, then the first ranging value is selected as the candidate observation value. If the first velocity entropy is less than the second velocity entropy, then the second ranging value is selected as the candidate observation value.

4. The dual-base station gantry crane positioning method based on event logging according to claim 2, characterized in that, The step of performing a longitudinal consistency check based on the candidate observations and the state estimate from the previous time step to determine the valid observations includes: Calculate the absolute value of the difference between the candidate observation and the state estimate at the previous time step; If the absolute value is less than or equal to a preset motion threshold, then the candidate observation value is taken as a valid observation value. If the absolute value is greater than the preset motion threshold, the state prediction value is taken as the valid observation value. The state prediction value is a prediction value determined based on the historical state estimation value sequence.

5. The dual-base station gantry crane positioning method based on event logging according to claim 2, characterized in that, The step of determining the filtered observations based on the effective observations through an adaptive adjustment of noise parameters using a Kalman filter update step includes: The observation noise covariance is adjusted based on the entropy ratio of the first velocity entropy to the second velocity entropy. In response to the detection of a sudden acceleration change in the target device, the noise parameter corresponding to the acceleration state in the process noise covariance is increased; The Kalman filter is then applied using the adjusted noise parameters to update the filtered observations.

6. The dual-base station gantry crane positioning method based on event logging according to claim 1, characterized in that, The step of identifying events and determining the location status of the target device based on the current state estimate and the historical state estimate sequence of the target device includes: Using the current state estimate as the time endpoint, trace back a sequence of historical state estimates over a preset time period. Calculate the change in position in the historical state estimate sequence; If the change in position is less than the static determination threshold, and the duration of the preset time length is greater than or equal to the shortest duration, then the positioning state of the target device is determined to be static. Otherwise, the target device's location status is determined to be in a moving state.

7. The dual-base station gantry crane positioning method based on event logging according to claim 1, characterized in that, The step of outputting the positioning result of the target device based on the positioning state and the current state estimate includes: Based on the current state estimate, determine the current timestamp and the location coordinates of the target device; The positioning result is generated based on the current timestamp, the location coordinates of the target device, and the positioning status.

8. The dual-base station gantry crane positioning method based on event logging according to claim 1, characterized in that, The event-driven trajectory compression and storage of the positioning results includes: Based on the positioning results of the target device over a continuous period of time, a sequence of location coordinates is generated. Based on the position coordinate sequence, the Douglas-Peucker algorithm is used for compression to extract key feature points; Based on the positioning status and timestamp corresponding to the key feature points, identify and merge adjacent events that are both stationary or both in motion. The merged events are stored in the form of event metadata, which includes at least the event type, start time, end time, and location coordinates of key feature points. The event metadata is then used to replace the original continuous positioning results for persistent storage.

9. The dual-base station gantry crane positioning method based on event logging according to any one of claims 1 to 8, characterized in that, The method further includes: Based on the first and second ranging values, and the current state estimate, the residual energy is determined; the residual energy is used to reflect the correction magnitude of the filter. Based on the historical state estimate sequence in which the current state estimate is located, the standard deviation of the position change at adjacent time points is calculated as a jitter index, which reflects the smoothness of the trajectory. Based on the current state estimate, the sum of the distances between the two base stations is inverted, and the deviation between the sum of the distances between the two base stations and the known baseline distance is calculated as the geometric closure error, which reflects the smoothness of the trajectory. The number of times the position change at adjacent moments exceeds the motion rationality threshold in the historical state estimate sequence to which the current state estimate is located is counted as the time consistency violation frame number, which reflects the temporal continuity of the trajectory.

10. A dual-base station gantry crane positioning system based on event logging, characterized in that, The system includes an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor is configured to invoke and run the computer program stored in the memory to perform the method as described in any one of claims 1 to 9.

Images