Method and system for positioning cargo carrying table of stacking machine based on multi-sensor fusion

Abstract

The invention discloses a stacking machine cargo carrying table positioning method and system based on multi-sensor fusion. According to the method, original distance data are obtained through laser distance sensors arranged on a stacking machine stand column and a cargo carrying table, data fusion is carried out through a complementary relation, and acceleration information is collected through an inertial sensor; kalman filtering is carried out on the laser data before and after fusion, the data credibility is evaluated based on the variance of a filtering value and an original value, and an abnormal data segment of which the tail end is subjected to instantaneous interference is identified; the speed is calculated based on the fused laser data, the inertial acceleration is integrated to obtain the speed, and the more reliable fusion speed is obtained through dynamic weighted fusion according to the acceleration change trend; and when detecting that the laser data is abnormal, predicting and correcting positioning in an abnormal time period by using the fusion speed, and outputting a final positioning result. According to the invention, the advantages of high precision of laser and inertia shielding resistance are effectively fused, and the robustness and reliability of positioning in a complex industrial environment are improved.

Description

Technical Field

[0001] This invention relates to the field of automated warehousing and logistics technology, and in particular to a stacker crane loading platform positioning method and system based on multi-sensor fusion. Background Technology

[0002] In automated storage and retrieval systems (AS / RS), stacker cranes are the core equipment for achieving automated storage and retrieval of goods. To accurately locate goods, it is typically necessary to obtain the three-dimensional coordinates of the loading platform in real time. Currently, the mainstream solution relies on real-time communication with the stacker crane's own control system to obtain internal positioning data calculated by the stacker crane's encoders or servo systems. However, this solution has significant limitations: First, a stable, low-latency data link needs to be established between the moving loading platform and the stacker crane's main controller. In complex industrial wireless environments, signals are susceptible to interference, making reliability difficult to guarantee. Second, not all stacker crane manufacturers provide open internal positioning data interfaces or standard communication protocols. Third, if external systems request positioning data via high-frequency polling, it may impose additional load on the stacker crane's main controller that was not considered in its design, posing a risk to the stability and real-time performance of the stacker crane's core control.

[0003] To reduce reliance on data from within the stacker crane, autonomous positioning using independent sensors has become an alternative. Common technologies include laser distance sensors and inertial sensors. Laser distance sensors offer high accuracy and fast response, achieving absolute positioning by measuring the distance to a fixed reflection point. However, their measurements rely on optical paths and are susceptible to environmental dust, temporary obstructions, and uneven reflective surface materials, leading to irregular fluctuations or momentary inaccuracies in the data. Inertial sensors (such as MEMS gyroscopes and accelerometers) obtain displacement by integrating acceleration, independent of external reference objects. However, they suffer from high measurement noise, and integration errors accumulate rapidly over time. For cost-sensitive applications, positioning drift generated by low-precision inertial sensors within seconds can exceed acceptable limits, while high-precision fiber optic or laser gyroscopes are difficult to popularize due to their high cost.

[0004] Therefore, existing technologies struggle to achieve both high precision and robustness against environmental interference for independent positioning of stacker crane loading platforms in complex industrial scenarios such as warehousing and logistics, while controlling costs. Developing a technology that effectively integrates the advantages of different sensors and compensates for their respective shortcomings, thereby achieving stable and reliable positioning without stacker crane communication support, has become an urgent technical problem to be solved in this field. Summary of the Invention

[0005] The technical problem to be solved by this invention is: how to achieve high-precision and robust real-time positioning of the stacker crane loading platform in complex industrial environments without relying on the internal communication of the stacker crane, through low-cost sensor fusion technology.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a stacker crane loading platform positioning method based on multi-sensor fusion, comprising: At least one pair of laser distance sensors installed on the stacker crane's loading platform and columns are used to obtain the original distance measurement data of the loading platform relative to a preset fixed reflection point; at the same time, the motion acceleration data of the loading platform is obtained by an inertial sensor installed on the loading platform. Complementary verification and fusion processing are performed on two raw distance measurement data from the same pair of laser distance sensors to obtain a set of optimized distance data; The original distance measurement data and the intermediate data obtained by fusion processing are processed based on Kalman filtering, and the reliability of the data is evaluated by calculating the variance between the filtered data and the original data, and abnormal data segments caused by instantaneous interference are identified. First velocity information is calculated based on the preferred distance data, and second velocity information is obtained by integrating the acceleration data of the inertial sensor. The first velocity information and the second velocity information are then fused according to a preset weighting rule to obtain fused velocity information. When an abnormal data segment is detected, the fusion speed information is used to predict and correct the preferred distance data located within the abnormal data segment, so as to obtain the corrected real-time positioning data of the loading platform.

[0007] Furthermore, the acquisition of raw distance measurement data of the loading platform relative to a preset fixed reflection point through at least one pair of laser distance sensors installed on the loading platform and column of the stacker crane specifically includes: The raw distance measurement data in the horizontal direction is obtained by using a pair of horizontal laser distance sensors that are fixed at the bottom of the stacker crane columns and point to the front and rear ends along the direction of the stacker crane's travel. The raw distance measurement data in the vertical direction is obtained by using a pair of vertical laser distance sensors installed at the bottom of the stacker crane's loading platform, pointing vertically to the top and bottom fixed reflection points.

[0008] Furthermore, the process of performing complementary verification and fusion processing on two raw distance measurement data from the same pair of laser distance sensors to obtain a set of preferred distance data specifically includes: For each pair of laser distance sensors, a theoretical constant is determined based on the sum of the two original distance values ​​measured after installation and stabilization; For each sampling, the original measurement value of one sensor is recorded as the first original value, and the difference between the theoretical set value and the original measurement value of the other sensor is recorded as the second original value. Subsequent fusion and processing based on the first and second original values ​​yields a set of preferred distance data.

[0009] Furthermore, the process of processing the original distance measurement data and the intermediate data obtained through fusion based on Kalman filtering, and evaluating the reliability of the data by calculating the variance between the filtered data and the original data, specifically includes: Kalman filtering is performed on the first and second raw value sequences generated by the same pair of sensors to obtain the corresponding first and second filtered value sequences. The variance of the first filtered value sequence and its corresponding first original value sequence is calculated as the first deviation value, and the variance of the second filtered value sequence and its corresponding second original value sequence is calculated as the second deviation value. Based on the magnitude of the first deviation value and the second deviation value, the preferred distance value for final acceptance is determined, wherein the smaller the deviation value, the higher the weight of the corresponding data.

[0010] Furthermore, determining the preferred distance value for final acceptance based on the magnitudes of the first and second deviation values ​​specifically includes: When the first deviation value is zero, the filtered value corresponding to the first original value is adopted as the preferred distance value; Otherwise, the reciprocal absolute values ​​of the first deviation value and the second deviation value are used as the weights of the filter values ​​corresponding to the first original value and the second original value, respectively, and the weighted average value is calculated as the preferred distance value.

[0011] Furthermore, the identification of abnormal data segments caused by transient interference includes: Set a time window length and a variance threshold; Examine the data segment at the end of the filtered value sequence, whose length is equal to the length of the time window, and calculate the variance between the filtered value and the original value within this data segment; If the variance is greater than the variance threshold, the data segment is determined to be an unreliable data anomaly segment caused by transient interference.

[0012] Furthermore, the step of calculating the first velocity information based on the preferred distance data, integrating the acceleration data from the inertial sensor to obtain the second velocity information, and fusing the first velocity information and the second velocity information according to a preset weighting rule to obtain the fused velocity information specifically includes: The first instantaneous velocity is calculated based on the difference between adjacent sample values ​​of the preferred distance data and the sampling time interval; The acceleration data collected by the inertial sensor is integrated to obtain the second instantaneous velocity, and the predicted instantaneous velocity in the near future is predicted based on the real-time fitting model of the acceleration data. Dynamic weights are set according to the acceleration change trend, and the first instantaneous velocity and the second instantaneous velocity or the predicted instantaneous velocity are weighted and averaged to obtain the fused velocity information. When the acceleration change trend increases, the weight of the first instantaneous velocity is increased.

[0013] Furthermore, when an abnormal data segment is detected, the step of using the fusion speed information to predict and correct the preferred distance data located within the abnormal data segment specifically includes: The reliable preferred distance data preceding the starting point of the data anomaly segment is used as the reference value; The predicted displacement within the time period is calculated by integrating the fusion speed information with the duration corresponding to the abnormal data segment. The reference value is added to the predicted displacement to generate corrected positioning data for each time point within the data anomaly segment.

[0014] The present invention also provides a stacker crane loading platform positioning system based on multi-sensor fusion, for implementing the positioning method described above, the system comprising: The laser ranging module includes at least one pair of laser distance sensors for acquiring raw distance measurement data of the loading platform; An inertial measurement module, including electronic inertial sensors, is used to acquire motion acceleration data of the loading platform; The data processing and control module is configured to perform the complementary verification and fusion processing, credibility assessment and anomaly detection, speed information fusion and positioning data correction, so as to output the corrected real-time positioning data.

[0015] Furthermore, the laser ranging module includes a horizontal ranging unit and a vertical ranging unit; The horizontal ranging unit includes a pair of horizontal laser distance sensors installed at the bottom of the stacker crane column and pointing towards fixed reflection points along the direction of travel; The vertical ranging unit includes a pair of vertical laser distance sensors installed at the bottom of the stacker crane's loading platform, pointing vertically to fixed reflection points at the top and bottom.

[0016] The beneficial effects of this invention are as follows: Through a closed-loop technology chain of multi-source data acquisition, complementary verification fusion, reliability assessment, and collaborative correction of velocity information, it effectively solves the inherent defects of a single sensor in complex industrial environments. Utilizing the complementary measurement relationship of paired laser sensors, verification fusion significantly improves the anti-interference capability and fault tolerance of the original distance data. The introduction of Kalman filtering and variance analysis mechanisms enables dynamic quantitative assessment of data reliability and accurate identification of instantaneous interference segments. Weighted fusion of the velocity calculated from the laser data and the integral velocity of the inertial sensor, with dynamic weight adjustments based on the motion state, forms a highly reliable velocity estimate with complementary advantages. Finally, this fused velocity is used to predict and correct the identified abnormal laser positioning data, thereby achieving high-precision, highly robust, and real-time self-consistent positioning results without relying on stacker crane communication, significantly improving the system's stability and practicality under complex working conditions such as dust obstruction and signal fluctuations. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, 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 the structures shown in these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the stacker crane loading platform positioning method based on multi-sensor fusion according to an embodiment of the present invention. Figure 2 This is a flowchart illustrating the stacker crane loading platform positioning method based on multi-sensor fusion, according to an embodiment of the present invention. Figure 3 This is a schematic diagram of the sensor positions on the horizontal stacker crane loading platform according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the sensor positions on the vertical stacker crane loading platform according to an embodiment of the present invention; Figure 5 This is a structural diagram of a stacker crane loading platform positioning system based on multi-sensor fusion, according to an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0020] It should be noted that the descriptions involving "first," "second," etc., in this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature. Furthermore, the technical solutions of the various embodiments can be combined with each other, but only on the basis of being achievable by those skilled in the art. When the combination of technical solutions is contradictory or impossible to implement, such a combination of technical solutions should be considered non-existent and not within the scope of protection claimed by this invention.

[0021] like Figure 1 As shown, an embodiment of the present invention is: a stacker crane loading platform positioning method based on multi-sensor fusion. The overall steps of the solution include steps S1-S5, and the specific complete execution steps are as follows: Figure 2 As shown.

[0022] S1. Obtain the original distance measurement data of the loading platform relative to the preset fixed reflection point by using at least a pair of laser distance sensors installed on the loading platform and the column of the stacker crane; at the same time, obtain the motion acceleration data of the loading platform by using the inertial sensor installed on the loading platform.

[0023] In a specific embodiment, the step of acquiring the original distance measurement data of the loading platform relative to a preset fixed reflection point through at least a pair of laser distance sensors installed on the loading platform and the column of the stacker crane specifically includes: The raw distance measurement data in the horizontal direction is obtained by using a pair of horizontal laser distance sensors that are fixed at the bottom of the stacker crane columns and point to the front and rear ends along the direction of the stacker crane's travel. The raw distance measurement data in the vertical direction is obtained by using a pair of vertical laser distance sensors installed at the bottom of the stacker crane's loading platform, pointing vertically to the top and bottom fixed reflection points.

[0024] In this embodiment, as Figure 3 , Figure 4 As shown, two laser distance sensors (horizontal distance sensors) are installed in pairs at the bottom of the stacker crane's columns, along its horizontal travel direction, pointing to fixed reflection points at the front and rear ends of the track, respectively. At the bottom of the loading platform, two laser distance sensors (vertical distance sensors) are installed in pairs, along the vertical lifting direction, pointing to fixed reflection points at the top overhead track and the bottom foundation, respectively. Theoretically, the sum of the real-time measurements of each pair of sensors should be a fixed installation distance constant. A three-axis electronic inertial sensor (IMU) is installed on the loading platform, with its x-axis parallel to the stacker crane's horizontal travel direction and its z-axis perpendicular to the ground. When collecting data, the local gravitational acceleration is subtracted from the raw z-axis reading to zero, thus directly corresponding to the horizontal and vertical accelerations, respectively, on the x-axis and z-axis.

[0025] S2. Perform complementary verification and fusion processing on two raw distance measurement data from the same pair of laser distance sensors to obtain a set of optimized distance data.

[0026] In a specific embodiment, the step of performing complementary verification and fusion processing on two raw distance measurement data from the same pair of laser distance sensors to obtain a set of preferred distance data specifically includes: S21. For each pair of laser distance sensors, a theoretical constant value is determined based on the sum of the two original distance values ​​measured after installation and stabilization. S22. For each sampling, the original measurement value of one sensor is recorded as the first original value, and the difference between the theoretical value and the original measurement value of the other sensor is recorded as the second original value. S23. Based on the first and second original values, subsequent fusion and processing are performed to obtain a set of preferred distance data.

[0027] In this embodiment, for each pair of laser sensors, taking the horizontal direction as an example: Let the reading of the front-end sensor be x, the reading of the back-end sensor be y, and the theoretical value measured after installation be S (S = x + y).

[0028] In each sampling, x is the first original value. Based on the complementarity relationship, the second original value x' = S - y can be calculated from y.

[0029] Ideally, x and x' should be equal, but due to interference, they differ. Subsequent processing is based on the x and x' sequences.

[0030] S3. The original distance measurement data and the intermediate data obtained by fusion processing are processed based on Kalman filtering, and the reliability of the data is evaluated by calculating the variance between the filtered data and the original data, and abnormal data segments caused by instantaneous interference are identified.

[0031] In a specific embodiment, the process of processing the original distance measurement data and the intermediate data obtained by fusion based on Kalman filtering, and evaluating the reliability of the data by calculating the variance between the filtered data and the original data, specifically includes: S31. Perform Kalman filtering on the first and second raw value sequences generated by the same pair of sensors to obtain the corresponding first and second filtered value sequences. S32. Calculate the variance of the first filtered value sequence and its corresponding first original value sequence as the first deviation value, and calculate the variance of the second filtered value sequence and its corresponding second original value sequence as the second deviation value. S33. Based on the magnitude of the first deviation value and the second deviation value, determine the preferred distance value for final acceptance, wherein the smaller the deviation value, the higher the weight of the corresponding data.

[0032] In a specific embodiment, determining the preferred distance value for final acceptance based on the magnitudes of the first deviation value and the second deviation value specifically includes: When the first deviation value is zero, the filtered value corresponding to the first original value is adopted as the preferred distance value; Otherwise, the reciprocal absolute values ​​of the first deviation value and the second deviation value are used as the weights of the filter values ​​corresponding to the first original value and the second original value, respectively, and the weighted average value is calculated as the preferred distance value.

[0033] In this embodiment, a fixed-length sliding window is used to perform Kalman filtering on the x sequence and the x' sequence respectively to obtain the filtered sequences x_f and x'_f.

[0034] The variance of the x_f sequence and the original x sequence is calculated and used as the first deviation value D_x, which reflects the degree of continuous interference to the x data.

[0035] Similarly, the variance between the x'_f sequence and the original x' sequence is calculated and used as the second deviation value D_x'.

[0036] The smaller the deviation value, the less interference the data is subjected to, and the more reliable it is.

[0037] If D_x is zero, the terminal value of x_f is directly adopted as the current preferred distance L_opt. Otherwise, L_opt is obtained by taking a weighted average of the terminal values ​​of x_f and x'_f using 1 / |D_x| and 1 / |D_x'| as weights.

[0038] In a specific embodiment, identifying abnormal data segments caused by transient interference includes: S34. Set a time window length and variance threshold; S35. Check the data segment at the end of the filtered value sequence with a length equal to the length of the time window, and calculate the variance between the filtered value and the original value in the data segment. S36. If the variance is greater than the variance threshold, the data segment is determined to be an unreliable data anomaly segment caused by transient interference.

[0039] In this embodiment, since transient interference (such as drifting particles) may affect recent measurements, it is necessary to identify unreliable end-point data. The specific identification process is as follows: Set parameters: End-point inspection data length d, variance threshold dg.

[0040] Examine the last d data points of the x_f sequence and calculate the variance Var_end between these d x_f values ​​and the corresponding original x values.

[0041] If Var_end > dg, then the last d data points are considered an anomalous segment, and the L_opt values ​​within the most recent d / p seconds (p is the sampling rate) cannot be directly trusted. In this case, the time point of the last trustworthy L_opt value lags behind the current time by d / p seconds.

[0042] S4. Calculate the first velocity information based on the preferred distance data, and perform integral processing on the acceleration data of the inertial sensor to obtain the second velocity information. Then, fuse the first velocity information and the second velocity information according to the preset weighting rules to obtain the fused velocity information.

[0043] In a specific embodiment, the process of calculating first velocity information based on the preferred distance data, integrating the acceleration data from the inertial sensor to obtain second velocity information, and fusing the first velocity information and second velocity information according to a preset weighting rule to obtain fused velocity information specifically includes: S41. Calculate the first instantaneous velocity based on the difference between adjacent sampled values ​​of the preferred distance data and the sampling time interval; S42. Integrate the acceleration data collected by the inertial sensor to obtain the second instantaneous velocity, and predict the instantaneous velocity in the near future based on the real-time fitting model of the acceleration data. S43. Set dynamic weights according to the acceleration change trend, and perform a weighted average of the first instantaneous velocity and the second instantaneous velocity or the predicted instantaneous velocity to obtain the fused velocity information, wherein when the acceleration change trend increases, the weight of the first instantaneous velocity is increased.

[0044] In this embodiment, inertial sensor data is used to compensate for missing or unreliable laser data during abnormal segments. Specifically, this includes: Laser velocity calculation: Based on the continuous L_opt values, the displacement difference between adjacent sampling points is calculated and divided by time to obtain the first instantaneous velocity V_laser based on the laser.

[0045] Inertial velocity calculation and prediction: The acceleration data (corresponding direction) acquired by the IMU is integrated to obtain the second instantaneous velocity V_imu. At the same time, real-time curve fitting (such as polynomial fitting) is performed on the acceleration sequence to predict the acceleration trend in the short term, and then the predicted instantaneous velocity V_pred is obtained by integration.

[0046] Velocity fusion: A weighted average of V_laser and V_imu (or V_pred) is taken to obtain the fused velocity V_fused. Dynamic weight adjustment: When the IMU detects a gradual change in acceleration (small jerk), the weight of V_imu can be appropriately increased; when the acceleration changes drastically, the weight of V_laser should be increased, because inertial sensors are prone to larger errors during abrupt changes. Initial weights are set according to the sensor's nominal accuracy, with laser typically having a higher weight.

[0047] S5. When an abnormal data segment is detected, the fusion speed information is used to predict and correct the preferred distance data located within the abnormal data segment to obtain the corrected real-time positioning data of the loading platform.

[0048] In a specific embodiment, when an abnormal data segment is detected, the step of using the fusion speed information to predict and correct the preferred distance data located within the abnormal data segment specifically includes: S51. Use the reliable preferred distance data preceding the starting point of the data anomaly segment as the reference value; S52. Integrate the duration corresponding to the abnormal data segment using the fusion speed information to calculate the predicted displacement within that time period. S53. Add the reference value to the predicted displacement to generate corrected positioning data for each time point within the data anomaly segment.

[0049] In this embodiment, when an abnormal segment of length d is detected at the end, the last reliable L_opt value before the start of the abnormal segment is taken as the correction base L_base; the fused velocity information V_fused within the corresponding time period [t_base, t_current] is taken, and the fused velocity information V_fused may be a sequence or function of multiple velocity values; V_fused is integrated within this time period to calculate the predicted displacement ΔL_pred; the final corrected real-time positioning value L_corrected = L_base + ΔL_pred.

[0050] By performing the above steps and calculating the horizontal and vertical directions respectively, the corrected real-time two-dimensional coordinates of the loading platform can be obtained. This invention, through ingenious sensor deployment and a hierarchical data fusion algorithm, effectively suppresses random interference from laser sensors and cumulative errors from inertial sensors, achieving high-precision and high-reliability autonomous positioning without the need for stacker crane communication support.

[0051] like Figure 5As shown, this embodiment of the invention also provides a stacker crane loading platform positioning system based on multi-sensor fusion, used to implement the positioning method described above. The system includes: a laser ranging module 10, an inertial measurement module 20, and a data processing and control module 30.

[0052] The laser ranging module 10 includes at least one pair of laser distance sensors for acquiring raw distance measurement data of the loading platform.

[0053] In a specific embodiment, the laser ranging module 10 includes a horizontal ranging unit and a vertical ranging unit; The horizontal ranging unit includes a pair of horizontal laser distance sensors installed at the bottom of the stacker crane column and pointing towards fixed reflection points along the direction of travel; The vertical ranging unit includes a pair of vertical laser distance sensors installed at the bottom of the stacker crane's loading platform, pointing vertically to fixed reflection points at the top and bottom.

[0054] The inertial measurement module 20 includes an electronic inertial sensor for acquiring motion acceleration data of the loading platform; The data processing and control module 30 is configured to perform the complementary verification and fusion processing, credibility assessment and anomaly detection, speed information fusion and positioning data correction, so as to output the corrected real-time positioning data.

[0055] In a specific embodiment, the data processing and control module 30 is configured to perform complementary verification and fusion processing, reliability assessment and anomaly detection, speed information fusion, and positioning data correction to output corrected real-time positioning data, as follows: In a specific embodiment, the data processing and control module 30 is configured to perform complementary verification and fusion processing, specifically including: For each pair of laser distance sensors, a theoretical constant is determined based on the sum of the two original distance values ​​measured after installation and stabilization; For each sampling, the original measurement value of one sensor is recorded as the first original value, and the difference between the theoretical set value and the original measurement value of the other sensor is recorded as the second original value. Subsequent fusion and processing based on the first and second original values ​​yields a set of preferred distance data.

[0056] In a specific embodiment, the data processing and control module 30 is configured to perform a credibility assessment, specifically including: Kalman filtering is performed on the first and second raw value sequences generated by the same pair of sensors to obtain the corresponding first and second filtered value sequences. The variance of the first filtered value sequence and its corresponding first original value sequence is calculated as the first deviation value, and the variance of the second filtered value sequence and its corresponding second original value sequence is calculated as the second deviation value. Based on the magnitude of the first deviation value and the second deviation value, the preferred distance value for final acceptance is determined, wherein the smaller the deviation value, the higher the weight of the corresponding data.

[0057] Specifically, determining the preferred distance value for final acceptance based on the magnitudes of the first and second deviation values ​​includes: When the first deviation value is zero, the filtered value corresponding to the first original value is adopted as the preferred distance value; Otherwise, the reciprocal absolute values ​​of the first deviation value and the second deviation value are used as the weights of the filter values ​​corresponding to the first original value and the second original value, respectively, and the weighted average value is calculated as the preferred distance value.

[0058] In a specific embodiment, the data processing and control module 30 is configured to perform anomaly detection, including: Set a time window length and a variance threshold; Examine the data segment at the end of the filtered value sequence, whose length is equal to the length of the time window, and calculate the variance between the filtered value and the original value within this data segment; If the variance is greater than the variance threshold, the data segment is determined to be an unreliable data anomaly segment caused by transient interference.

[0059] In a specific embodiment, the data processing and control module 30 is configured to perform speed information fusion, specifically including: The first instantaneous velocity is calculated based on the difference between adjacent sample values ​​of the preferred distance data and the sampling time interval; The acceleration data collected by the inertial sensor is integrated to obtain the second instantaneous velocity, and the predicted instantaneous velocity in the near future is predicted based on the real-time fitting model of the acceleration data. Dynamic weights are set according to the acceleration change trend, and the first instantaneous velocity and the second instantaneous velocity or the predicted instantaneous velocity are weighted and averaged to obtain the fused velocity information. When the acceleration change trend increases, the weight of the first instantaneous velocity is increased.

[0060] In a specific embodiment, the data processing and control module 30 is configured to perform positioning data correction, specifically including: The reliable preferred distance data preceding the starting point of the data anomaly segment is used as the reference value; The predicted displacement within the time period is calculated by integrating the fusion speed information with the duration corresponding to the abnormal data segment. The reference value is added to the predicted displacement to generate corrected positioning data for each time point within the data anomaly segment.

[0061] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A stacker crane loading platform positioning method based on multi-sensor fusion, characterized in that, include: At least one pair of laser distance sensors installed on the stacker crane's loading platform and columns are used to obtain the original distance measurement data of the loading platform relative to a preset fixed reflection point; at the same time, the motion acceleration data of the loading platform is obtained by an inertial sensor installed on the loading platform. Complementary verification and fusion processing are performed on two raw distance measurement data from the same pair of laser distance sensors to obtain a set of optimized distance data; The original distance measurement data and the intermediate data obtained by fusion processing are processed based on Kalman filtering, and the reliability of the data is evaluated by calculating the variance between the filtered data and the original data, and abnormal data segments caused by instantaneous interference are identified. First velocity information is calculated based on the preferred distance data, and second velocity information is obtained by integrating the acceleration data of the inertial sensor. The first velocity information and the second velocity information are then fused according to a preset weighting rule to obtain fused velocity information. When an abnormal data segment is detected, the fusion speed information is used to predict and correct the preferred distance data located within the abnormal data segment, so as to obtain the corrected real-time positioning data of the loading platform.

2. The stacker crane loading platform positioning method based on multi-sensor fusion according to claim 1, characterized in that, The process of acquiring the original distance measurement data of the loading platform relative to a preset fixed reflection point through at least one pair of laser distance sensors installed on the loading platform and column of the stacker crane specifically includes: The raw distance measurement data in the horizontal direction is obtained by using a pair of horizontal laser distance sensors that are fixed at the bottom of the stacker crane columns and point to the front and rear ends along the direction of the stacker crane's travel. The raw distance measurement data in the vertical direction is obtained by using a pair of vertical laser distance sensors installed at the bottom of the stacker crane's loading platform, pointing vertically to the top and bottom fixed reflection points.

3. The stacker crane loading platform positioning method based on multi-sensor fusion according to claim 2, characterized in that, The process of performing complementary verification and fusion on two raw distance measurement data from the same pair of laser distance sensors to obtain a set of preferred distance data specifically includes: For each pair of laser distance sensors, a theoretical constant is determined based on the sum of the two original distance values ​​measured after installation and stabilization; For each sampling, the original measurement value of one sensor is recorded as the first original value, and the difference between the theoretical set value and the original measurement value of the other sensor is recorded as the second original value. Subsequent fusion and processing based on the first and second original values ​​yields a set of preferred distance data.

4. The stacker crane loading platform positioning method based on multi-sensor fusion according to claim 3, characterized in that, The process of processing the original distance measurement data and the intermediate data obtained through fusion based on Kalman filtering, and evaluating the reliability of the data by calculating the variance between the filtered data and the original data, specifically includes: Kalman filtering is performed on the first and second raw value sequences generated by the same pair of sensors to obtain the corresponding first and second filtered value sequences. The variance of the first filtered value sequence and its corresponding first original value sequence is calculated as the first deviation value, and the variance of the second filtered value sequence and its corresponding second original value sequence is calculated as the second deviation value. Based on the magnitude of the first deviation value and the second deviation value, the preferred distance value for final acceptance is determined, wherein the smaller the deviation value, the higher the weight of the corresponding data.

5. The stacker crane loading platform positioning method based on multi-sensor fusion according to claim 4, characterized in that, The process of determining the preferred distance value for final acceptance based on the magnitudes of the first and second deviation values ​​specifically includes: When the first deviation value is zero, the filtered value corresponding to the first original value is adopted as the preferred distance value; Otherwise, the reciprocal absolute values ​​of the first deviation value and the second deviation value are used as the weights of the filter values ​​corresponding to the first original value and the second original value, respectively, and the weighted average value is calculated as the preferred distance value.

6. The stacker crane loading platform positioning method based on multi-sensor fusion according to claim 4, characterized in that, The identification of abnormal data segments caused by transient interference includes: Set a time window length and a variance threshold; Examine the data segment at the end of the filtered value sequence, whose length is equal to the length of the time window, and calculate the variance between the filtered value and the original value within this data segment; If the variance is greater than the variance threshold, the data segment is determined to be an unreliable data anomaly segment caused by transient interference.

7. The stacker crane loading platform positioning method based on multi-sensor fusion according to claim 6, characterized in that, The process of calculating first velocity information based on the preferred distance data, integrating the acceleration data from the inertial sensor to obtain second velocity information, and fusing the first and second velocity information according to a preset weighting rule to obtain fused velocity information specifically includes: The first instantaneous velocity is calculated based on the difference between adjacent sample values ​​of the preferred distance data and the sampling time interval; The acceleration data collected by the inertial sensor is integrated to obtain the second instantaneous velocity, and the predicted instantaneous velocity in the near future is predicted based on the real-time fitting model of the acceleration data. Dynamic weights are set according to the acceleration change trend, and the first instantaneous velocity and the second instantaneous velocity or the predicted instantaneous velocity are weighted and averaged to obtain the fused velocity information. When the acceleration change trend increases, the weight of the first instantaneous velocity is increased.

8. The stacker crane loading platform positioning method based on multi-sensor fusion according to claim 7, characterized in that, The step of using the fusion speed information to predict and correct the preferred distance data located within the abnormal data segment when an abnormal data segment is detected specifically includes: The reliable preferred distance data preceding the starting point of the data anomaly segment is used as the reference value; The predicted displacement within the time period is calculated by integrating the fusion speed information with the duration corresponding to the abnormal data segment. The reference value is added to the predicted displacement to generate corrected positioning data for each time point within the data anomaly segment.

9. A stacker crane loading platform positioning system based on multi-sensor fusion, characterized in that, The system for implementing the positioning method according to any one of claims 1-8 includes: The laser ranging module includes at least one pair of laser distance sensors for acquiring raw distance measurement data of the loading platform; An inertial measurement module, including electronic inertial sensors, is used to acquire motion acceleration data of the loading platform; The data processing and control module is configured to perform the complementary verification and fusion processing, credibility assessment and anomaly detection, speed information fusion and positioning data correction, so as to output the corrected real-time positioning data.

10. The stacker crane loading platform positioning system based on multi-sensor fusion according to claim 9, characterized in that, The laser ranging module includes a horizontal ranging unit and a vertical ranging unit; The horizontal ranging unit includes a pair of horizontal laser distance sensors installed at the bottom of the stacker crane column and pointing towards fixed reflection points along the direction of travel; The vertical ranging unit includes a pair of vertical laser distance sensors installed at the bottom of the stacker crane's loading platform, pointing vertically to fixed reflection points at the top and bottom.

Images