Automated data fitting method and system for equipment position calibration

Abstract

This invention provides an automated data fitting method and system that combines equipment position calibration. First, it collects a multi-source sensor data set of the target equipment in three-dimensional space, including equipment motion trajectory data, position coordinate offset, and environmental interference parameters. Next, it extracts features from the multi-source sensor data set to generate a target feature set containing spatial position correlation features, dynamic trajectory fluctuation features, and environmental coupling features. Then, it calls a pre-trained adaptive fitting model, inputs the target feature set for dynamic parameter matching, and generates an initial calibration parameter set. Based on the initial calibration parameter set, it generates a dynamic path sequence for equipment position calibration, triggering the equipment to perform automated position calibration according to the sequence. The calibrated verification data is then fed back to the model to update the parameters, thereby improving the accuracy, real-time performance, and adaptability of equipment position calibration.

Description

Technical Field

[0001] This invention relates to the field of sensor technology, and more specifically, to an automated data fitting method and system that incorporates device position calibration. Background Technology

[0002] In numerous fields such as industrial automation, robot navigation, and smart warehousing, there are extremely high requirements for the precise positioning of target equipment in three-dimensional space. Currently, traditional equipment positioning calibration methods mainly rely on manual measurement and fixed parameter settings. For example, in some simple industrial scenarios, workers use measuring tools to manually record the initial position information of the equipment and set fixed calibration parameters based on this information.

[0003] However, these traditional methods have many limitations. On the one hand, manual measurement is easily affected by human factors, such as measurement errors and improper operation, leading to low calibration accuracy. On the other hand, in actual operation, equipment is affected by various factors, such as changes in ambient temperature, mechanical vibration, and uneven ground. These factors can cause the equipment's position to shift, and fixed calibration parameters cannot adapt to these changes in real time, thus affecting the normal operation and operational accuracy of the equipment. Furthermore, as the complexity of the equipment's working environment and tasks continues to increase, traditional calibration methods are unable to meet the requirements for high accuracy, real-time performance, and adaptability in equipment position calibration. There is an urgent need for a method that can combine equipment position calibration with automated data fitting capabilities to solve these problems. Summary of the Invention

[0004] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide an automated data fitting method incorporating device location calibration, the method comprising:

[0005] Collect a set of multi-source sensor data of the target device in three-dimensional space. The set of multi-source sensor data includes device motion trajectory data, position coordinate offset and environmental interference parameters.

[0006] Feature extraction is performed on the multi-source sensor data set to generate a target feature set, which includes spatial location correlation features, dynamic trajectory fluctuation features, and environmental coupling features.

[0007] The pre-trained adaptive fitting model is invoked, and the target feature set is input into the adaptive fitting model for dynamic parameter matching processing to generate the initial calibration parameter set of the target device.

[0008] A dynamic path sequence for device location calibration is generated based on the initial calibration parameter set, and the dynamic path sequence includes spatial coordinate adjustment parameters for multiple consecutive calibration points;

[0009] The target device is triggered to perform an automated position calibration operation according to the dynamic path sequence, and the calibrated verification data set is collected and fed back to the adaptive fitting model to update the model parameters.

[0010] In another aspect, embodiments of the present invention also provide an automated data fitting system that combines device location calibration, including a processor and a machine-readable storage medium connected to the processor. The machine-readable storage medium is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the machine-readable storage medium to implement the above-described method.

[0011] Based on the above, this embodiment of the invention collects a multi-source sensor data set of the target device in three-dimensional space, covering information such as device motion trajectory data, position coordinate offset, and environmental interference parameters. Feature extraction is performed on the multi-source sensor data set to generate a target feature set, which includes spatial position correlation features, dynamic trajectory fluctuation features, and environmental coupling features. This accurately characterizes the device's state in three-dimensional space and its interaction with the environment. A pre-trained adaptive fitting model is invoked to perform dynamic parameter matching processing on the target feature set, generating an initial calibration parameter set. Utilizing the adaptive capability of the adaptive fitting model, parameters can be dynamically adjusted according to different feature information, improving the accuracy and adaptability of the calibration parameters. Based on the initial calibration parameter set, a dynamic path sequence for device position calibration is generated, enabling the device to perform automated position calibration operations according to this sequence. This achieves real-time dynamic adjustment of the device position, further improving the accuracy and real-time performance of position calibration. The collected and calibrated verification data set is fed back to the adaptive fitting model to update the model parameters, allowing the adaptive fitting model to continuously learn and improve, better adapting to the position calibration needs of the device in different environments and operating states, thereby significantly improving the accuracy, real-time performance, and adaptability of device position calibration. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the execution flow of the automated data fitting method combined with device location calibration provided in an embodiment of the present invention.

[0013] Figure 2 This is a schematic diagram of exemplary hardware and software components of an automated data fitting system that combines device location calibration, provided in an embodiment of the present invention. Detailed Implementation

[0014] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1This is a flowchart illustrating an automated data fitting method combining device location calibration according to an embodiment of the present invention. The following is a detailed description of this automated data fitting method combining device location calibration.

[0015] Step S110: Collect a set of multi-source sensor data of the target device in three-dimensional space. The set of multi-source sensor data includes device motion trajectory data, position coordinate offset and environmental interference parameters.

[0016] Taking industrial automation production scenarios as an example, specifically, taking an AGV (Automated Guided Vehicle) that performs cargo handling tasks in an intelligent warehousing system as an example, in order to achieve accurate position calibration and data fitting in the future, it is necessary to collect its multi-source sensor data set.

[0017] Step S111: The displacement increment data of the target device in the X-axis, Y-axis and Z-axis are collected in real time by a laser ranging sensor array deployed on the motion plane of the target device.

[0018] A laser ranging sensor array is deployed on the AGV's movement plane. The sensors in the array are distributed according to a set layout to ensure comprehensive and accurate acquisition of the AGV's displacement information in three-dimensional space. Each laser ranging sensor emits a laser beam, which is reflected back when it encounters the AGV. By measuring the emission and reception time difference of the laser and combining this with the laser's propagation speed in air, the sensor can calculate the distance between itself and the AGV. As the AGV moves, the sensors continuously repeat this process, thereby monitoring the AGV's displacement changes in the X, Y, and Z axes in real time.

[0019] To ensure data accuracy and reliability, the laser rangefinder sensor needs to be calibrated and adjusted. During sensor installation, the accuracy and stability of its mounting position must be guaranteed to avoid measurement errors caused by improper installation. Simultaneously, the sensor parameters need to be set and adjusted to adapt to different working environments and measurement requirements. For example, the sensor's measurement accuracy may be affected under different lighting conditions; therefore, parameters such as sensitivity and threshold need to be adjusted according to the actual situation.

[0020] During data acquisition, the sensors collect data at a set sampling frequency. The selection of the sampling frequency needs to comprehensively consider the AGV's movement speed and data processing requirements. If the sampling frequency is too low, data loss or incompleteness may occur, failing to accurately reflect the AGV's movement trajectory; if the sampling frequency is too high, it can generate a large amount of data, increasing the burden on data processing. Therefore, it is necessary to rationally select the sampling frequency based on the actual movement of the AGV to ensure that the collected data accurately reflects the AGV's movement status without placing excessive pressure on subsequent data processing.

[0021] The collected displacement increment data is a time-varying sequence, recording the displacement changes of the AGV in the X, Y, and Z axes at different times. This data will be stored in a data storage device for subsequent analysis and processing.

[0022] Step S112: Call the inertial measurement unit to obtain the angular velocity offset and acceleration offset of the target device, and record the temperature interference parameters and electromagnetic interference intensity at each sampling time.

[0023] An inertial measurement unit (IMU) is installed on the AGV, which is a sensor capable of measuring the angular velocity and acceleration of an object. The IMU can acquire the AGV's angular velocity offset and acceleration offset. The angular velocity offset reflects the speed change of the AGV during rotation, while the acceleration offset reflects the speed change of the AGV during linear motion.

[0024] The IMU (Inertial Measurement Unit) works based on the measurement of inertial forces. It contains sensitive components such as accelerometers and gyroscopes. The accelerometers measure the AGV's acceleration in various directions, while the gyroscopes measure its angular velocity. During the AGV's movement, the IMU continuously collects this data and converts it into digital signals for output.

[0025] While acquiring angular velocity and acceleration offsets, it is also necessary to record temperature interference parameters and electromagnetic interference intensity at each sampling moment. In intelligent warehousing systems, temperature variations can affect the performance of the IMU, leading to measurement errors. For example, an increase in temperature may cause thermal expansion of the electronic components inside the IMU, thus affecting its measurement accuracy. Therefore, it is necessary to use a temperature sensor to monitor and record the ambient temperature in real time.

[0026] Electromagnetic interference (EMI) is also a significant factor affecting the measurement accuracy of IMUs. In warehouse environments, various electrical devices and communication signals may exist, all of which generate EMI. To reduce the impact of EMI on the IMU, shielding measures need to be implemented around the IMU, and EMI sensors should be used to monitor and record the intensity of EMI in real time.

[0027] The purpose of recording these temperature interference parameters and electromagnetic interference intensity is to compensate and correct them during subsequent data processing in order to improve the accuracy and reliability of the data.

[0028] Step S113: Perform timestamp alignment processing on the displacement increment data, the angular velocity offset, and the acceleration offset to obtain synchronized motion trajectory data.

[0029] Because the sampling times of the laser rangefinder and the inertial measurement unit may differ, the timestamps of the acquired displacement increment data, angular velocity offset, and acceleration offset may be inconsistent. To effectively integrate and analyze these data, timestamp alignment is necessary.

[0030] The specific method for timestamp alignment is to first determine a unified time reference. For example, the sampling time of the laser rangefinder or the sampling time of the inertial measurement unit can be chosen as the reference. In this embodiment, the sampling time of the laser rangefinder is chosen as the reference.

[0031] Then, based on this time reference, interpolation is performed on the displacement increment data, angular velocity offset, and acceleration offset. The purpose of interpolation is to obtain the values ​​of these data at a uniform point in time. For example, if the laser rangefinder collects displacement increment data at a certain point in time, but the inertial measurement unit does not collect the corresponding angular velocity offset and acceleration offset at that point in time, the angular velocity offset and acceleration offset at that point in time can be estimated by interpolation based on the angular velocity offset and acceleration offset data from before and after that point in time.

[0032] Synchronized motion trajectory data was obtained through timestamp alignment. This data is consistent in time and can accurately reflect the motion trajectory of the AGV in three-dimensional space.

[0033] Step S114: Normalize the temperature interference parameters and electromagnetic interference intensity to obtain standardized environmental interference parameters.

[0034] The collected temperature interference parameters and electromagnetic interference intensity values ​​may have different ranges. To facilitate subsequent data analysis and processing, they need to be normalized. The purpose of normalization is to map these data to a uniform numerical range, making different types of environmental interference parameters comparable.

[0035] The normalization process can be performed using linear normalization. First, determine the maximum and minimum values ​​of the temperature interference parameter and the electromagnetic interference intensity. Then, for each temperature interference parameter and electromagnetic interference intensity data, normalize it using the following formula: Normalized value = (Original value - Minimum value) / (Maximum value - Minimum value).

[0036] Using the above method, the temperature interference parameters and electromagnetic interference intensity data were normalized to the range of 0 to 1, and standardized environmental interference parameters were obtained.

[0037] Step S115: Perform spatial coordinate mapping between the synchronized motion trajectory data and the standardized environmental interference parameters to generate a multi-source sensor data set containing device motion trajectory data, position coordinate offset, and environmental interference parameters.

[0038] After obtaining synchronized motion trajectory data and standardized environmental interference parameters, they need to be mapped to spatial coordinates to generate a multi-source sensor data set containing device motion trajectory data, position coordinate offsets, and environmental interference parameters.

[0039] A specific method for spatial coordinate mapping is to associate the displacement increment data and position coordinate offset in the synchronized motion trajectory data with standardized environmental disturbance parameters. These can be mapped using timestamps, ensuring that the displacement increment data and position coordinate offset at each moment correspond to the relevant standardized environmental disturbance parameters.

[0040] For example, at a certain point in time, the synchronized motion trajectory data records the displacement increments and position coordinate offsets of the AGV in the X, Y, and Z axis directions, while the standardized environmental interference parameters record the temperature interference parameters and electromagnetic interference intensity at that point in time. Correlating these data results in a multi-source sensor data set containing equipment motion trajectory data, position coordinate offsets, and environmental interference parameters.

[0041] This multi-source sensor dataset is a multi-dimensional dataset that includes the AGV's motion information in three-dimensional space and environmental interference information.

[0042] Step S120: Extract features from the multi-source sensor data set to generate a target feature set, which includes spatial location correlation features, dynamic trajectory fluctuation features, and environmental coupling features.

[0043] After obtaining the multi-source sensor data set, feature extraction is required to generate a target feature set in order to better analyze and process this data. The target feature set includes spatial location correlation features, dynamic trajectory fluctuation features, and environmental coupling features. These features can more effectively reflect the motion state of the AGV and the environmental influence.

[0044] Step S121: Extract the trajectory change rate within a continuous time window from the device motion trajectory data, and construct a spatial location association feature based on the trajectory change rate. The spatial location association feature includes the coordinate transformation matrix between adjacent calibration points.

[0045] In equipment motion trajectory data, the trajectory change rate reflects the changes in the AGV's speed and direction within a continuous time window. To extract the trajectory change rate, a suitable time window can be selected, such as a time window of 10 sampling moments. Within this time window, the ratio of the AGV's displacement change in the X, Y, and Z axes to time is calculated to obtain the trajectory change rate.

[0046] Based on the trajectory change rate, spatial location association features can be constructed. These features include the coordinate transformation matrix between adjacent calibration points. Calibration points are representative points selected on the AGV's movement trajectory. By calculating the coordinate transformation matrix between adjacent calibration points, the motion relationship of the AGV at different positions can be described.

[0047] The coordinate transformation matrix can be calculated using the rigid body transformation method. Assume that the AGV undergoes translation and rotation between two adjacent calibration points. The coordinate transformation matrix can be formed by calculating the translation vector and the rotation matrix. The translation vector describes the translation distance of the AGV along the three coordinate axes, while the rotation matrix describes the rotation angle and direction of the AGV.

[0048] By extracting the trajectory change rate within a continuous time window and constructing spatial location association features based on this, the positional changes and motion relationships of AGVs in three-dimensional space can be described more accurately.

[0049] Step S122: Perform frequency domain transformation on the position coordinate offset, extract the low-frequency component as the reference position feature, and extract the high-frequency component as the dynamic trajectory fluctuation feature.

[0050] The position coordinate offset reflects the deviation of the AGV from its ideal position during movement. To better analyze these deviations, the position coordinate offset needs to be processed by frequency domain transformation.

[0051] Frequency domain transformation can be performed using the Fourier transform method. The Fourier transform can convert a time-domain signal into a frequency-domain signal, thereby decomposing the position coordinate offset into components of different frequencies.

[0052] In the frequency domain signal, the low-frequency component represents the slowly changing portion of the position coordinate offset. This change is usually caused by the long-term motion trend of the AGV or systematic errors. Extracting the low-frequency component as a reference position feature can reflect the basic position state of the AGV.

[0053] High-frequency components represent the rapidly changing portion of the position coordinate offset, which is usually caused by jitter, vibration, or external interference during the AGV's movement. Extracting high-frequency components as dynamic trajectory fluctuation features can reflect the dynamic changes of the AGV during its movement.

[0054] By performing frequency domain transformation on the position coordinate offset, low-frequency components are extracted as reference position features, and high-frequency components are extracted as dynamic trajectory fluctuation features, which allows for a more in-depth analysis of the AGV's motion state and position deviation.

[0055] Step S123: Perform coupling analysis on the environmental interference parameters and the dynamic trajectory fluctuation characteristics, calculate the contribution weight of the environmental parameters to the trajectory fluctuation, and generate environmental coupling characteristics based on the contribution weight.

[0056] There may be a certain coupling relationship between environmental disturbance parameters and dynamic trajectory fluctuation characteristics. In order to analyze this coupling relationship, coupling degree analysis is required.

[0057] Coupling analysis can be performed using correlation analysis. By calculating the correlation coefficient between environmental disturbance parameters and dynamic trajectory fluctuation characteristics, the degree of coupling between them can be assessed. The closer the correlation coefficient is to 1, the higher the degree of coupling; the closer the correlation coefficient is to 0, the lower the degree of coupling.

[0058] Based on the coupling degree analysis, the contribution weights of environmental parameters to trajectory fluctuations are calculated. These contribution weights reflect the degree of influence of environmental disturbance parameters on the dynamic trajectory fluctuation characteristics. The contribution weights can be obtained by normalizing the correlation coefficients.

[0059] Based on the contribution weights, environmental coupling features are generated. These environmental coupling features are feature vectors that comprehensively consider environmental disturbance parameters and dynamic trajectory fluctuation characteristics. They are weighted combinations of these parameters, ensuring that the impact of environmental disturbance parameters on trajectory fluctuations is more accurately reflected in the feature vector.

[0060] By performing coupling analysis between environmental disturbance parameters and dynamic trajectory fluctuation characteristics, calculating the contribution weight of environmental parameters to trajectory fluctuations, and generating environmental coupling characteristics based on this, the influence of environmental factors on AGV motion trajectory can be considered more comprehensively.

[0061] Step S124: Perform dimensional alignment processing on the spatial location association features, the dynamic trajectory fluctuation features, and the environmental coupling features to generate a target feature set with a unified timestamp and spatial coordinate reference system.

[0062] Since spatial location association features, dynamic trajectory fluctuation features, and environmental coupling features may have different dimensions and timestamps, dimension alignment processing is required to combine spatial location association features, dynamic trajectory fluctuation features, and environmental coupling features into a unified target feature set.

[0063] The first step in dimensional alignment is to unify timestamps. A unified time base can be selected, and the timestamps of spatial location association features, dynamic trajectory fluctuation features, and environmental coupling features can be adjusted to this time base, thereby ensuring that they are consistent in time.

[0064] The second step is to unify the spatial coordinate reference system. Spatial location association features, dynamic trajectory fluctuation features, and environmental coupling features may be defined based on different spatial coordinate reference systems, and they need to be transformed into the same spatial coordinate reference system. A fixed spatial coordinate reference system can be selected as the benchmark, and the coordinates of other features can be transformed to that benchmark coordinate system.

[0065] Through dimensional alignment processing, spatial location association features, dynamic trajectory fluctuation features, and environmental coupling features are combined into a target feature set with a unified timestamp and spatial coordinate reference system. This target feature set is a multi-dimensional feature vector that contains the AGV's position, motion, and environmental information in three-dimensional space.

[0066] Step S130: Call the pre-trained adaptive fitting model, input the target feature set into the adaptive fitting model for dynamic parameter matching processing, and generate the initial calibration parameter set of the target device.

[0067] After obtaining the target feature set, a pre-trained adaptive fitting model needs to be invoked for dynamic parameter matching to generate the initial calibration parameter set for the target device. The adaptive fitting model is an artificial intelligence model trained on a large amount of data. It can automatically adjust the model's parameters based on the input target feature set to achieve the best fit to the target device.

[0068] Step S131: Input the spatial location association features into the first feature encoding layer of the adaptive fitting model to generate a spatial location encoding vector.

[0069] The first feature encoding layer of the adaptive fitting model is specifically designed to process spatial location-related features. After the spatial location-related features are input into the first feature encoding layer, the layer encodes them, converting them into spatial location encoded vectors.

[0070] The first feature encoding layer typically contains multiple neurons and a weight matrix. During the encoding process, spatial location-related features are multiplied and summed with the weight matrix, and then subjected to a non-linear transformation through an activation function to finally obtain the spatial location encoding vector.

[0071] The spatial location encoding vector is a multi-dimensional vector that effectively compresses and represents spatial location-related features. This vector contains the AGV's positional relationships and motion information in three-dimensional space, and will be used for subsequent feature fusion and parameter decoding.

[0072] Step S132: Input the dynamic trajectory fluctuation features into the second feature encoding layer of the adaptive fitting model to generate a trajectory fluctuation encoding vector.

[0073] The second feature encoding layer of the adaptive fitting model is used to process dynamic trajectory fluctuation features. After the dynamic trajectory fluctuation features are input into the second feature encoding layer, the second feature encoding layer encodes them to generate trajectory fluctuation encoding vectors.

[0074] The second feature coding layer works in a similar way to the first feature coding layer. It also contains multiple neurons and weight matrices. By performing operations and nonlinear transformations on the dynamic trajectory fluctuation features, it converts them into trajectory fluctuation coding vectors.

[0075] The trajectory fluctuation encoding vector is a multi-dimensional vector that can more effectively represent the dynamic changes of the AGV during its movement. This trajectory fluctuation encoding vector will be fused with the spatial position encoding vector and the environment coupling encoding vector to achieve more accurate parameter matching.

[0076] Step S133: Input the environmental coupling features into the third feature encoding layer of the adaptive fitting model to generate an environmental coupling encoding vector.

[0077] The third feature encoding layer of the adaptive fitting model is used to process environmental coupling features. After the environmental coupling features are input into the third feature encoding layer, the third feature encoding layer can encode them to generate environmental coupling encoded vectors.

[0078] The third feature encoding layer also contains multiple neurons and weight matrices. By performing operations and nonlinear transformations on the environmental coupling features, it converts them into environmental coupling encoding vectors.

[0079] The environment coupling encoding vector is a multi-dimensional vector that integrates information from environmental disturbance parameters and dynamic trajectory fluctuation characteristics, reflecting the impact of environmental factors on the AGV's motion trajectory. This environment coupling encoding vector, along with the spatial position encoding vector and the trajectory fluctuation encoding vector, will participate in subsequent attention weight allocation processing.

[0080] Step S134: Perform attention weight allocation processing on the spatial location encoding vector, trajectory fluctuation encoding vector and environment coupling encoding vector to obtain a weighted fusion feature vector.

[0081] After obtaining the spatial location encoding vector, trajectory fluctuation encoding vector, and environment coupling encoding vector, they need to be processed by attention weight allocation to obtain a weighted fused feature vector. The purpose of attention weight allocation is to weight and fuse them according to the importance of different features, thereby improving the performance of the model.

[0082] Step S1341: Calculate the first similarity score between the spatial location encoding vector and the trajectory fluctuation encoding vector.

[0083] In this embodiment, to calculate the first similarity score between the spatial location encoding vector and the trajectory fluctuation encoding vector, a cosine similarity method can be used, for example. Cosine similarity measures the similarity between two vectors by calculating the cosine of the angle between them.

[0084] The spatial location encoding vector and the trajectory fluctuation encoding vector are understood as two vectors in a high-dimensional space. By aligning the dimensions of the spatial location encoding vector and the trajectory fluctuation encoding vector, their dot product is calculated, and then divided by the product of their magnitudes to obtain the first similarity score. The closer the first similarity score is to 1, the more similar the two vectors are; the closer the score is to 0, the less similar the two vectors are.

[0085] Step S1342: Calculate the second similarity score between the trajectory fluctuation encoding vector and the environment coupling encoding vector.

[0086] In this embodiment, the cosine similarity method is also used to calculate the second similarity score between the trajectory fluctuation coding vector and the environment coupling coding vector. The second similarity score is obtained by calculating the dot product of these two vectors and dividing by the product of their magnitudes.

[0087] The second similarity score reflects the degree of similarity between trajectory fluctuation characteristics and environmental coupling characteristics. A higher second similarity score indicates that environmental factors have a greater impact on trajectory fluctuations; a lower second similarity score indicates that environmental factors have a smaller impact on trajectory fluctuations.

[0088] Step S1343: Calculate the third similarity score between the environment coupling encoding vector and the spatial location encoding vector.

[0089] In this embodiment, the cosine similarity method is still used to calculate the third similarity score between the environment coupling encoding vector and the spatial location encoding vector. The third similarity score is obtained by calculating the dot product of these two vectors and dividing by the product of their magnitudes.

[0090] The third similarity score reflects the degree of similarity between environmental factors and spatial location, which can help the model better understand the impact of environmental factors on AGV position changes.

[0091] Step S1344: Perform softmax normalization on the first similarity score, the second similarity score, and the third similarity score to obtain the spatial location attention weight, trajectory fluctuation attention weight, and environment coupling attention weight.

[0092] After obtaining the first, second, and third similarity scores, softmax normalization is needed to convert these scores into appropriate attention weights. Softmax normalization maps different similarity scores to a probability distribution, ensuring that the sum of all attention weights is 1.

[0093] The specific processing procedure is as follows: First, an exponential operation is performed on each similarity score to highlight the differences between scores. Then, each exponentially calculated score is divided by the sum of all exponentially calculated scores to obtain the corresponding attention weight. For example, for the first similarity score, after exponential operation, it is divided by the sum of the scores after three exponential operations to obtain the spatial location attention weight. Similarly, trajectory fluctuation attention weights and environmental coupling attention weights can be obtained, which reflect the importance of different features in the subsequent fusion process.

[0094] Step S1345: Scale the spatial position encoding vector according to the spatial position attention weight, scale the trajectory fluctuation encoding vector according to the trajectory fluctuation attention weight, scale the environment coupling encoding vector according to the environment coupling attention weight, and fuse the scaled spatial position encoding vector, trajectory fluctuation encoding vector, and environment coupling encoding vector element by element to generate a weighted fusion feature vector.

[0095] In this embodiment, after obtaining the spatial location attention weight, trajectory fluctuation attention weight, and environment coupling attention weight, the corresponding encoding vectors are scaled using these weights respectively. For the spatial location encoding vector, each element is multiplied by the spatial location attention weight; for the trajectory fluctuation encoding vector, each element is multiplied by the trajectory fluctuation attention weight; and for the environment coupling encoding vector, each element is multiplied by the environment coupling attention weight.

[0096] The three scaled encoding vectors are dimensionally matched, and then element-wise fusion is performed. Element-wise fusion combines corresponding elements from the three scaled encoding vectors to form a weighted fused feature vector. This weighted fused feature vector integrates information from spatial location, dynamic trajectory fluctuations, and environmental coupling, providing a more comprehensive reflection of the AGV's motion state and environmental influences.

[0097] Step S135: Call the parameter decoder of the adaptive fitting model to perform inverse mapping processing on the weighted fusion feature vector to generate an initial calibration parameter set containing position compensation parameters, motion acceleration correction coefficients and environmental interference suppression coefficients.

[0098] The parameter decoder of the adaptive fitting model receives the weighted fused feature vector and performs inverse mapping on it. The parameter decoder is an important module in the adaptive fitting model, and its role is to convert the weighted fused feature vector into the initial set of calibration parameters needed in practice.

[0099] During the inverse mapping process, the neurons and weight matrix inside the parameter decoder perform a series of operations and transformations on the weighted fused feature vector. Through this method, the information contained in the weighted fused feature vector is decoded into position compensation parameters, motion acceleration correction coefficients, and environmental interference suppression coefficients. The position compensation parameters are used to correct the actual position of the AGV, making it closer to the ideal position; the motion acceleration correction coefficients are used to adjust the motion acceleration of the AGV to improve its motion stability; and the environmental interference suppression coefficients are used to suppress the influence of environmental interference on the AGV's motion.

[0100] Step S140: Generate a dynamic path sequence for device position calibration based on the initial calibration parameter set, wherein the dynamic path sequence includes spatial coordinate adjustment parameters for multiple consecutive calibration points.

[0101] After obtaining the initial set of calibration parameters, a dynamic path sequence for device position calibration needs to be generated based on these parameters. The dynamic path sequence describes the movement path of the AGV from its current position to the target position and includes spatial coordinate adjustment parameters for multiple consecutive calibration points.

[0102] Step S141: Calculate the coordinate difference between the starting point and the ending point of the calibration path of the target device in three-dimensional space based on the position compensation parameters.

[0103] The position compensation parameters include position correction information for the AGV in the X, Y, and Z axis directions. Based on this information, the coordinate differences between the start and end points of the calibration path in these three axes can be calculated. For example, for the X-axis direction, the X-coordinate of the starting point is subtracted from the X-coordinate of the end point, and then the X-axis correction value from the position compensation parameters is added to obtain the coordinate difference in the X-axis direction. Similarly, the coordinate differences in the Y and Z axis directions can be calculated, reflecting the distance the AGV needs to move in these three axes during calibration.

[0104] Step S142: Perform piecewise linear interpolation on the coordinate difference based on the motion acceleration correction coefficient to generate theoretical coordinate values ​​for multiple intermediate calibration points.

[0105] In this embodiment, after obtaining the coordinate difference, piecewise linear interpolation is performed using a motion acceleration correction coefficient. The purpose of piecewise linear interpolation is to divide the calibration path into multiple segments, each corresponding to an intermediate calibration point. During the interpolation process, the length and slope of each segment are determined based on the motion acceleration correction coefficient, ensuring that the AGV's acceleration during movement meets the requirements.

[0106] For example, the specific interpolation process involves first dividing the coordinate difference according to a set rule based on the motion acceleration correction coefficient, resulting in multiple segments of varying lengths. Then, based on the lengths of these segments and the starting coordinates, the theoretical coordinate values ​​of each intermediate calibration point are calculated sequentially. For instance, in the X-axis direction, starting from the X-coordinate of the starting point, the theoretical X-coordinate values ​​of each intermediate calibration point are obtained by sequentially accumulating the values ​​according to the lengths of the segments. Similarly, the theoretical coordinate values ​​in the Y-axis and Z-axis directions can be calculated.

[0107] Step S143: Combine the environmental interference suppression coefficient to dynamically adjust the theoretical coordinate values ​​of each intermediate calibration point, and generate a dynamic path sequence that includes the actual coordinate values, adjustment time interval and motion direction parameters.

[0108] Step S1431: Obtain the theoretical coordinates of the current intermediate calibration point and the corresponding environmental interference suppression coefficient. Perform scalar multiplication on the environmental interference suppression coefficient and each spatial dimension component of the theoretical coordinates to generate a three-dimensional environmental compensation vector for each intermediate calibration point.

[0109] For each intermediate calibration point, its theoretical coordinates and corresponding environmental interference suppression coefficient are first obtained. The environmental interference suppression coefficient is a coefficient that comprehensively considers various environmental interference factors, reflecting the degree of influence of environmental interference on the AGV's position.

[0110] The environmental interference suppression coefficient is multiplied by the theoretical coordinate values ​​along the X, Y, and Z axes using scalar multiplication. For example, for the X-axis, the environmental interference suppression coefficient is multiplied by the X-axis component of the theoretical coordinate value to obtain the environmental compensation component in the X-axis direction; similarly, the environmental compensation components in the Y and Z axes are obtained. These three environmental compensation components form a three-dimensional environmental compensation vector. The three-dimensional environmental compensation vector reflects the correction information of environmental interference on the position of the intermediate calibration point.

[0111] Step S1432: Determine the dynamic adjustment priority of each axis according to the component amplitude of the three-dimensional environment compensation vector in the X-axis, Y-axis and Z-axis, and superimpose the three-dimensional environment compensation vector on the corresponding axial components of the theoretical coordinate value in order of priority from high to low to generate the actual coordinate value.

[0112] The amplitudes of the three-dimensional environmental compensation vector components along the X, Y, and Z axes are analyzed. Larger amplitudes indicate more severe environmental interference along that axis, and thus higher priority. The corresponding components of the three-dimensional environmental compensation vector are then superimposed onto the corresponding axial components of the theoretical coordinate values ​​in descending order of priority.

[0113] For example, if the amplitude of the component along the X-axis is the largest, followed by the Y-axis, and the smallest along the Z-axis, then first add the X-axis component of the 3D environment compensation vector to the X-axis component of the theoretical coordinate value to obtain the adjusted actual X-axis coordinate value; then add the Y-axis component of the 3D environment compensation vector to the Y-axis component of the theoretical coordinate value to obtain the adjusted actual Y-axis coordinate value; finally, add the Z-axis component of the 3D environment compensation vector to the Z-axis component of the theoretical coordinate value to obtain the adjusted actual Z-axis coordinate value. This yields the actual coordinate value of each intermediate calibration point.

[0114] Step S1433: Based on the ratio of the total amplitude of the three-dimensional environmental compensation vector to the preset reference adjustment speed, calculate the adjustment time interval for each intermediate calibration point, wherein the total amplitude of the three-dimensional environmental compensation vector is the result of the square root of the sum of the squares of each axial component.

[0115] The total magnitude of the three-dimensional environmental compensation vector is calculated, which reflects the overall impact of environmental disturbances on the intermediate calibration point position. The preset baseline adjustment speed is a pre-defined constant, representing the basic speed of the AGV when adjusting its position.

[0116] Dividing the total magnitude of the 3D environmental compensation vector by the preset baseline adjustment speed yields the adjustment time interval for each intermediate calibration point. The adjustment time interval represents the time required for the AGV to adjust from its theoretical coordinates to its actual coordinates. The calculation of the adjustment time interval takes into account the degree of environmental interference; the greater the environmental interference, the longer the adjustment time interval.

[0117] Step S1434: Determine the motion direction parameters based on the positive or negative signs of each axial component of the three-dimensional environment compensation vector. The motion direction parameters include positive or negative movement indicators for the X-axis, Y-axis, and Z-axis.

[0118] Observe the signs of the components of the 3D environment compensation vector along the X, Y, and Z axes. If the X-axis component is positive, the AGV needs to move in the positive direction along the X-axis, and the corresponding X-axis movement indicator is positive. If the X-axis component is negative, the AGV needs to move in the negative direction along the X-axis, and the corresponding X-axis movement indicator is negative. Similarly, the movement indicators in the Y and Z axes can be determined. The motion direction parameters clearly define the direction in which the AGV needs to move at each intermediate calibration point.

[0119] Step S1435: Arrange the actual coordinate values, adjustment time intervals, and motion direction parameters in chronological order to generate an initial dynamic path sequence.

[0120] The actual coordinates, adjustment time intervals, and movement direction parameters of each intermediate calibration point are arranged in chronological order to form an initial dynamic path sequence. This initial dynamic path sequence records the position information, adjustment time, and movement direction of each intermediate calibration point during the AGV's calibration process, providing detailed guidance for the AGV's actual movement.

[0121] Step S1436: Perform kinematic feasibility verification on each actual coordinate value in the initial dynamic path sequence. When the displacement between adjacent actual coordinate values ​​exceeds the product of the maximum movement speed of the device and the adjustment time interval, scale the adjustment time interval proportionally until the displacement constraint condition is met.

[0122] The kinematic feasibility of each actual coordinate value in the initial dynamic path sequence is verified, mainly by checking whether the displacement between adjacent actual coordinate values ​​meets the AGV's motion capability. The AGV has a maximum speed limit; if the displacement between adjacent actual coordinate values ​​exceeds the product of the device's maximum speed and the adjustment time interval, it indicates that the AGV cannot complete the displacement within that time interval.

[0123] When a non-compliance with displacement constraints is detected, the adjustment time interval needs to be scaled proportionally. This scaling process involves increasing the adjustment time interval by a set ratio until the displacement between adjacent actual coordinate values ​​satisfies the constraint that the maximum speed of the device is multiplied by the adjustment time interval.

[0124] Step S1437: Update the time parameters in the initial dynamic path sequence according to the scaled adjustment time interval to generate an intermediate dynamic path sequence.

[0125] After scaling the adjustment time interval proportionally, the time parameters in the initial dynamic path sequence are updated according to the scaled adjustment time interval. The adjustment time of each intermediate calibration point is replaced with the scaled adjustment time, thus obtaining the intermediate dynamic path sequence. The intermediate dynamic path sequence arranges the AGV's movement time more rationally while satisfying kinematic feasibility.

[0126] Step S1438: Insert device attitude stabilization detection points into the intermediate dynamic path sequence. The insertion position of the attitude stabilization detection points is the median time of the adjustment time interval between every two adjacent intermediate calibration points. Add attitude maintenance commands to the attitude stabilization detection points to suppress inertial drift.

[0127] To ensure the attitude stability of the AGV during movement, attitude stability detection points are inserted into the intermediate dynamic path sequence. The insertion position of the attitude stability detection point is the median of the adjustment time interval between every two adjacent intermediate calibration points.

[0128] An attitude maintenance command is added at each attitude stability detection point. The purpose of the attitude maintenance command is to suppress the inertial drift of the AGV. During the movement of the AGV, due to inertia, the attitude may deviate. The attitude maintenance command can adjust the attitude control parameters of the AGV to maintain a stable attitude.

[0129] Step S1439: Redistribute the adjustment time interval of each intermediate calibration point according to the number of insertions of the attitude stabilization detection points, to ensure that the execution time of the attitude maintenance command is included in the total adjustment time interval.

[0130] After inserting attitude stabilization detection points, the adjustment time intervals for each intermediate calibration point need to be reallocated based on the number of insertions. The purpose of this reallocation is to ensure that the execution time of attitude maintenance commands is included within the total adjustment time interval, while maintaining the total time of the entire dynamic path sequence.

[0131] The specific redistribution process involves appropriately adjusting the adjustment time interval between adjacent intermediate calibration points based on the insertion position and number of attitude stabilization detection points. This ensures that each attitude stabilization detection point has sufficient time to execute attitude maintenance commands without affecting the total time for the AGV to reach the target position.

[0132] Step S1440: Merge the updated intermediate dynamic path sequence with the attitude stabilization detection points to generate a dynamic path sequence containing actual coordinate values, adjustment time intervals, and motion direction parameters.

[0133] The updated intermediate dynamic path sequence is merged with the attitude stabilization detection points, meaning the information from the attitude stabilization detection points is incorporated into the intermediate dynamic path sequence. The merged sequence contains the actual coordinates, adjustment time intervals, and motion direction parameters of all intermediate calibration points and attitude stabilization detection points, forming the final dynamic path sequence. This dynamic path sequence provides complete guidance for the automated position calibration operation of the AGV.

[0134] Step S144: Perform collision detection and path smoothing on multiple consecutive calibration points in the dynamic path sequence to generate an optimized dynamic path sequence.

[0135] Collision detection is performed on multiple consecutive calibration points in a dynamic path sequence to check whether the AGV will collide with surrounding obstacles as it moves along the dynamic path sequence. This can be achieved by establishing a geometric model of the AGV and surrounding obstacles, and then determining whether the AGV will enter the space of the obstacles during its movement based on the position information of each calibration point in the dynamic path sequence.

[0136] If a collision risk is detected, the dynamic path sequence needs to be adjusted to avoid obstacles. Path smoothing is used to make the AGV's movement smoother and reduce unnecessary rapid acceleration and deceleration. Curve fitting can be used to fit the calibration points in the dynamic path sequence to generate a smooth path.

[0137] After collision detection and path smoothing, an optimized dynamic path sequence is obtained. The optimized dynamic path sequence ensures both the safety of AGV movement and improves the smoothness of movement.

[0138] Step S150: Trigger the target device to perform an automated position calibration operation according to the dynamic path sequence, and feed back the collected calibration verification data set to the adaptive fitting model to update the model parameters.

[0139] Step S151: Convert the dynamic path sequence into a device control instruction set, which includes motor drive parameters and servo control parameters for each calibration point.

[0140] The dynamic path sequence contains detailed information about the AGV during the calibration process. To ensure the AGV moves along this path, the dynamic path sequence needs to be converted into a device control instruction set. The device control instruction set includes the motor drive parameters and servo control parameters for each calibration point.

[0141] Motor drive parameters are used to control the AGV's motor, causing it to rotate at a specified speed and direction. Servo control parameters are used to precisely control the AGV's position and attitude, ensuring that the AGV can accurately reach each calibration point.

[0142] The conversion process involves calculating the corresponding motor drive parameters and servo control parameters based on the actual coordinates of each calibration point in the dynamic path sequence, the adjustment time interval, and the motion direction parameters, combined with the kinematic and dynamic models of the AGV.

[0143] Step S152: Monitor in real time the actual motion trajectory data of the target device when it performs calibration operation according to the control instruction set.

[0144] During the calibration process of the AGV according to the control instruction set, its actual motion trajectory data is monitored in real time. Real-time position and attitude information of the AGV during its movement can be obtained through sensors installed on the AGV, such as laser rangefinders and inertial measurement units.

[0145] These actual motion trajectory data are time-varying sequences, recording the AGV's actual position and posture at different moments. By monitoring the actual motion trajectory data in real time, it is possible to promptly detect any deviations that may occur during the AGV's movement.

[0146] Step S153: Collect the actual position coordinate data after calibration, and calculate the deviation with the theoretical coordinate values ​​in the dynamic path sequence to generate position calibration error data.

[0147] After calibration, the actual position coordinate data of the AGV is collected. The deviation is then calculated by comparing this actual position coordinate data with the theoretical coordinate values ​​in the dynamic path sequence. The deviation calculation involves subtracting the theoretical coordinate values ​​from the actual position coordinate data along the X, Y, and Z axes respectively, obtaining the deviation value for each coordinate axis.

[0148] These deviation values ​​constitute the position calibration error data, which reflects the accuracy of the AGV during the position calibration process.

[0149] Step S154: Combine the actual motion trajectory data, position calibration error data, and real-time environmental interference parameters into a verification data set.

[0150] The actual motion trajectory data obtained from real-time monitoring, the calculated position calibration error data, and the collected real-time environmental interference parameters are combined to form a verification dataset. The verification dataset contains various information about the AGV during the position calibration process.

[0151] Step S155: Input the verification data set into the error backpropagation link of the adaptive fitting model to update the weight parameters of the first feature encoding layer, the second feature encoding layer and the parameter decoder.

[0152] Step S1551: Extract the root mean square error between the actual motion trajectory data and the theoretical trajectory data from the verification data set as the first loss function.

[0153] Actual motion trajectory data and theoretical trajectory data from the dynamic path sequence are extracted from the validation dataset. The root mean square error (RMSE) between them is calculated, which measures the overall deviation between the actual and theoretical trajectories. This RMS error is used as the first loss function, reflecting the AGV's error on the motion trajectory.

[0154] Step S1552: Calculate the maximum offset of the position calibration error data in different spatial dimensions as the second loss function.

[0155] The maximum offsets in the X, Y, and Z axes are identified for the position calibration error data. These maximum offsets reflect the maximum deviation of the AGV in different spatial dimensions during position calibration. These maximum offsets are used as a second loss function, which focuses on the maximum error in position calibration.

[0156] Step S1553: Generate a third loss function based on the prediction error of the environmental coupling coding vector according to the real-time environmental interference parameters.

[0157] The real-time environmental disturbance parameters contain the latest information on environmental disturbances. These parameters are compared with the environmental coupling encoding vector to calculate the prediction error. The prediction error reflects the accuracy of the adaptive fitting model in predicting the impact of environmental disturbances. A third loss function is generated based on this prediction error, which reflects the model's performance in handling environmental disturbances.

[0158] Step S1554: Perform a weighted summation of the first loss function, the second loss function, and the third loss function to obtain the comprehensive loss value.

[0159] To comprehensively consider motion trajectory error, maximum position calibration error, and environmental interference prediction error, a weighted sum is performed on the first, second, and third loss functions. Each loss function is assigned a weight, the magnitude of which is determined based on the importance of each function.

[0160] The three loss functions are multiplied by their respective weights and then summed to obtain the overall loss value. The overall loss value reflects the total error of the adaptive fitting model throughout the entire position calibration process.

[0161] Step S1555: The adaptive momentum optimization algorithm is used to update the weights of the fully connected layers and the attention allocation parameters of the adaptive fitting model according to the comprehensive loss value.

[0162] The Adaptive Momentum Optimization (AMO) algorithm is an optimization algorithm used to adjust model parameters. It combines the concepts of momentum and adaptive learning rate, enabling more efficient updates to model parameters. In this embodiment, this algorithm is used to update the weights of the fully connected layers and the attention allocation parameters of the adaptively fitted model based on the comprehensive loss value.

[0163] Step S1555-1: Calculate the first gradient component of the integrated loss value on the spatial location encoding vector.

[0164] The overall loss is a function of the model parameters, while the spatial location encoding vector is the result of encoding the model's input features, and it affects the magnitude of the overall loss. To update the model parameters, the gradient of the overall loss with respect to the spatial location encoding vector needs to be calculated. The gradient represents the rate of change of the overall loss relative to the spatial location encoding vector. Using the backpropagation algorithm, starting from the overall loss, the gradient is calculated backward along the model's computational path to obtain the first gradient component of the overall loss with respect to the spatial location encoding vector. This first gradient component reflects the direction and extent of the influence of changes in the spatial location encoding vector on the overall loss.

[0165] Step S1555-2: Calculate the second gradient component of the trajectory fluctuation encoding vector with respect to the comprehensive loss value.

[0166] Similarly, for the trajectory fluctuation encoding vector, it is also necessary to calculate the gradient of the comprehensive loss value with respect to it. Using the backpropagation algorithm, in the computational graph of the model, starting from the comprehensive loss value, the gradient is calculated step-by-step in reverse to obtain the second gradient component of the comprehensive loss value with respect to the trajectory fluctuation encoding vector. The second gradient component reflects the impact of changes in the trajectory fluctuation encoding vector on the comprehensive loss value.

[0167] Step S1555-3: Calculate the third gradient component of the integrated loss value with respect to the environment-coupled encoding vector.

[0168] Similar to the previous steps, the third gradient component of the integrated loss value with respect to the environment-coupled encoding vector is calculated. By backpropagating along the computational path of the model, the rate of change of the integrated loss value relative to the environment-coupled encoding vector is determined, yielding the third gradient component. This third gradient component reflects the effect of changes in the environment-coupled encoding vector on the integrated loss value, and helps optimize environment-coupled parameters in the model.

[0169] Step S1555-4: Determine the momentum update direction for each parameter node based on the exponential weighted sum of the historical gradient moving average and the current gradient.

[0170] The adaptive momentum optimization algorithm introduces the concept of momentum, utilizing historical gradient information to smooth the parameter update process. First, for each parameter node, the moving average of its historical gradients is calculated. This historical gradient moving average is a weighted average of gradients over a past period, with the weights decaying exponentially over time. Then, the currently calculated gradient is exponentially weighted and summed with the historical gradient moving average. This summation comprehensively considers information from both historical and current gradients, and based on this result, the momentum update direction for each parameter node can be determined. The momentum update direction guides the parameters towards a more optimal direction, avoiding drastic oscillations during the parameter update process.

[0171] Step S1555-5: Iteratively update the weights of the fully connected layers of the adaptive fitting model according to the momentum update direction, and adjust the temperature coefficient in the attention allocation parameter to control the smoothness of softmax normalization.

[0172] After determining the momentum update direction for each parameter node, the weights of the fully connected layers in the adaptive fitting model are iteratively updated according to this direction. The iterative update process involves gradually adjusting the size of the fully connected layer weights to gradually reduce the overall loss value. Simultaneously, the temperature coefficient in the attention allocation parameters also needs to be adjusted. The temperature coefficient plays a crucial role in the softmax normalization process, controlling the smoothness of the softmax function's output. A larger temperature coefficient results in a smoother softmax output, reducing differences between different categories; a smaller temperature coefficient results in a sharper softmax output, highlighting categories with significant differences. By adjusting the temperature coefficient, the attention allocation can be made more reasonable, increasing the model's focus on different features.

[0173] Through the above series of steps, the parameters of the adaptive fitting model were updated. The updated model can better adapt to the actual situation of the target equipment, improving the accuracy and reliability of equipment position calibration.

[0174] Throughout the data acquisition and processing process, the protection of privacy-sensitive data is crucial. For example, device motion trajectory data and environmental interference parameters may contain sensitive information. To prevent these data leaks, various privacy protection and leak prevention technologies are employed. First, during the data acquisition phase, the data is encrypted using a symmetric encryption algorithm to ensure security during transmission. Second, for data storage, a secure storage system is used with access control, allowing only authorized personnel to access the data. Regular data backups are also performed to prevent data loss. Furthermore, during data processing, the data is anonymized to remove potentially personally identifiable information, protecting user privacy.

[0175] In practical applications, adaptive fitting models are closely integrated with intelligent warehousing systems. The input data for the adaptive fitting model is a collection of multi-source sensor data after acquisition and processing, including equipment motion trajectory data, position coordinate offsets, and environmental interference parameters. The output of the adaptive fitting model is the initial calibration parameter set and dynamic path sequence for equipment position calibration. This output data is directly used to control the movement of the AGV, enabling it to accurately complete cargo handling tasks. By continuously collecting and merging the calibrated validation dataset to update the model parameters, the adaptive fitting model can continuously learn and optimize, improving its ability to predict and control the AGV's motion state, thereby enhancing the operational efficiency and reliability of the entire intelligent warehousing system.

[0176] Figure 2 The diagram illustrates exemplary hardware and software components of an automated data fitting system 100 incorporating device location calibration, which can implement the inventive concept according to some embodiments of the present invention. For example, a processor 120 may be used in the automated data fitting system 100 incorporating device location calibration and to perform the functions of the present invention.

[0177] The automated data fitting system 100, which incorporates device location calibration, can be a general-purpose server or a special-purpose server; both can be used to implement the automated data fitting method incorporating device location calibration of this invention. Although only one server is shown in this invention, for convenience, the functions described herein can be implemented in a distributed manner on multiple similar platforms to balance the load.

[0178] For example, the automated data fitting system 100 incorporating device location calibration may include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and various forms of storage media 140, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the automated data fitting system 100 incorporating device location calibration may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The method of the present invention can be implemented according to these program instructions. The automated data fitting system 100 incorporating device location calibration also includes an I / O interface 150 between the computer and other input / output devices.

[0179] For ease of explanation, only one processor is described in the automated data fitting system 100 for device location calibration. However, it should be noted that the automated data fitting system 100 for device location calibration of the present invention may also include multiple processors, and therefore the steps performed by one processor described in the present invention may also be performed jointly by multiple processors or individually. For example, if the processor of the automated data fitting system 100 for device location calibration performs steps A and B, it should be understood that steps A and B may also be performed jointly by two different processors or individually by one processor. For example, the first processor performs step A, the second processor performs step B, or the first processor and the second processor jointly perform steps A and B.

[0180] Furthermore, embodiments of the present invention also provide a readable storage medium, wherein computer-executable instructions are preset in the readable storage medium, and when the processor executes the computer-executable instructions, the above-mentioned automated data fitting method combined with device location calibration is implemented.

[0181] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. An automated data fitting method combining equipment position calibration, characterized in that, The method includes: Collect a set of multi-source sensor data of the target device in three-dimensional space. The set of multi-source sensor data includes device motion trajectory data, position coordinate offset and environmental interference parameters. Feature extraction is performed on the multi-source sensor data set to generate a target feature set, which includes spatial location correlation features, dynamic trajectory fluctuation features, and environmental coupling features. The pre-trained adaptive fitting model is invoked, and the target feature set is input into the adaptive fitting model for dynamic parameter matching processing to generate the initial calibration parameter set of the target device. A dynamic path sequence for device location calibration is generated based on the initial calibration parameter set, and the dynamic path sequence includes spatial coordinate adjustment parameters for multiple consecutive calibration points; The target device is triggered to perform an automated position calibration operation according to the dynamic path sequence, and the collected calibration verification data set is fed back to the adaptive fitting model to update the model parameters. The step involves calling a pre-trained adaptive fitting model, inputting the target feature set into the adaptive fitting model for dynamic parameter matching, and generating an initial calibration parameter set for the target device, including: The spatial location association features are input into the first feature encoding layer of the adaptive fitting model to generate a spatial location encoding vector. The dynamic trajectory fluctuation features are input into the second feature encoding layer of the adaptive fitting model to generate a trajectory fluctuation encoding vector. The environmental coupling features are input into the third feature encoding layer of the adaptive fitting model to generate an environmental coupling encoding vector. Attention weight allocation is performed on the spatial location encoding vector, trajectory fluctuation encoding vector, and environment coupling encoding vector to obtain a weighted fusion feature vector. The parameter decoder of the adaptive fitting model is invoked to perform inverse mapping processing on the weighted fusion feature vector to generate an initial calibration parameter set containing position compensation parameters, motion acceleration correction coefficients, and environmental interference suppression coefficients.

2. The automated data fitting method combining equipment position calibration according to claim 1, characterized in that, The collection of multi-source sensor data of the target device in three-dimensional space includes: The displacement increment data of the target device in the X-axis, Y-axis and Z-axis are collected in real time by a laser ranging sensor array deployed on the motion plane of the target device; The inertial measurement unit is invoked to obtain the angular velocity offset and acceleration offset of the target device, and the temperature interference parameters and electromagnetic interference intensity at each sampling time are recorded; The displacement increment data, the angular velocity offset, and the acceleration offset are time-stamp aligned to obtain synchronized motion trajectory data; The temperature interference parameters and electromagnetic interference intensity are normalized to obtain standardized environmental interference parameters. The synchronized motion trajectory data is spatially mapped to the standardized environmental interference parameters to generate a multi-source sensor data set containing device motion trajectory data, position coordinate offset, and environmental interference parameters.

3. The automated data fitting method combining equipment position calibration according to claim 1, characterized in that, The step of extracting features from the multi-source sensor data set to generate a target feature set includes: Extract the trajectory change rate within a continuous time window from the device motion trajectory data, and construct a spatial location association feature based on the trajectory change rate. The spatial location association feature includes the coordinate transformation matrix between adjacent calibration points. The position coordinate offset is subjected to frequency domain transformation processing to extract the low frequency component as the reference position feature and the high frequency component as the dynamic trajectory fluctuation feature. The environmental disturbance parameters are coupled with the dynamic trajectory fluctuation characteristics to perform coupling analysis, calculate the contribution weight of the environmental parameters to the trajectory fluctuation, and generate environmental coupling characteristics based on the contribution weight. The spatial location association features, the dynamic trajectory fluctuation features, and the environmental coupling features are dimensionally aligned to generate a target feature set with a unified timestamp and spatial coordinate reference system.

4. The automated data fitting method combining equipment position calibration according to claim 1, characterized in that, The attention weight allocation process performed on the spatial location encoding vector, trajectory fluctuation encoding vector, and environment coupling encoding vector to obtain a weighted fusion feature vector includes: Calculate the first similarity score between the spatial location encoding vector and the trajectory fluctuation encoding vector; Calculate the second similarity score between the trajectory fluctuation encoding vector and the environment coupling encoding vector; Calculate the third similarity score between the environment coupling encoding vector and the spatial location encoding vector; The first similarity score, the second similarity score, and the third similarity score are subjected to softmax normalization to obtain spatial location attention weight, trajectory fluctuation attention weight, and environment coupling attention weight. The spatial location encoding vector is scaled according to the spatial location attention weight, the trajectory fluctuation encoding vector is scaled according to the trajectory fluctuation attention weight, and the environment coupling encoding vector is scaled according to the environment coupling attention weight. The scaled spatial location encoding vector, trajectory fluctuation encoding vector and environment coupling encoding vector are then fused element by element to generate a weighted fusion feature vector.

5. The automated data fitting method combining equipment position calibration according to claim 1, characterized in that, The generation of the dynamic path sequence for device location calibration based on the initial calibration parameter set includes: Calculate the coordinate difference between the starting point and the ending point of the calibration path of the target device in three-dimensional space based on the position compensation parameters; Based on the motion acceleration correction coefficient, the coordinate difference is subjected to piecewise linear interpolation to generate theoretical coordinate values ​​for multiple intermediate calibration points; The theoretical coordinates of each intermediate calibration point are dynamically adjusted based on the environmental interference suppression coefficient to generate a dynamic path sequence that includes the actual coordinates, adjustment time interval, and motion direction parameters. Collision detection and path smoothing are performed on multiple consecutive calibration points in the dynamic path sequence to generate an optimized dynamic path sequence.

6. The automated data fitting method combining equipment position calibration according to claim 5, characterized in that, The step of dynamically adjusting the theoretical coordinates of each intermediate calibration point based on the environmental interference suppression coefficient to generate a dynamic path sequence containing actual coordinates, adjustment time intervals, and motion direction parameters includes: Obtain the theoretical coordinates of the current intermediate calibration point and the corresponding environmental interference suppression coefficient. Perform scalar multiplication on the environmental interference suppression coefficient and each spatial dimension component of the theoretical coordinates to generate a three-dimensional environmental compensation vector for each intermediate calibration point. The dynamic adjustment priority of each axis is determined based on the component amplitudes of the three-dimensional environmental compensation vector in the X, Y, and Z axes. The three-dimensional environmental compensation vector is then superimposed on the corresponding axial components of the theoretical coordinate values ​​in descending order of priority to generate the actual coordinate values. Based on the ratio of the total amplitude of the three-dimensional environmental compensation vector to the preset reference adjustment speed, the adjustment time interval of each intermediate calibration point is calculated, wherein the total amplitude of the three-dimensional environmental compensation vector is the result of the square root of the sum of the squares of each axial component. The motion direction parameters are determined based on the positive or negative sign of each axial component of the three-dimensional environment compensation vector. The motion direction parameters include positive or negative movement indicators for the X-axis, Y-axis, and Z-axis. Arrange the actual coordinate values, adjustment time intervals, and motion direction parameters in chronological order to generate an initial dynamic path sequence; The kinematic feasibility of each actual coordinate value in the initial dynamic path sequence is verified. When the displacement between adjacent actual coordinate values ​​exceeds the product of the maximum movement speed of the device and the adjustment time interval, the adjustment time interval is scaled proportionally until the displacement constraint condition is met. The time parameters in the initial dynamic path sequence are updated according to the scaled adjustment time interval to generate an intermediate dynamic path sequence; Insert device attitude stabilization detection points into the intermediate dynamic path sequence. The insertion position of the attitude stabilization detection points is the median time of the adjustment time interval between every two adjacent intermediate calibration points. Add attitude maintenance commands to the attitude stabilization detection points to suppress inertial drift. The adjustment time interval of each intermediate calibration point is redistributed according to the number of insertions of the attitude stabilization detection points to ensure that the execution time of the attitude maintenance command is included in the total adjustment time interval. The updated intermediate dynamic path sequence is merged with the attitude stabilization detection points to generate a dynamic path sequence that includes actual coordinate values, adjustment time intervals, and motion direction parameters.

7. The automated data fitting method combining equipment position calibration according to claim 1, characterized in that, The step of triggering the target device to perform an automated position calibration operation according to the dynamic path sequence, and collecting the calibration verification data set and feeding it back to the adaptive fitting model to update the model parameters, includes: The dynamic path sequence is converted into a device control instruction set, which includes motor drive parameters and servo control parameters for each calibration point. Real-time monitoring of the actual motion trajectory data of the target device when it performs calibration operations according to the control instruction set; Collect the actual position coordinate data after calibration, and calculate the deviation with the theoretical coordinate values ​​in the dynamic path sequence to generate position calibration error data; The actual motion trajectory data, position calibration error data, and real-time environmental interference parameters are combined into a verification data set. The validation dataset is input into the error backpropagation path of the adaptive fitting model to update the weight parameters of the first feature encoding layer, the second feature encoding layer, and the parameter decoder.

8. The automated data fitting method combining equipment position calibration according to claim 7, characterized in that, The step of inputting the verification data set into the error backpropagation path of the adaptive fitting model to update the weight parameters of the first feature encoding layer, the second feature encoding layer, and the parameter decoder includes: The root mean square error between the actual motion trajectory data and the theoretical trajectory data is extracted from the verification dataset and used as the first loss function. The maximum offset of the position calibration error data in different spatial dimensions is calculated as the second loss function; A third loss function is generated based on the prediction error of the environmental coupling coding vector according to the real-time environmental interference parameters; The first loss function, the second loss function, and the third loss function are weighted and summed to obtain the comprehensive loss value. An adaptive momentum optimization algorithm is used to update the weights of the fully connected layers and the attention allocation parameters of the adaptive fitting model based on the comprehensive loss value; The parameter update process of the adaptive momentum optimization algorithm includes: Calculate the first gradient component of the integrated loss value with respect to the spatial location encoding vector; Calculate the second gradient component of the trajectory fluctuation encoding vector with respect to the comprehensive loss value; Calculate the third gradient component of the integrated loss value with respect to the environment-coupled encoding vector; The momentum update direction for each parameter node is determined by the exponentially weighted sum of the historical gradient moving average and the current gradient. The weights of the fully connected layers of the adaptive fitting model are iteratively updated according to the momentum update direction, while the temperature coefficient in the attention allocation parameter is adjusted to control the smoothness of softmax normalization.

9. An automated data fitting system combining equipment position calibration, characterized in that, The automated data fitting system combining device location calibration includes a processor and a memory, the memory and the processor being connected. The memory is used to store programs, instructions or code, and the processor is used to execute the programs, instructions or code in the memory to implement the automated data fitting method combining device location calibration as described in any one of claims 1-8.

Images