Comprehensive method for correcting parabola trajectory deviation of movement speed of stacking machine

Abstract

The invention discloses a stacking machine motion speed parabolic trajectory deviation correction comprehensive method, and relates to the technical field of stacking machine trajectory deviation correction, and the method comprises the following steps: carrying out the data collection of the real-time motion trajectory of a stacking machine through an acceleration sensor, an encoder and a visual recognition system, building a trajectory deviation detection model, and carrying out the calculation of the trajectory deviation detection model; the motion speed, the acceleration and the position information are extracted, and the deviation value of the current trajectory deviating from the ideal parabolic trajectory is calculated. According to the invention, through multi-sensor fusion and Kalman filtering, the accuracy and stability of track correction of the stacker are improved; an LSTM neural network is adopted to predict inertial errors, and correction failures are reduced in combination with an adaptive compensation strategy; a double-closed-loop control and anomaly detection mechanism is introduced, intelligent safety protection is achieved, the fault recovery capacity is improved, it is ensured that the stacking machine stably operates in a complex environment, and the reliability and working efficiency of an automatic warehousing system are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of stacker trajectory deviation correction, and particularly to a comprehensive method for correcting the parabolic trajectory deviation of the stacker movement speed. Background Art

[0002] "Comprehensive correction of the parabolic trajectory deviation of the stacker movement speed" refers to a technical method for comprehensively correcting the parabolic deviation between the movement speed and the trajectory of the stacker during its operation. During the operation of the stacker, especially during high-speed operation, start-up, stop or turning, its movement trajectory may deviate from the ideal parabolic trajectory, resulting in reduced positioning accuracy, decreased operation efficiency, and even possible impact on the stability of the stacking operation. This technology comprehensively considers the speed curve, movement trajectory, load change, frictional resistance and response characteristics of the control system of the stacker, and uses intelligent control algorithms, feedback correction mechanisms or dynamic compensation technologies to monitor and correct the deviation in real time, ensuring that the stacker can maintain an accurate parabolic movement trajectory during operation, improving operation efficiency and accuracy, and reducing energy consumption and mechanical wear.

[0003] The prior art has the following deficiencies:

[0004] In the prior art, when correcting the parabolic trajectory deviation of the stacker movement speed, the cumulative inertial error may lead to the failure of trajectory correction. During the trajectory correction process, the system usually relies on acceleration sensors, encoders or vision recognition systems to detect and adjust the movement trajectory of the stacker. However, when the stacker starts and stops frequently, changes direction rapidly or operates for a long time, the inertial effect will cause the trajectory deviation to gradually accumulate. Due to the response lag of the control system or the deviation of the error compensation model, the correction instruction may not be able to accurately adjust the movement trajectory, resulting in trajectory oscillation, overcorrection or even out of control, causing the stacker to deviate from the predetermined path. In severe cases, the equipment may collide with the shelves, the goods may fall or the machinery may be damaged, affecting the safety and stability of the automated warehousing system.

[0005] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and therefore it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The object of the present invention is to provide a comprehensive method for correcting the parabolic trajectory deviation of the stacker crane's movement speed. Through multi-sensor fusion technology and the Kalman filtering algorithm, high-precision trajectory deviation detection is achieved, and fuzzy adaptive PID control is used to optimize the correction process, improving trajectory stability. Based on the inertial error prediction model of the LSTM neural network, trajectory deviation is compensated in advance to avoid the correction failure caused by the accumulation of inertial errors, and an online learning mechanism is used to adapt to different working conditions, improving long-term stability. In addition, double closed-loop control and an intelligent anomaly detection mechanism are adopted to monitor the correction effect in real time, automatically trigger safety measures, enhance the system's fault recovery ability, and improve the safety, accuracy, and reliability of the stacker crane's operation, so as to solve the problems in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions: A comprehensive method for correcting the parabolic trajectory deviation of the stacker crane's movement speed, including the following steps:

[0008] Using an acceleration sensor, an encoder, and a vision recognition system, collect data on the real-time movement trajectory of the stacker crane, establish a trajectory deviation detection model, extract movement speed, acceleration, and position information, and calculate the deviation amount of the current trajectory from the ideal parabolic trajectory.

[0009] Based on historical operation data and real-time trajectory data, construct an inertial error prediction model, analyze the cumulative impact of inertial effects on the stacker crane's trajectory, combine the Kalman filtering algorithm and neural network optimization methods, predict the trend of inertial errors, and calculate the inertial deviation compensation value.

[0010] Combined with the inertial error prediction results, use the fuzzy control algorithm and the adaptive PID control strategy to calculate the correction parameters, and generate speed adjustment instructions and movement trajectory correction strategies in real time to ensure the smoothness and accuracy of trajectory adjustment.

[0011] Use a closed-loop control system to monitor the actual operation trajectory of the stacker crane in real time, and adjust the motor drive signal according to the trajectory deviation, so that the stacker crane's movement trajectory gradually approaches the ideal parabolic trajectory, and at the same time dynamically adjust the running speed to reduce deviation accumulation.

[0012] Analyze historical trajectory correction data through deep learning algorithms, optimize the inertial error compensation model, enable the correction algorithm to have self-adaptive capabilities, automatically adjust compensation parameters under different operating conditions, and improve the stability and accuracy of trajectory correction.

[0013] When the trajectory correction deviation exceeds the set threshold, trigger the safety protection mechanism, perform deceleration and recalibration operations on the stacker crane, prevent trajectory out-of-control caused by the accumulation of inertial errors, and restore the normal trajectory correction function after the abnormal state is lifted.

[0014] Preferably, the specific steps of trajectory deviation detection are as follows:

[0015] By adopting the multi-sensor fusion technology, the data obtained by the acceleration sensor, encoder and visual recognition system are fused to improve the accuracy of trajectory deviation detection;

[0016] The acceleration sensor is used to detect the instantaneous acceleration and vibration information during the operation of the stacker, the encoder is used to obtain the rotation angle and traveling speed of the stacker wheel set, and the visual recognition system detects the trajectory change in real time through lidar, and combines machine learning algorithms to establish a trajectory prediction model;

[0017] The data fusion method adopts a weighted average fusion strategy and combines Bayesian estimation for noise suppression to improve the robustness of the measurement data, so that the trajectory deviation detection accuracy can reach the sub-millimeter level, reduce the system error in the correction process, and improve the correction accuracy and stability.

[0018] Preferably, the specific steps of inertial error prediction are as follows:

[0019] Use a long short-term memory neural network to train and predict the trajectory data, and combine a dynamic model adjustment strategy to enable the prediction system to adaptively adjust the prediction parameters according to different working conditions;

[0020] The input of the LSTM network includes historical trajectory deviation data, current speed, current acceleration and environmental variables, and the output is the inertial error prediction value. This neural network model uses the data set for multi-round iterative optimization during training, and uses the Adam optimizer to minimize the loss function, thereby improving the accuracy of inertial error prediction;

[0021] By introducing an attention mechanism, focusing on the moments when the inertial error changes violently, the prediction result is more stable and reliable, effectively avoiding misjudgment caused by environmental disturbances, and improving the real-time performance and accuracy of the correction strategy.

[0022] Preferably, the specific steps of inertial error calculation in the inertial error prediction steps are as follows:

[0023] Set the motion state equation of the stacker, and the formula is as follows:

[0024] x t+1 = Ax t + Bu t + w t ,

[0026] y t = Cx t + v t

[0027] , where x t is the system state vector, representing the state of the stacker at time t, x t+1is the state of the stacker at time t+1, u t is the control input, representing the control input at time t, A is the state transition matrix, B is the control matrix, w t is the process noise, y t is the observed value, C is the observation matrix, v t is the measurement noise;

[0028] Calculate the correction value of the prediction error, and the calculation expression is as follows:

[0029]

[0030] , where e t is the prediction error, is the estimated state value;

[0031] Calculate the optimal estimate through the Kalman gain, and the calculation expression is as follows:

[0032] K t =P t C T (CP t C T +R) -1

[0033] , where K t is the Kalman gain, P t is the state covariance matrix, R is the measurement noise covariance matrix, and T is the transpose of the observation matrix;

[0034] Calculate the final inertial error compensation value, and the calculation expression is as follows:

[0035]

[0036] , where is the final inertial error compensation value.

[0037] Preferably, the specific steps for generating the dynamic correction strategy are as follows:

[0038] Adopt the fuzzy adaptive PID control algorithm to enable the correction strategy to dynamically adjust the control parameters and maintain the optimal correction effect under different working conditions;

[0039] Adopt fuzzy logic rules to adaptively adjust the proportional, integral, and derivative parameters of the PID controller. Among them: when the trajectory deviation is large, increase the P parameter to improve the correction response speed; when the trajectory oscillates, adjust the D parameter to suppress overshoot; when there is a long-term error, optimize the I parameter to reduce the steady-state error;

[0040] The input of the fuzzy controller is the current trajectory deviation and the deviation change rate, and the output is the PID parameter adjustment value;

[0041] Effectively reduce the over-compensation or under-compensation problems caused by fixed parameters, improve the adaptive ability of trajectory correction, make the stacker trajectory smoother, and avoid additional oscillations caused by excessive adjustment.

[0042] Preferably, the specific steps for generating the trajectory optimization control strategy by the dynamic correction strategy are as follows:

[0043] Set the error minimization objective function, and the expression is as follows:

[0044]

[0045] , where J is the optimization objective function, z t is the current trajectory position, is the ideal trajectory position, q t is the weight coefficient of the trajectory deviation, which controls the influence of the trajectory error on the objective function, r t is the weight coefficient of the control input, which prevents the system from being unstable due to excessive control amount, M is the total optimization time, representing the total time points in the optimization process;

[0046] Optimize the control input using the gradient descent method, and the expression is as follows:

[0047]

[0048] [[ID=, where u t ′ is the optimized control input, and η is the learning rate;

[0049] Calculate the trajectory adjustment amount and correct the motion trajectory, and the expression is as follows:

[0050] [[ID=k t+1 ′=x t +αu t ′

[0051]

[0052] , where x t+1 is the corrected trajectory adjustment amount, and α is the adjustment coefficient.

[0053] Preferably, the specific steps for feedback control correction are as follows:

[0054] Adopt a double closed-loop control structure, including an outer-loop trajectory control loop and an inner-loop speed control loop, to achieve more accurate trajectory correction;

[0055] The outer-loop control loop calculates the control amount based on the trajectory deviation data and generates a reference speed command to ensure that the stacker moves along the optimal trajectory;

[0056] The inner-loop control loop is responsible for performing speed adjustment. By real-time monitoring the wheel group speed, ground friction and load conditions of the stacker, the motor drive signal is dynamically adjusted;

[0056] The double - closed - loop control system adjusts the motor torque and speed in real - time, keeping the trajectory deviation within the minimum range, effectively suppressing the accumulation of inertial errors, and improving the stability and response speed of trajectory correction.

[0057] Preferably, the specific steps of adaptive error compensation are as follows:

[0058] Adopt an online learning mechanism to continuously optimize the historical trajectory data, and use a convolutional neural network (CNN) to extract the trajectory feature pattern to improve the adaptability of error compensation;

[0059] The input of the CNN network includes the trajectory deviation information of multiple time steps. After multiple convolutional and pooling operations, the time - dependent features of the trajectory error are extracted;

[0060] By introducing a residual connection, the problem of gradient disappearance is avoided, and the stability of network training is improved;

[0061] The online learning mechanism optimizes the model parameters continuously during operation through incremental training, enabling the error compensation model to adapt to different environmental changes and load conditions, and improving the long - term stability of trajectory correction.

[0062] Preferably, the specific steps of adaptive error compensation are as follows: The specific steps of abnormal state handling are as follows:

[0063] Adopt a fuzzy decision - making algorithm, combine the abnormal detection threshold and multi - sensor data fusion to achieve fast - response abnormal detection and processing;

[0064] Set multiple abnormal detection thresholds, including but not limited to the trajectory deviation exceeding the set range, abnormal correction oscillation frequency, and abnormal increase in inertial error. When an abnormal state is detected, calculate the abnormal level, and trigger corresponding protection measures according to different levels, such as deceleration, emergency stop, or recalibration;

[0065] After the abnormality is lifted, automatically restore the trajectory correction function, and record the abnormal data for subsequent optimization to improve the system's fault self - diagnosis ability.

[0066] In the above technical solutions, the technical effects and advantages provided by the present invention:

[0067] Through multi-sensor fusion technology, the present invention realizes high-precision detection of the trajectory deviation of the stacker crane, and combines the Kalman filtering algorithm for error optimization, making the trajectory correction more accurate and reliable. Traditional trajectory correction methods usually rely on single-sensor data and are easily affected by environmental interference, noise, or sensor accuracy limitations, resulting in a decrease in the accuracy of trajectory correction. The present invention combines an acceleration sensor, an encoder, and a vision recognition system to obtain the motion data of the stacker crane from multiple dimensions and improve the accuracy of trajectory deviation detection through data fusion. In addition, the present invention adopts a fuzzy adaptive PID control strategy to dynamically adjust the PID parameters according to the real-time trajectory deviation, enabling the correction process to adapt to different working conditions and avoiding trajectory oscillation problems caused by over-adjustment or correction lag. Compared with traditional methods, the present invention not only improves the accuracy of trajectory correction, enabling the stacker crane to operate more stably on the set trajectory, but also significantly reduces the long-term impact of trajectory error accumulation, ensuring the continuity and consistency of trajectory correction.

[0068] The present invention adopts an inertial error prediction model based on deep learning. By analyzing the historical trajectory data of the stacker crane through a long short-term memory (LSTM) neural network, it predicts the cumulative trend of inertial errors in advance and compensates before trajectory correction, so that inertial errors will not cause the failure of trajectory correction after long-term operation. Traditional trajectory correction methods often compensate after the error occurs, resulting in trajectory correction lag, or even over-compensation or correction oscillation problems. The present invention can predict the development trend of errors in advance and combine an adaptive compensation strategy to optimize the control parameters before correction, making the trajectory adjustment smoother and more accurate. In addition, the present invention adopts an online learning mechanism to continuously optimize the inertial error compensation model during operation, enabling the system to adapt to different operating environments and load conditions and improving the long-term stability of trajectory correction. Compared with traditional methods, the present invention can not only effectively reduce the problem of correction failure caused by inertial error accumulation, but also maintain a high-precision trajectory correction ability under different loads and speed conditions, enabling the stacker crane to maintain a stable operating state in a complex operating environment.

[0069] The present invention adopts a double - closed - loop control structure and adds an abnormal detection mechanism during the trajectory correction process. It can monitor the execution effect of the trajectory correction in real - time and take corresponding safety measures when abnormal situations occur, avoiding equipment damage or safety accidents caused by out - of - control trajectories. Traditional trajectory correction methods often lack intelligent abnormal detection capabilities. Once the correction fails, it may cause the stacker crane to deviate from the normal operating trajectory, and even pose safety risks such as collisions and cargo dropping. The present invention intelligently evaluates the abnormal state through a fuzzy decision - making algorithm and automatically triggers corresponding safety handling measures according to the abnormal level, such as deceleration, suspension of operation, and recalibration, enabling the system to respond quickly and resume normal operation. In addition, the present invention adopts a rule - learning method to optimize the decision - making model by analyzing historical abnormal data, enabling the system to automatically adjust the abnormal handling strategy according to different operating environments and fault types, improving the system's self - recovery ability. Compared with traditional methods, the present invention greatly improves the safety and fault - tolerance ability of trajectory correction, ensuring that the stacker crane can still operate safely and stably even in high - dynamic environments or complex working conditions, and improving the overall reliability and working efficiency of the automated warehousing system. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0071] Figure 1 It is a method flow chart of the comprehensive method for correcting the parabolic trajectory deviation of the movement speed of the stacker crane of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0072] Now, the exemplary embodiments will be described more comprehensively with reference to the accompanying drawings. However, the exemplary embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these exemplary embodiments are provided so that the present disclosure will be more thorough and complete, and the concept of the exemplary embodiments will be fully conveyed to those skilled in the art.

[0073] The present invention provides a comprehensive method for correcting the parabolic trajectory deviation of the movement speed of the stacker crane as shown in Figure 1 and includes the following steps:

[0074] Using an acceleration sensor, an encoder, and a vision recognition system, collect data on the real - time movement trajectory of the stacker crane, establish a trajectory deviation detection model, extract information on movement speed, acceleration, and position, and calculate the deviation amount of the current trajectory from the ideal parabolic trajectory;

[0075] The specific steps of the trajectory deviation detection are as follows:

[0076] By adopting multi-sensor fusion technology, the data obtained from the acceleration sensor, encoder and vision recognition system are fused to improve the accuracy of trajectory deviation detection;

[0077] The acceleration sensor is used to detect the instantaneous acceleration and vibration information during the operation of the stacker crane, the encoder is used to obtain the rotation angle and traveling speed of the stacker crane's wheel set, and the vision recognition system detects the trajectory change in real time through lidar, and combines machine learning algorithms to establish a trajectory prediction model;

[0078] The data fusion method adopts a weighted average fusion strategy and combines Bayesian estimation for noise suppression to improve the robustness of the measurement data, so that the accuracy of trajectory deviation detection can reach the sub-millimeter level, reduce the systematic error in the correction process, and improve the correction accuracy and stability.

[0079] Based on historical operation data and real-time trajectory data, an inertial error prediction model is constructed to analyze the cumulative impact of inertial effects on the stacker crane's trajectory. Combining the Kalman filter algorithm and neural network optimization method, the trend of inertial error is predicted, and the inertial deviation compensation value is calculated;

[0080] The specific steps of inertial error prediction are as follows:

[0081] The long short-term memory (LSTM) neural network is used to train and predict the trajectory data, and combined with a dynamic model adjustment strategy, so that the prediction system can adaptively adjust the prediction parameters according to different working conditions;

[0082] The input of the LSTM network includes historical trajectory deviation data, current speed, current acceleration and environmental variables, and the output is the inertial error prediction value. This neural network model uses the data set for multi-round iterative optimization during training, and uses the Adam optimizer to minimize the loss function, thereby improving the accuracy of inertial error prediction;

[0083] By introducing the attention mechanism, focus on the moments when the inertial error changes violently, make the prediction results more stable and reliable, effectively avoid misjudgment caused by environmental disturbances, and improve the real-time performance and accuracy of the correction strategy.

[0084] The specific steps of inertial error calculation in the inertial error prediction steps are as follows:

[0085] Set the motion state equation of the stacker crane, and the formula is as follows:

[0086] x t+1 = Ax t + Bu t + w t ,

[0088] yt = Cx t + v t

[0089] , where x t is the system state vector, representing the state of the stacker crane at time t, and x t+1 is the state of the stacker crane at time t + 1, and u t is the control input, representing the control input at time t. A is the state transition matrix, which is used to describe the relationship of the system state evolution over time, and B is the control matrix, reflecting the influence of the control input u t on the system state, and w t is the process noise, representing the random error during the system operation, and y t is the observation value, which is the stacker crane trajectory data measured by the sensor. C is the observation matrix, reflecting the relationship between the sensor measurement value and the system state, and v t is the measurement noise;

[0090] Calculate the correction value of the prediction error, and the calculation expression is as follows:

[0091]

[0092] , where e t is the prediction error, is the estimated state value;

[0093] Calculate the optimal estimate through the Kalman gain, and the calculation expression is as follows:

[0094] K t = P t C T (CP t C T + R) -1

[0095] , where K t is the Kalman gain, and P t is the state covariance matrix, representing the uncertainty of the state estimate. The smaller it is, the higher the estimation accuracy. R is the measurement noise covariance matrix, and T is the transpose of the observation matrix;

[0096] Calculate the final inertial error compensation value, and the calculation expression is as follows:

[0097]

[0098] , where is the final inertial error compensation value.

[0099] Combined with the prediction results of inertial errors, a fuzzy control algorithm and an adaptive PID control strategy are adopted to calculate the correction parameters, and speed adjustment instructions and motion trajectory correction strategies are generated in real time to ensure the smoothness and accuracy of trajectory adjustment;

[0100] The specific steps for generating the dynamic correction strategy are as follows:

[0101] The fuzzy adaptive PID control algorithm is adopted to enable the correction strategy to dynamically adjust the control parameters and maintain the optimal correction effect under different working conditions;

[0102] The proportional (P), integral (I), and derivative (D) parameters of the PID controller are adaptively adjusted through fuzzy logic rules. Among them: when the trajectory deviation is large, the P parameter is increased to improve the correction response speed; when the trajectory oscillates, the D parameter is adjusted to suppress overshoot; when there is a long-term error, the I parameter is optimized to reduce the steady-state error;

[0103] The input of the fuzzy controller is the current trajectory deviation and the deviation change rate, and the output is the PID parameter adjustment value;

[0104] Effectively reduce the over-compensation or under-compensation problems caused by fixed parameters, improve the adaptive ability of trajectory correction, make the stacker trajectory smoother, and avoid additional oscillations caused by over-adjustment.

[0105] The specific steps for generating the trajectory optimization control strategy by the dynamic correction strategy are as follows:

[0106] Set the error minimization objective function, and the expression is as follows:

[0107]

[0108] , where J is the optimization objective function, z t is the current trajectory position, is the ideal trajectory position, q t is the weight coefficient of the trajectory deviation, which controls the influence of the trajectory error on the objective function, r t is the weight coefficient of the control input, which prevents the control quantity from being too large and causing system instability, M is the total optimization time, representing the total time points in the optimization process;

[0109] The gradient descent method is adopted to optimize the control input, and the expression is as follows:

[0110]

[0111] , where u t ′ is the optimized control input, and η is the learning rate;

[0112] Calculate the trajectory adjustment amount and correct the motion trajectory, and the expression is as follows:

[0113] k t+1 ′ = x t + αu t ′

[0114] , where x t+1 is the corrected trajectory adjustment amount, and α is the adjustment coefficient to ensure the smoothness of the trajectory adjustment.

[0115] The actual operating trajectory of the stacker is monitored in real time by using a closed-loop control system, and the motor drive signal is adjusted according to the trajectory deviation, so that the movement trajectory of the stacker gradually approaches the ideal parabolic trajectory, and the operating speed is dynamically adjusted at the same time to reduce the deviation accumulation;

[0116] The specific steps of feedback control correction are as follows:

[0117] A dual closed-loop control structure is adopted, including an outer-loop trajectory control loop and an inner-loop speed control loop, to achieve more accurate trajectory correction;

[0118] The outer-loop control loop calculates the control quantity based on the trajectory deviation data and generates a reference speed command to ensure that the stacker moves along the optimal trajectory;

[0119] The inner-loop control loop is responsible for performing speed adjustment. By monitoring the wheel set speed, ground friction and load conditions of the stacker in real time, the motor drive signal is dynamically adjusted;

[0120] The dual closed-loop control system keeps the trajectory deviation within the minimum range by adjusting the motor torque and speed in real time, effectively suppressing the accumulation of inertial errors and improving the stability and response speed of trajectory correction.

[0121] By analyzing the historical trajectory correction data through a deep learning algorithm, the inertial error compensation model is optimized, so that the correction algorithm has the ability of self-adaptation, automatically adjusts the compensation parameters under different operating conditions, and improves the stability and accuracy of trajectory correction;

[0122] The specific steps of adaptive error compensation are as follows:

[0123] An online learning mechanism is adopted to continuously optimize the historical trajectory data, and a convolutional neural network (CNN) is used to extract the trajectory feature patterns to improve the self-adaptability of error compensation;

[0124] The input of the CNN network includes the trajectory deviation information of multiple time steps. After multiple convolutional and pooling operations, the time-dependent features of the trajectory error are extracted;

[0125] By introducing a residual connection, the problem of gradient disappearance is avoided, and the stability of network training is improved;

[0126] The online learning mechanism optimizes the model parameters during operation through incremental training, enabling the error compensation model to adapt to different environmental changes and load conditions, and improving the long-term stability of trajectory correction.

[0127] When the trajectory correction deviation exceeds the set threshold, the safety protection mechanism is triggered to perform deceleration and recalibration operations on the stacker, preventing trajectory out-of-control caused by the accumulation of inertial errors, and restoring the normal trajectory correction function after the abnormal state is lifted;

[0128] The specific steps for handling abnormal states are as follows:

[0129] Adopt a fuzzy decision-making algorithm, combine the anomaly detection threshold and multi-sensor data fusion to achieve fast-response anomaly detection and processing;

[0130] Set multiple anomaly detection thresholds, including but not limited to the trajectory deviation exceeding the set range, abnormal correction oscillation frequency, and abnormal increase in inertial error. When an abnormal state is detected, calculate the anomaly level and trigger corresponding protection measures according to different levels, such as deceleration, emergency stop, or recalibration;

[0131] After the anomaly is lifted, automatically restore the trajectory correction function and record the anomaly data for subsequent optimization to improve the system's fault self-diagnosis ability.

[0132] Embodiment 1: This embodiment adopts multi-sensor fusion technology to accurately detect the trajectory deviation generated during the movement of the stacker, and combines an adaptive control strategy for real-time correction to improve the operation accuracy and stability of the automated warehousing system. When the stacker operates in a complex environment, its trajectory may be affected by various factors, such as uneven load distribution, ground friction change, motor drive error, and inertial effect. These factors may cause the trajectory to deviate from the set parabolic trajectory, thereby affecting the cargo access efficiency and equipment safety. To solve these problems, this solution adopts a multi-sensor fusion method to collect the movement data of the stacker from multiple dimensions and uses advanced data processing algorithms for trajectory deviation detection and correction.

[0133] First, during the operation trajectory monitoring of the stacker crane, a combination of an acceleration sensor, an encoder, and a vision recognition system is adopted to obtain real-time motion state data. The acceleration sensor is mainly used to detect the instantaneous acceleration and vibration information of the stacker crane, which can reflect the dynamic response characteristics of the stacker crane under different motion states; the encoder is installed on the wheel set or drive shaft of the stacker crane and is used to measure the rotation angle, rotation speed, and linear velocity of the wheel set, so as to infer the overall motion of the stacker crane; the vision recognition system is based on lidar or camera technology, and by identifying ground markings, shelf edges, or other environmental feature points, it obtains the actual position and trajectory information of the stacker crane. The data acquisition frequencies of the above sensors are high and can complement each other, providing multi-dimensional motion information to improve the accuracy of trajectory detection.

[0134] To enhance the reliability and stability of the data, this solution uses the Kalman filter algorithm to fuse and process multi-sensor data, eliminates the random noise in the sensor data, and performs high-precision calculations on the trajectory deviation. The Kalman filter continuously optimizes the motion state estimation of the stacker crane through two processes: state prediction and measurement update, so as to achieve precise calculation of the trajectory deviation. Specifically, the system first establishes a trajectory prediction equation based on the historical motion data and physical model of the stacker crane, and uses the current sensor data for state update to calculate the optimal trajectory estimation value. Once the system detects that the trajectory of the stacker crane deviates from the preset parabolic trajectory, the trajectory correction algorithm is triggered to ensure that the stacker crane runs along the ideal path.

[0135] During the trajectory correction process, a fuzzy adaptive PID control strategy is adopted. By dynamically adjusting the proportional (P), integral (I), and derivative (D) parameters, the trajectory correction is made more precise and stable. This control strategy first determines the P parameter based on the magnitude of the trajectory deviation. When the deviation is large, the P value is increased to accelerate the correction response speed; when the trajectory deviation oscillates, the D parameter is optimized to suppress overshoot; for the case of long-term deviation accumulation, the I parameter is appropriately adjusted to reduce the steady-state error. The fuzzy controller will adjust the PID parameters in real time according to the change trend of the trajectory deviation and the correction effect, so that the stacker crane can maintain the best trajectory correction effect under different working conditions. In addition, when performing trajectory correction, the system will adapt different correction strategies according to the current environmental state. For example, a progressive correction strategy is adopted during low-speed operation, and a mandatory correction strategy is adopted during high-speed operation or when the inertia is large, to ensure the stability and safety of the correction process.

[0136] In summary, this embodiment realizes high-precision detection and correction of the trajectory deviation of the stacker crane through multi-sensor fusion, Kalman filter trajectory prediction, and fuzzy adaptive PID control, which can effectively reduce the influence of inertial error on trajectory control and improve the accuracy and stability of trajectory correction.

[0137] Embodiment 2: This embodiment uses deep learning technology to accurately predict the inertial error generated during the movement of the stacker crane, and combines an adaptive compensation mechanism to optimize the trajectory correction process. Since the inertial error will gradually accumulate after long-term operation, making it difficult for the correction instructions to effectively adjust the trajectory, this solution proposes an inertial error prediction method based on a long short-term memory (LSTM) neural network to calculate the error compensation value in advance and improve the accuracy and real-time performance of trajectory correction.

[0138] First, the system trains the LSTM neural network using historical trajectory data and real-time sensor data, and extracts the change pattern of the inertial error. The LSTM network is a deep learning model suitable for processing time series data, which can effectively learn the time correlation of the stacker crane trajectory error. The inputs of the network include the current trajectory deviation, speed, acceleration, load information, etc., and the output is the predicted inertial error value. To improve the prediction accuracy, the system adopts a supervised learning method, conducts multiple rounds of training through historical error data, and uses the Adam optimizer to minimize the loss function, enabling the network to provide accurate error predictions under different operating conditions.

[0139] Based on the prediction results of the LSTM network, the system adopts an adaptive compensation strategy to perform error pre-compensation before trajectory correction to prevent the correction failure caused by the accumulation of inertial error. Specifically, when the LSTM predicts that the trajectory deviation is about to exceed the safety range, the system will automatically adjust the control parameters, such as increasing the correction amplitude, shortening the correction interval, or optimizing the PID control parameters, so that the correction instructions can more accurately adjust the movement trajectory. In addition, to enhance the adaptive ability of the compensation strategy, the system introduces an online learning mechanism to continuously optimize the LSTM network parameters through incremental training, enabling it to dynamically adjust according to the latest trajectory data and improve the generalization ability of the model.

[0140] Through this embodiment, the system can predict the inertial error in advance and perform compensation before trajectory correction, reduce trajectory oscillation and over-compensation problems, and improve the stability and accuracy of trajectory adjustment.

[0141] Embodiment 3: This embodiment uses a double closed-loop control system to adjust the movement trajectory of the stacker crane in real time, and combines an intelligent anomaly detection mechanism to ensure the safety and stability of trajectory correction. During the trajectory adjustment process, the outer-loop trajectory control loop calculates the correction amount based on the trajectory deviation and generates a speed adjustment instruction; the inner-loop speed control loop is responsible for executing the correction instruction and dynamically adjusting the motor drive signal to improve the accuracy and response speed of the correction.

[0142] During actual operation, various abnormal situations may occur during the trajectory correction process, such as abnormal trajectory oscillation frequency, excessive inertial error, correction failure, etc. This solution uses a fuzzy decision-making algorithm to detect abnormal states and takes corresponding safety measures according to the abnormal levels. For example, when the trajectory correction fails, the system will reduce the speed of the stacker and pause the operation for recalibration if necessary. In addition, the system also adopts an abnormal handling method based on rule learning, optimizing the decision-making model through historical abnormal data, enabling the system to automatically adapt to different operating environments and improving the reliability of trajectory correction.

[0143] Through this embodiment, the stacker can maintain precise trajectory control under different operating conditions and respond quickly in case of abnormalities, improving the safety and stability of the automated warehousing system.

[0144] The present invention realizes high-precision detection of the trajectory deviation of the stacker through multi-sensor fusion technology and combines the Kalman filtering algorithm for error optimization, making the trajectory correction more accurate and reliable. Traditional trajectory correction methods usually rely on single-sensor data and are easily affected by environmental interference, noise, or sensor accuracy limitations, resulting in a decrease in trajectory correction accuracy. The present invention combines an acceleration sensor, an encoder, and a vision recognition system to obtain the motion data of the stacker from multiple dimensions and improves the accuracy of trajectory deviation detection through data fusion. In addition, the present invention adopts a fuzzy adaptive PID control strategy to dynamically adjust the PID parameters according to the real-time trajectory deviation, enabling the correction process to adapt to different working conditions and avoiding trajectory oscillation problems caused by over-adjustment or correction lag. Compared with traditional methods, the present invention not only improves the accuracy of trajectory correction, enabling the stacker to operate more stably on the set trajectory, but also significantly reduces the long-term impact of trajectory error accumulation, ensuring the continuity and consistency of trajectory correction.

[0145] The present invention adopts an inertial error prediction model based on deep learning, analyzes the historical trajectory data of the stacker through a long short-term memory (LSTM) neural network, predicts in advance the cumulative trend of inertial errors, and compensates before trajectory correction, so that inertial errors will not cause the failure of trajectory correction after long-term operation. Traditional trajectory correction methods often compensate after the error occurs, resulting in lagging trajectory correction, and even problems such as overcompensation or correction oscillation. The present invention can predict the error development trend in advance, and combine with an adaptive compensation strategy to optimize the control parameters before correction, making the trajectory adjustment smoother and more accurate. In addition, the present invention adopts an online learning mechanism to continuously optimize the inertial error compensation model during operation, enabling the system to adapt to different operating environments and load conditions, and improving the long-term stability of trajectory correction. Compared with traditional methods, the present invention can not only effectively reduce the problem of correction failure caused by the accumulation of inertial errors, but also maintain a high-precision trajectory correction ability under different loads and speed conditions, enabling the stacker to maintain a stable operating state in a complex operating environment.

[0146] The present invention adopts a double-closed-loop control structure and adds an abnormal detection mechanism during the trajectory correction process, which can monitor the execution effect of the trajectory correction in real time and take corresponding safety measures in case of abnormal situations to avoid equipment damage or safety accidents caused by out-of-control trajectories. Traditional trajectory correction methods often lack the ability of intelligent abnormal detection. Once the correction fails, it may cause the stacker to deviate from the normal operating trajectory, and even pose safety risks such as collisions and cargo drops. The present invention intelligently evaluates the abnormal state through a fuzzy decision-making algorithm and automatically triggers corresponding safety handling measures according to the abnormal level, such as deceleration, suspension of operation, recalibration, etc., enabling the system to respond quickly and resume normal operation. In addition, the present invention adopts a rule learning method to optimize the decision-making model by analyzing historical abnormal data, enabling the system to automatically adjust the abnormal handling strategy according to different operating environments and fault types, and improving the autonomous recovery ability of the system. Compared with traditional methods, the present invention greatly improves the safety and fault tolerance of trajectory correction, ensures that the stacker can operate safely and stably even in a high-dynamic environment or complex working conditions, and improves the overall reliability and working efficiency of the automated warehousing system.

[0147] The above formulas are all dimensionless and take their numerical values for calculation. The formula is obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0148] Only some exemplary embodiments of the present invention have been described by way of illustration. Undoubtedly, for those of ordinary skill in the art, without departing from the spirit and scope of the present invention, the described embodiments can be modified in various different ways. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of the claims of the present invention.

[0149] It should be noted that in this text, if there are relational terms such as first and second, they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the element.

[0150] It should be understood that in various embodiments of the present application, the magnitude of the sequence numbers of the above processes does not mean the order of execution is prior or subsequent. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0151] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this text can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0152] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0153] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0154] In addition, in each embodiment of the present application, each functional unit may be integrated into one processing unit, may exist separately as individual physical units, or two or more units may be integrated into one unit.

[0155] As described above, the foregoing is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

[0156] Only some exemplary embodiments of the present invention have been described above by way of illustration. Undoubtedly, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

Claims

1. Comprehensive method for correcting the parabolic trajectory deviation of the stacker's movement speed, characterized in that, The steps include the following: Utilize an acceleration sensor, an encoder, and a vision recognition system to collect data on the real-time motion trajectory of the stacker crane, establish a trajectory deviation detection model, extract motion speed, acceleration, and position information, and calculate the deviation amount of the current trajectory from the ideal parabolic trajectory; Based on historical operation data and real-time trajectory data, construct an inertial error prediction model, analyze the cumulative impact of the inertial effect on the stacker crane's trajectory, combine the Kalman filter algorithm and neural network optimization method, predict the trend of inertial error, and calculate the inertial deviation compensation value; Combined with the inertial error prediction results, adopt a fuzzy control algorithm and an adaptive PID control strategy to calculate the correction parameters, and generate speed adjustment instructions and motion trajectory correction strategies in real time to ensure the smoothness and accuracy of trajectory adjustment; Use a closed-loop control system to monitor the actual operation trajectory of the stacker crane in real time, and adjust the motor drive signal according to the trajectory deviation, so that the stacker crane's motion trajectory gradually approaches the ideal parabolic trajectory, and at the same time dynamically adjust the running speed to reduce the accumulation of deviation; Analyze the historical trajectory correction data through a deep learning algorithm to optimize the inertial error compensation model, enable the correction algorithm to have an adaptive ability, automatically adjust the compensation parameters under different operating conditions, and improve the stability and accuracy of trajectory correction; When the trajectory correction deviation exceeds the set threshold, trigger the safety protection mechanism, perform deceleration and recalibration operations on the stacker crane to prevent trajectory out-of-control caused by the accumulation of inertial error, and restore the normal trajectory correction function after the abnormal state is lifted.

2. The comprehensive method for correcting the parabolic trajectory deviation of the stacker moving speed according to claim 1, characterized in that, The specific steps of trajectory deviation detection are as follows: Adopt multi-sensor fusion technology to fuse the data obtained by the acceleration sensor, encoder, and vision recognition system to improve the accuracy of trajectory deviation detection; The acceleration sensor is used to detect the instantaneous acceleration and vibration information during the operation of the stacker crane, the encoder is used to obtain the rotation angle and traveling speed of the stacker crane's wheel set, and the vision recognition system detects the trajectory change in real time through lidar, and combines machine learning algorithms to establish a trajectory prediction model; The data fusion method adopts a weighted average fusion strategy and combines Bayesian estimation for noise suppression to improve the robustness of the measurement data, so that the accuracy of trajectory deviation detection can reach the sub-millimeter level, reduce the systematic error in the correction process, and improve the correction accuracy and stability.

3. The comprehensive method for correcting the parabolic trajectory deviation of the stacker movement speed according to claim 1, characterized in that The specific steps of inertial error prediction are as follows: Adopt a long short-term memory neural network to train and predict the trajectory data, and combine a dynamic model adjustment strategy to enable the prediction system to adaptively adjust the prediction parameters according to different working conditions; The input of the LSTM network includes historical trajectory deviation data, current speed, current acceleration, and environmental variables, and the output is the inertial error prediction value. This neural network model is optimized through multiple rounds of iteration using a dataset during training, and the Adam optimizer is used to minimize the loss function, thereby improving the accuracy of inertial error prediction; By introducing an attention mechanism, focus on the moments when the inertial error changes violently, make the prediction results more stable and reliable, effectively avoid misjudgment caused by environmental disturbances, and improve the real-time performance and accuracy of the correction strategy.

4. The comprehensive method for correcting the parabolic trajectory deviation of the stacker crane movement speed according to claim 1, characterized in that The specific steps for calculating the inertial error in the inertial error prediction step are as follows: Set the motion state equation of the stacker crane, and the formula is as follows: x t+1 = Ax t + Bu t + w t , y t = Cx t + v t , where x t is the system state vector, representing the state of the stacker crane at time t, and x t+1 is the state of the stacker crane at time t + 1, u t is the control input, representing the control input at time t, A is the state transition matrix, B is the control matrix, w t is the process noise, y t is the observation value, C is the observation matrix, v t is the measurement noise; Calculate the correction value of the prediction error, and the calculation expression is as follows: ， where, e t is the prediction error, is the estimated state value; Calculate the optimal estimate through the Kalman gain, and the calculation expression is as follows: K t = P t C T (CP t C T + R) -1 where K t is the Kalman gain, P t is the state covariance matrix, R is the measurement noise covariance matrix, and T is the transpose of the observation matrix; Calculate the final inertial error compensation value, and the calculation expression is as follows: ， In the formula, is the final inertial error compensation value.

5. The comprehensive method for correcting the parabolic trajectory deviation of the stacker moving speed according to claim 1, characterized in that, The specific steps for generating the dynamic correction strategy are as follows: Adopt the fuzzy adaptive PID control algorithm to enable the correction strategy to dynamically adjust the control parameters and maintain the optimal correction effect under different working conditions; Adopt fuzzy logic rules to adaptively adjust the proportional, integral, and derivative parameters of the PID controller. Among them: when the trajectory deviation is large, increase the P parameter to improve the correction response speed; when the trajectory oscillates, adjust the D parameter to suppress overshoot; when there is a long-term error, optimize the I parameter to reduce the steady-state error; The input of the fuzzy controller is the current trajectory deviation and the deviation change rate, and the output is the PID parameter adjustment value; Effectively reduce the over-compensation or under-compensation problems caused by fixed parameters, improve the adaptive ability of trajectory correction, make the stacker crane trajectory smoother, and avoid additional oscillations caused by excessive adjustment.

6. The comprehensive method for correcting the deviation of the parabolic trajectory of the stacking machine movement speed according to claim 1, characterized in that The specific steps for generating the trajectory optimization control strategy of the dynamic correction strategy are as follows: Set the error minimization objective function, and the expression is as follows: ， where J is the optimization objective function, z t is the current trajectory position, is the ideal trajectory position, q t is the weight coefficient of the trajectory deviation, controlling the influence of the trajectory error on the objective function, r t is the weight coefficient of the control input, preventing the control quantity from being too large and causing system instability, M is the total optimization time, representing the total time points in the optimization process; Adopt the gradient descent method to optimize the control input, and the expression is as follows: ， where \(u\) t ′ is the optimized control input and \(\eta\) is the learning rate; Calculate the trajectory adjustment amount and correct the motion trajectory, and the expression is as follows: k t+1 ′ = x t + αu t ′ , where x t+1 is the adjusted trajectory correction amount, and α is the adjustment coefficient.

7. The comprehensive method for correcting the parabolic trajectory deviation of the stacker movement speed according to claim 1, characterized in that The specific steps for feedback control correction are as follows: Adopt a double closed-loop control structure, including an outer-loop trajectory control loop and an inner-loop speed control loop, to achieve more accurate trajectory correction; The outer-loop control loop calculates the control amount based on the trajectory deviation data and generates a reference speed command to ensure that the stacker crane moves along the optimal trajectory; The inner-loop control loop is responsible for performing speed adjustment. By real-time monitoring the wheel group speed, ground friction, and load conditions of the stacker crane, the motor drive signal is dynamically adjusted; The double closed-loop control system keeps the trajectory deviation within the minimum range by real-time adjusting the motor torque and speed, effectively suppressing the accumulation of inertial errors, and improving the stability and response speed of trajectory correction.

8. The comprehensive method for correcting the parabolic trajectory deviation of the stacker movement speed according to claim 1, wherein, The specific steps for adaptive error compensation are as follows: Adopt an online learning mechanism to continuously optimize the historical trajectory data, and use a convolutional neural network (CNN) to extract the trajectory feature pattern to improve the adaptability of error compensation; The input of the CNN network includes the trajectory deviation information of multiple time steps. After multiple convolutional and pooling operations, the time-dependent features of the trajectory error are extracted; By introducing a residual connection, the problem of gradient disappearance is avoided, and the network training stability is improved; The online learning mechanism optimizes the model parameters during the operation through incremental training, enabling the error compensation model to adapt to different environmental changes and load conditions, and improving the long-term stability of trajectory correction.

9. The comprehensive method for correcting the parabolic trajectory deviation of the stacker moving speed according to claim 1, characterized in that The specific steps for adaptive error compensation are as follows: The specific steps for abnormal state handling are as follows: Adopt a fuzzy decision-making algorithm, combined with the abnormal detection threshold and multi-sensor data fusion, to achieve fast-response abnormal detection and processing; Set multiple anomaly detection thresholds, including but not limited to trajectory deviation exceeding the set range, abnormal correction oscillation frequency, and abnormal increase in inertial error. When an abnormal state is detected, calculate the anomaly level and trigger corresponding protection measures according to different levels, such as deceleration, emergency stop, or recalibration; After the anomaly is lifted, automatically resume the trajectory correction function and record the anomaly data for subsequent optimization to improve the system's fault self-diagnosis ability.

Images