A posture control system for pallet handling collaborative scheduling

Abstract

This invention belongs to the field of commercial data processing technology, specifically disclosing an attitude control system for collaborative scheduling of material frame handling. The system includes a control body, which is implemented as follows: Position coordinate data and velocity vector data are collected from the composite motion state of the material frame via a sensor network. The composite motion state includes a superposition of translation and rotation. A particle filter algorithm is used to filter noise from the position coordinate data and velocity vector data. The input to the particle filter algorithm is the position coordinate data and velocity vector data, and the output is the filtered position coordinates and the filtered velocity vector. Based on the filtered position coordinates and the filtered velocity vector, the instantaneous displacement vector of the material frame in space is determined, and the product of the instantaneous displacement vector and a preset time interval is obtained. The purpose of this invention is to solve the problems of difficulty in predicting collision risks and inaccurate real-time attitude adjustment during handling in complex environments.

Description

Technical Field

[0001] This invention relates to the field of commercial data processing technology, specifically to a posture control system for collaborative scheduling of material handling. Background Technology

[0002] In modern intelligent manufacturing and automated logistics, the material handling and scheduling system plays a crucial role in connecting various links of the production line. Its operational efficiency and safety directly affect the stability and economic benefits of the entire production process. As factory automation levels continue to rise, material handling tasks are becoming increasingly complex, placing higher demands on the system's precise control and intelligent decision-making capabilities. Current material handling systems generally suffer from insufficient adaptability when dealing with complex spatial environments, especially in scenarios involving dense shelving layouts and multi-device collaborative operations. Existing methods often rely on preset fixed paths and attitude parameters, lacking real-time response capabilities to dynamic environmental changes. This rigid control approach is prone to efficiency losses when encountering changes in spatial constraints, failing to fully utilize the equipment's motion potential. The core technical challenge in this field lies primarily in the accuracy of spatial interference prediction for material frames under complex motion states. During handling, material frames need to perform multiple complex actions such as rotation, translation, and tilting. These combinations of actions generate complex spatial occupancy trajectories, which traditional collision detection methods struggle to accurately capture. More critically, due to the lack of systematic accumulation and analysis of historical collision risk data, the system cannot form an effective safety assessment mechanism. For example, when a robotic arm needs to precisely place a material frame within narrow gaps in a shelf, if the minimum distance between each face of the frame and the shelf uprights during movement cannot be accurately predicted, it may lead to collisions or overly conservative safety margin settings, severely impacting operational efficiency. Therefore, establishing a real-time detection mechanism capable of accurately tracking spatial coordinate changes in the frame's complex motion state, and constructing a reliable collision risk assessment system through extensive simulation data, becomes a key issue in achieving collaborative scheduling and attitude control management for material frame handling. Summary of the Invention

[0003] This invention provides an attitude control system for collaborative scheduling of material frame handling, aiming to solve the problems of difficulty in predicting collision risks and inaccurate real-time attitude adjustment during handling in complex environments.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0005] A posture control system for coordinated scheduling of material frame handling includes a control system body. The control system operates through the following method: acquiring position coordinate data and velocity vector data from the composite motion state of the material frame via a sensor network, wherein the composite motion state includes a superposition of translation and rotation; filtering the position coordinate data and velocity vector data for noise using a particle filter algorithm, wherein the input of the particle filter algorithm is the position coordinate data and velocity vector data, and the output is the filtered position coordinates and the filtered velocity vector; determining the instantaneous displacement vector of the material frame in space based on the filtered position coordinates and the filtered velocity vector; and obtaining the instantaneous displacement vector and a preset time interval. The product of the values ​​is used to obtain the displacement increment of the material frame. If the displacement increment exceeds a preset threshold, it is judged as an abnormal motion event. For the abnormal motion event, additional acceleration data is collected from the sensor network to obtain an abnormal acceleration sequence. By integrating the abnormal acceleration sequence with the filtered velocity vector, the abnormal acceleration sequence is accumulated on the time axis to the filtered velocity vector to calculate the displacement change, and the predicted trajectory offset of the material frame is determined. The difference between the predicted trajectory offset and the real-time trajectory change derived from the filtered position coordinates is obtained to obtain the trajectory correction factor. The filtered position coordinates are adjusted according to the trajectory correction factor, and the continuity of the adjusted position coordinates is judged to obtain the real-time spatial trajectory change of the material frame.

[0006] In one aspect of this disclosure, the real-time spatial trajectory change of the material frame includes:

[0007] By collecting real-time trajectory changes of the material frame through sensors and combining them with a preset environmental geometric model, the Euclidean distance from the coordinate points of each surface of the material frame to the surface points of surrounding obstacles is calculated, and the distance matrix is ​​obtained.

[0008] For the distance matrix, if the minimum value is lower than a preset threshold, it is marked as a potential interference region, and the initial interference point in the material frame movement process is obtained;

[0009] Based on the initial interference point, the relative position of the material frame velocity vector and the obstacle is obtained. By dividing the velocity vector by the relative position distance, the velocity adjustment coefficient is determined, and the adjusted trajectory change is obtained.

[0010] Using the adjusted trajectory changes, the distance matrix is ​​recalculated and potential interference regions are marked to obtain a set of spatial interference risk points.

[0011] In one aspect of this disclosure, the set of spatial interference risk points includes:

[0012] Acquire spatial position data during the movement of the material frame, calculate position coordinate difference and velocity vector from the spatial position data, and extract motion trajectory features;

[0013] The motion trajectory features are queried from the historical database, and the trajectory similarity is calculated using Euclidean distance. Collision event records under similar trajectories are extracted from trajectories with similarity exceeding a preset threshold to determine the historical collision event sequence.

[0014] The support vector machine is used to construct feature vectors from the set of spatial interference risk points and the sequence of historical collision events as input. Classification training is performed through an optimization process that maximizes the classification margin to obtain the classification boundary parameters.

[0015] Based on the classification boundary parameters, the current composite motion state of the material frame is analyzed, and it is determined that if the current state exceeds the classification boundary, the collision probability increases, thus obtaining a preliminary probability value.

[0016] By integrating the initial probability values ​​with motion trajectory features, a weighted average method is used to calculate the probability adjustment and determine the collision probability distribution of the material frame under the current composite motion state.

[0017] In one aspect of this disclosure, the collision probability distribution includes:

[0018] Based on the composite motion state of the material frame, the real-time position coordinates and velocity vectors collected by the sensors are obtained, and the collision probability distribution is calculated using the Monte Carlo simulation method to determine the frequency of potential collision events.

[0019] Based on the collision probability distribution, multiple alternative attitude adjustment sequences are generated, and the influence of ambient lighting factors are incorporated to obtain an extended adjustment sequence;

[0020] A genetic algorithm is used to iteratively optimize the extended adjustment sequence. The input of the genetic algorithm is the individual of the extended adjustment sequence, and the output is the optimized individual after fitness evaluation, so as to obtain the preliminary optimization parameters.

[0021] If the initial optimization parameters deviate from the preset load balancing factor, the initial optimization parameters are adjusted, and the corrected attitude parameters are determined by a weighted fusion method.

[0022] Based on the corrected attitude parameters, the surrounding obstacle avoidance paths are integrated to obtain the optimized set of material frame attitude parameters.

[0023] In one aspect of this disclosure, the optimized set of material frame attitude parameters includes:

[0024] The space occupancy trajectory of the material frame after the attitude is adjusted is simulated and compared with the dynamic environmental change data. If there is no interference in the simulated trajectory, the attitude parameter set is confirmed to be valid, and the final attitude control scheme is obtained.

[0025] In one aspect of this disclosure, the final attitude control scheme includes:

[0026] The attitude control scheme is obtained, the equipment attitude adjustment parameters are extracted from the collaborative scheduling system, the parameters are determined to match the initial state of the material frame handling equipment, and the degree of matching is judged by comparing the difference between the parameter values ​​and the initial state indicators.

[0027] The attitude control scheme is integrated into the collaborative scheduling system. A preset threshold is used to judge the compatibility of the integration process. The compatibility score is calculated from the integration interface data to obtain the integrated scheduling framework.

[0028] For the integrated scheduling framework, control commands are issued to the material handling equipment, and real-time feedback data after the command execution is obtained;

[0029] Based on the real-time feedback data, if the data deviation exceeds the preset threshold, the control command is adjusted, the offset is subtracted from the deviation value, and the command is reissued to determine the stability of the adjusted feedback data.

[0030] From the adjusted feedback data, it is determined that the data meets the risk assessment requirements, and the final verification result is obtained.

[0031] In one aspect of this disclosure, the final verification result includes:

[0032] By verifying the feedback data, records that meet the risk requirements are obtained, resulting in an updated collision risk dataset.

[0033] Based on the collision risk dataset, a cumulative update mechanism is used to process the historical database. This mechanism extracts the cumulative collision risk from the historical data and combines it with the update mechanism to gradually add records that meet the risk requirements, thereby obtaining an enhanced safety assessment model.

[0034] For the enhanced security assessment model, if the parameter deviation exceeds a preset threshold, the threshold is adjusted to adjust the associated attributes. The deviation is obtained by calculating the difference between the enhanced model parameters and the enhanced security assessment, and the enhanced risk assessment system is determined.

[0035] Task planning details are obtained from the enhanced risk assessment system. These details are integrated with system architecture optimization and material handling task planning to obtain an execution feedback loop for subsequent material handling tasks.

[0036] In another aspect of this disclosure, the control system body includes:

[0037] The data acquisition and filtering module is used to acquire position coordinate data and velocity vector data from the composite motion state of the material frame through a sensor network, and to use a particle filter algorithm to filter the noise in the position coordinate data and velocity vector data, and output the filtered position coordinates and filtered velocity vectors.

[0038] The displacement increment calculation module is used to determine the instantaneous displacement vector of the material frame in space based on the filtered position coordinates and the filtered velocity vector, and obtain the product of the instantaneous displacement vector and the preset time interval to obtain the displacement increment of the material frame.

[0039] An abnormal event detection module is used to determine an abnormal motion event if the displacement increment of the material frame exceeds a preset threshold. For the abnormal motion event, additional acceleration data is collected from the sensor network to obtain an abnormal acceleration sequence.

[0040] The trajectory correction factor calculation module is used to determine the predicted trajectory offset of the material frame by integrating the abnormal acceleration sequence and the filtered velocity vector, and to obtain the difference between the predicted trajectory offset and the real-time trajectory change derived from the filtered position coordinates to obtain the trajectory correction factor.

[0041] The trajectory adjustment and continuity judgment module is used to adjust the filtered position coordinates according to the trajectory correction factor, judge the continuity of the adjusted position coordinates, and obtain the real-time spatial trajectory change of the material frame.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] This invention collects material frame motion data through a sensor network, uses a particle filter algorithm for state estimation to obtain a real-time trajectory, combines an environmental geometric model to calculate a distance matrix to identify potential interference regions, utilizes a support vector machine to classify and train risk point sets and historical collision records to determine the collision probability distribution, employs a genetic algorithm to iteratively optimize the attitude adjustment sequence to generate the optimal attitude parameter set, verifies the safety of the adjusted trajectory through simulation calculations, and finally integrates the optimized scheme into collaborative scheduling to achieve intelligent control, continuously strengthening the risk assessment model through a cumulative update mechanism. This invention achieves accurate prediction of collision risks and intelligent optimization and adjustment of attitude during material frame handling, significantly improving the safety and efficiency of handling operations. Attached Figure Description

[0044] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0045] Figure 1 This is a flowchart of an attitude control management method for collaborative scheduling of material frame handling according to the present invention. Detailed Implementation

[0046] The present invention will be further described below with reference to embodiments. These embodiments are merely some, not all, of the embodiments described. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the protection scope of the present invention.

[0047] Please see Figure 1 As shown in the figure, this embodiment discloses a posture control management method for collaborative scheduling of material frame handling, which may specifically include:

[0048] Step S1 involves acquiring position coordinate data and velocity vector data from the composite motion state of the material frame using a sensor network. A particle filter algorithm is then applied to filter noise and estimate the state of the acquired data, yielding the real-time spatial trajectory changes of the material frame. The composite motion state of the material frame includes a superposition of translational and rotational motions. The sensor network, through devices such as LiDAR, inertial measurement units, and vision sensors, acquires the position coordinate data and velocity vector data of the material frame in real time. The particle filter algorithm is a nonlinear filtering method based on Bayesian inference. It uses a large number of random particles to represent the probability distribution of the system state and filters noisy sensor data. The algorithm inputs the original position coordinate data and velocity vector data, and outputs the accurate position coordinates and velocity vectors after noise filtering. This processing method effectively eliminates sensor measurement errors and environmental interference, improving the accuracy of the material frame motion state estimation.

[0049] Step S11: Position coordinate data and velocity vector data are acquired from the composite motion state of the material frame using a sensor network. The sensor network consists of multiple sensor nodes, including an inertial measurement unit (IMU) mounted on the material frame, a LiDAR scanner in the surrounding environment, and a vision camera. The IMU measures the acceleration and angular velocity of the material frame and obtains position and velocity information through integration. The LiDAR emits a laser beam and receives reflected signals to measure the distance between the material frame and surrounding objects, calculating the spatial position of the material frame. The vision camera tracks the motion trajectory of the material frame using image recognition technology.

[0050] For example, in one embodiment, as the material frame moves between warehouse shelves, an inertial measurement unit collects data 100 times per second, a lidar scans the surrounding environment at a frequency of 10 Hz, and a vision camera records the material frame's movement at a rate of 30 frames per second. Acquiring this complex motion requires synchronizing data from multiple sensors, and timestamp alignment ensures data consistency.

[0051] Step S12 involves using a particle filter algorithm to filter noise from the position coordinate data and velocity vector data. The particle filter algorithm generates a large number of random particles to represent the probability distribution of the material frame's state, with each particle representing a possible motion state. The algorithm first predicts the particle's state at the next moment based on the system dynamics model, and then updates the particle weights using sensor observation data.

[0052] Specifically, the prediction step calculates the position and velocity of each particle at the next moment based on the motion equation of the material frame. The update step adjusts the particle weights by comparing the predicted state with the actual observed data; particles with higher weights represent the more likely true state.

[0053] For example, in one possible implementation, the particle filtering algorithm uses 1000 particles to represent the state of the material frame. After multiple iterations, the algorithm converges to the optimal state estimate. The noise filtering process can effectively handle random errors and systematic biases in sensor measurements, outputting filtered position coordinates and velocity vectors.

[0054] Step S13: Determine the instantaneous displacement vector of the material frame in space based on the filtered position coordinates and the filtered velocity vector. The instantaneous displacement vector is calculated by integrating the filtered velocity vector over a preset time interval, reflecting the positional change of the material frame in a short period of time.

[0055] Specifically, the instantaneous displacement vector is equal to the product of the velocity vector and the time interval, which is usually set as the sensor sampling period.

[0056] For example, in one embodiment, the filtered velocity vector of the material frame is 0.5 meters per second, the preset time interval is 0.1 seconds, and the calculated instantaneous displacement vector is 0.05 meters. By continuously calculating the instantaneous displacement vector, the movement trajectory of the material frame can be tracked in real time. This calculation method provides basic data for subsequent anomaly detection and trajectory prediction.

[0057] Step S14: Obtain the product of the instantaneous displacement vector and the preset time interval to get the material frame displacement increment. The material frame displacement increment represents the actual displacement distance of the material frame within the preset time interval, which is calculated using the modulus of the instantaneous displacement vector. The preset time interval is determined based on the characteristics of the material frame motion and the control accuracy requirements, and is usually set between 0.1 seconds and 1 second.

[0058] For example, in one possible implementation, when the material frame travels between shelves, the instantaneous displacement vector has a magnitude of 0.08 meters and a time interval of 0.2 seconds, resulting in a calculated displacement increment of 0.016 square meters per second. The calculation of the displacement increment provides a quantitative indicator for abnormal motion monitoring; when the increment exceeds the normal range, it indicates that the material frame may be experiencing abnormal motion.

[0059] Step S15: If the displacement increment of the material frame exceeds a preset threshold, it is determined to be an abnormal motion event. Additional acceleration data is then collected from the sensor network for this event. The preset threshold is determined based on the statistical characteristics of normal material frame movement and is typically set to 1.5 to 2 times the normal displacement increment. Abnormal motion events may be caused by external interference, equipment malfunction, or path deviation. When an abnormal motion event is detected, the system immediately initiates an additional data acquisition program to obtain more detailed acceleration data from the sensor network.

[0060] For example, in one embodiment, the displacement increment during normal movement of the material frame is approximately 0.01 square meters per second, and the preset threshold is set to 0.02 square meters per second. When the actual increment reaches 0.025 square meters per second, the system determines it to be an abnormal event. The additional acceleration data collected includes triaxial acceleration and angular acceleration, forming an abnormal acceleration sequence to provide data support for subsequent trajectory correction.

[0061] Step S16: The predicted trajectory offset of the material frame is determined by integrating the abnormal acceleration sequence with the filtered velocity vector. The integration operation accumulates the abnormal acceleration sequence onto the filtered velocity vector on the time axis to calculate the displacement change of the material frame under abnormal motion conditions.

[0062] Specifically, the integration process first converts the acceleration data into velocity increments, then superimposes them with the original velocity vector to obtain a corrected velocity vector. The predicted trajectory offset is calculated by further integrating the corrected velocity vector, reflecting the degree to which the material frame deviates from the original trajectory.

[0063] For example, in one possible implementation, the abnormal acceleration sequence shows that the material frame is subjected to a lateral force. After integration, the predicted trajectory offset is 0.15 meters, indicating that the material frame deviates significantly from its original trajectory. This calculation method can predict the movement trend of the material frame under abnormal conditions, providing a basis for trajectory correction.

[0064] Step S17: Obtain the difference between the predicted trajectory offset and the real-time trajectory change derived from the filtered position coordinates to obtain the trajectory correction factor. The real-time trajectory change is obtained through time series analysis of the filtered position coordinates, reflecting the actual movement trajectory of the material frame. The trajectory correction factor is equal to the difference between the predicted trajectory offset and the real-time trajectory change, used to quantify the deviation between the prediction and the actual situation.

[0065] For example, in one embodiment, the predicted trajectory offset is 0.15 meters, the real-time trajectory change is 0.12 meters, and the calculated trajectory correction factor is 0.03 meters. A positive trajectory correction factor indicates that the predicted trajectory offset is greater than the actual change, requiring correction in the opposite direction. A negative trajectory correction factor indicates that the actual change is greater than the predicted offset, requiring stronger correction. This difference calculation can accurately quantify the trajectory deviation, providing accurate parameters for subsequent position correction.

[0066] Step S18: Adjust the filtered position coordinates according to the trajectory correction factor, determine the continuity of the adjusted position coordinates, and obtain the real-time spatial trajectory change of the material frame. Position coordinate adjustment is achieved by applying the trajectory correction factor to the filtered position coordinates, with the adjustment direction opposite to the sign of the correction factor. Continuity is determined by checking the smoothness and consistency of the adjusted position coordinate sequence to ensure that the trajectory change conforms to the laws of physical motion.

[0067] For example, in one possible implementation, the trajectory correction factor is 0.03 meters, the adjusted position coordinates are shifted 0.03 meters to the left, and the continuity check shows that the adjusted trajectory is smooth with no abrupt changes. The real-time spatial trajectory changes of the material frame are output in time series form, including the adjusted position coordinates and the corresponding timestamp. This adjustment and judgment mechanism ensures the accuracy and reliability of the trajectory data, providing high-quality input for subsequent interference detection.

[0068] Step S2 involves calculating the distance matrix between each face of the material frame and surrounding obstacles based on the real-time spatial trajectory changes of the material frame and a preset environmental geometric model. If the minimum value in the distance matrix is ​​lower than a preset threshold, it is marked as a potential interference zone, thus obtaining a set of spatial interference risk points during the material frame's movement. The environmental geometric model is a three-dimensional geometric description of the warehouse or work area, including the position and shape information of obstacles such as shelves, walls, and equipment. The distance matrix is ​​generated by calculating the Euclidean distance from feature points on each face of the material frame to the obstacle surface; the rows of the matrix represent the faces of the material frame, and the columns represent the obstacles. The potential interference zone refers to the spatial area where the material frame is too close to an obstacle, potentially leading to a collision. This method allows for early identification of collision risks, providing a basis for path planning and attitude adjustment.

[0069] Step S21: Real-time trajectory changes of the material frame are collected by sensors, and a distance matrix is ​​calculated based on a pre-set environmental geometric model. The environmental geometric model is stored in the form of a 3D point cloud or mesh, containing geometric information of all fixed and moving obstacles in the warehouse. The model is constructed through a combination of LiDAR scanning and manual modeling, and is updated regularly to reflect environmental changes. The distance matrix calculation process first determines the feature points of each face of the material frame, typically selecting the center point and corner points of the face. Then, the Euclidean distance from each feature point to the nearest point on the obstacle surface is calculated. The Euclidean distance formula is the square root of the sum of the squares of the coordinate differences between the two points.

[0070] For example, in one embodiment, a cuboid frame has 6 faces, with 5 feature points selected on each face, and 3 surrounding shelf obstacles. The generated distance matrix is ​​30 rows and 3 columns. Each element in the distance matrix represents the shortest distance between the corresponding feature point and the obstacle, providing quantitative data for subsequent interference judgment.

[0071] Step S22: For the distance matrix, if the minimum value is lower than a preset threshold, it is marked as a potential interference region, thus obtaining the initial interference points during the material frame's movement. The preset threshold is determined based on the size of the material frame and safety requirements, typically set to 10% to 20% of the maximum size of the material frame. The minimum value search is achieved by traversing all elements of the distance matrix to find the closest distance between the material frame and the obstacle. When the minimum distance is lower than the preset threshold, the corresponding material frame surface and obstacle are marked as potential interference regions.

[0072] For example, in one possible implementation, the minimum value of the distance matrix is ​​0.08 meters, the preset threshold is 0.1 meters, and the system marks the right side of the material frame and shelf A as potential interference areas. Initial interference points are recorded in the form of coordinates and identifiers, including the material frame surface number, obstacle identification, and distance value. This marking method can accurately locate potential collision risk areas.

[0073] Step S23: Based on the initial interference point, obtain the relative position of the material frame velocity vector and the obstacle. Determine the velocity adjustment coefficient by dividing the velocity vector by the relative position distance. The relative position is obtained by calculating the position vectors of the center point of the material frame and the center point of the obstacle, reflecting their spatial relationship. The velocity adjustment coefficient is equal to the modulus of the material frame velocity vector divided by the relative position distance, used to quantify the speed at which the material frame approaches the obstacle.

[0074] For example, in one embodiment, the velocity vector magnitude of the material frame is 0.5 meters per second, and the relative position distance to the shelf is 2 meters. The calculated velocity adjustment coefficient is 0.25 per second. A larger velocity adjustment coefficient indicates that the material frame is rapidly approaching the obstacle, with a higher risk of collision. A smaller coefficient indicates that the material frame is slowly approaching or moving away from the obstacle, with a relatively lower risk. This calculation method allows for a dynamic assessment of the urgency of the collision risk.

[0075] Step S24 involves recalculating the distance matrix and marking potential interference areas using the adjusted trajectory change, thus obtaining the final set of spatial interference risk points. The adjusted trajectory change is obtained by correcting the original trajectory using a velocity adjustment coefficient, typically by reducing the velocity component in the direction approaching the obstacle. The recalculated distance matrix, based on the adjusted material frame position and trajectory, reflects the corrected spatial relationships.

[0076] For example, in one possible implementation, after speed adjustment, the approach speed of the material frame is reduced by 30%, and the recalculated distance matrix shows that the minimum distance has increased to 0.12 meters, exceeding a preset threshold. The final spatial interference risk point set includes all areas that still pose an interference risk after adjustment, as well as newly emerging potential risk points. This iterative calculation and adjustment mechanism can dynamically optimize the trajectory of the material frame, minimizing the risk of collision.

[0077] Step S3 involves acquiring the set of spatial interference risk points during the material frame's movement and extracting collision event records under similar trajectories from the historical database. A support vector machine (SVM) is then used to classify and train the point set and historical records to determine the collision probability distribution under the current composite motion state of the material frame. The historical database stores a large amount of trajectory data and collision event records for material frame handling tasks, providing training samples for machine learning. The SVM is a supervised learning algorithm that classifies data into different categories by finding the optimal classification boundary. The collision probability distribution describes the likelihood of a collision occurring in different motion states of the material frame, providing a quantitative basis for risk assessment and decision-making. This method allows for the prediction of collision risks under current conditions using historical experience.

[0078] Step S31: Acquire spatial position data during the movement of the material frame, calculate the position coordinate difference and velocity vector, and extract motion trajectory features. The spatial position data originates from the real-time trajectory changes in step S1, including the position coordinates of the material frame over time. The position coordinate difference is obtained by calculating the difference between position coordinates at adjacent moments, reflecting the instantaneous displacement of the material frame. The velocity vector is calculated by dividing the position coordinate difference by the time interval, representing the velocity and direction of the material frame's movement. Motion trajectory features include statistical characteristics such as trajectory curvature, rate of change of velocity, and direction of movement.

[0079] For example, in one embodiment, the material frame moves along a curved path, and the trajectory features extracted show an average curvature of 0.1 m / s and a velocity change rate of 0.05 m / s². These features provide a quantitative description for subsequent similarity matching.

[0080] Step S32: Query motion trajectory features from the historical database, calculate trajectory similarity using Euclidean distance, and extract collision event records for similar trajectories. The historical database is indexed according to trajectory features, supporting fast querying and matching. Euclidean distance is calculated by comparing the differences between current trajectory features and historical trajectory features; the smaller the distance, the more similar the trajectories.

[0081] Specifically, the Euclidean distance is equal to the square root of the sum of the squares of the differences in each feature dimension.

[0082] For example, in one possible implementation, the current trajectory feature vector is [0.1, 0.05, 1.2], the historical trajectory feature vector is [0.12, 0.04, 1.15], and the calculated Euclidean distance is 0.06. Trajectories with a similarity exceeding a preset threshold are selected, and the corresponding collision event records are extracted. The collision event records contain information such as the location, time, collision type, and severity of the event.

[0083] Step S33 involves using a Support Vector Machine (SVM) to construct feature vectors from the set of spatial interference risk points and the sequence of historical collision events, and then performing classification training to obtain classification boundary parameters. The feature vectors are constructed by combining the coordinates and distance values ​​of the spatial interference risk points and the features of historical collision events, forming a high-dimensional feature space. The SVM then divides the feature space into collision and non-collision classes by finding the hyperplane that maximizes the classification margin. The training process uses kernel functions to transform the nonlinear problem into a linear one; commonly used kernel functions include radial basis functions and polynomial functions.

[0084] For example, in one embodiment, the feature vector dimension is 15, the training samples contain 500 collision events and 1000 non-collision events, and the classification accuracy after training the support vector machine is 92%. The classification boundary parameters include the coordinates and weights of the support vectors, which are used for subsequent collision probability calculations.

[0085] Step S34: Analyze the current composite motion state of the material frame based on the classification boundary parameters. If the current state exceeds the classification boundary, the collision probability increases, and a preliminary probability value is obtained. The current composite motion state is represented by feature vectors extracted from the current trajectory and interference risk points, and input into the trained support vector machine model. The classification boundary is determined by calculating the distance from the feature vector to the classification hyperplane. A positive distance indicates a non-collision class, and a negative distance indicates a collision class. The absolute value of the distance reflects the confidence of the classification; the larger the absolute value, the more reliable the classification.

[0086] For example, in one possible implementation, the distance from the current state feature vector to the classification boundary is -0.3, indicating that it belongs to the collision class, and the initial probability value is set to 0.7. This analysis method can quickly assess the collision risk of the current state based on historical experience.

[0087] Step S35 involves fusing the initial probability values ​​with motion trajectory features and calculating the adjusted probability using a weighted average method to determine the collision probability distribution under the current composite motion state of the material frame. Motion trajectory features include dynamic parameters such as velocity, acceleration, and curvature, which have varying degrees of influence on the collision probability. The weighted average method calculates the adjusted collision probability by assigning weights to different features. The weights are determined based on the degree of influence of each feature on the collision risk, typically through statistical analysis or expert experience.

[0088] For example, in one embodiment, the velocity feature weight is 0.4, the curvature feature weight is 0.3, and the distance feature weight is 0.3, resulting in a weighted average collision probability of 0.65. The collision probability distribution is represented by a probability density function, describing the likelihood of different collision types and severity. This fusion method improves the accuracy and reliability of probability estimation.

[0089] Step S4: Based on the collision probability distribution of the material frame under its current composite motion state, multiple candidate posture adjustment sequences are generated. A genetic algorithm is then used to iteratively optimize these sequences, resulting in an optimized set of material frame posture parameters. The collision probability distribution, provided in step S3, describes the likelihood of the material frame colliding under its current motion state. The candidate posture adjustment sequences are generated by adjusting the position and angle of the material frame, aiming to reduce the risk of collision. The genetic algorithm is an optimization method based on the principle of natural selection. By simulating the population evolution process, iteratively optimizing the candidate sequences, it finds the optimal posture parameters. The optimized set of material frame posture parameters contains adjusted position coordinates and angle information, effectively reducing the possibility of interference between the material frame and obstacles.

[0090] Specifically, the genetic algorithm uses candidate posture sequences as the initial population and gradually improves the fitness of the population through selection, crossover, and mutation operations, ultimately outputting posture parameters that satisfy collision probability constraints. This method can find suitable material frame posture schemes in complex environments.

[0091] Step S41: Based on the composite motion state of the material frame, acquire the real-time position coordinates and velocity vectors collected by the sensors to determine the collision probability distribution. The real-time position coordinates and velocity vectors are continuously provided by the sensor network, reflecting the current motion state of the material frame. The collision probability distribution is calculated using the Monte Carlo simulation method to simulate the collision event frequency of the material frame under various possible paths.

[0092] Specifically, Monte Carlo simulation generates a large number of possible motion trajectories through random sampling, and combines the collision probability distribution in step S3 to calculate the collision risk of each trajectory.

[0093] For example, in one possible implementation, the material box moves near the shelving area in the warehouse. Monte Carlo simulation generates 1000 possible trajectories, some of which show that the distance to the shelving is too close, leading to an increased collision probability. The collision probability distribution is output as probability values, providing a basis for generating alternative posture sequences. This method can comprehensively consider the uncertainty of the material box's motion, improving the accuracy of probability estimation.

[0094] Step S42: Multiple candidate posture adjustment sequences are generated based on the collision probability distribution, and the influence of ambient lighting is incorporated to obtain an extended adjustment sequence. The candidate posture adjustment sequences are generated by adjusting the position coordinates and rotation angle of the material frame, with each sequence corresponding to a possible motion posture. The influence of ambient lighting is considered to account for the impact of changes in light conditions in the warehouse on the sensor's acquisition accuracy.

[0095] For example, strong light or shadow can cause deviations in LiDAR data. By incorporating lighting factors, the alternative sequences are expanded to include adjustments that incorporate more environmental constraints.

[0096] For example, in one embodiment, the material box moves near a window in a warehouse. Changes in lighting cause fluctuations in sensor data. The extended adjustment sequence adapts to motion requirements under lighting interference by adding attitude fine-tuning options. The extended adjustment sequence includes multiple attitude combinations, each corresponding to a set of position and angle parameters. This extension method enhances the adaptability of attitude adjustment and addresses uncertainties in complex environments.

[0097] Step S43: The extended adjustment sequence is iteratively optimized using a genetic algorithm to obtain preliminary optimization parameters. The genetic algorithm uses the extended adjustment sequence as the initial population, with each sequence as an individual. The fitness function evaluates the quality of individuals based on collision probability and motion stability.

[0098] Specifically, the fitness function comprehensively considers factors such as the distance between the material box and obstacles, trajectory smoothness, and energy consumption. The genetic algorithm selects individuals with high fitness, performs crossover and mutation operations to generate a new generation of population, and converges to the optimal solution after multiple iterations.

[0099] For example, in one possible implementation, the material box moves in a narrow channel. The initial population contains 100 posture adjustment sequences. After 10 iterations, the genetic algorithm outputs a posture parameter that increases the minimum distance between the material box and obstacles to more than 0.3 meters while maintaining a stable trajectory. The initial optimization parameters include the adjusted position coordinates and rotation angles, providing a basis for subsequent corrections. This optimization method can find a balanced posture scheme under multiple constraints.

[0100] Step S44: If the preliminary optimization parameters deviate from the preset load balance factor, adjust the preliminary optimization parameters and determine the corrected posture parameters using a weighted fusion method. The load balance factor considers the weight distribution and stability requirements of the material handling equipment to ensure that the adjusted posture will not cause the equipment to tilt or become unbalanced. The deviation is calculated by comparing the difference between the preliminary optimization parameters and the load balance requirements. If the deviation exceeds the preset range,

[0101] For example, if the angle deviation exceeds 5 degrees, further correction is required. The weighted fusion method generates corrected attitude parameters by weighting and summing position, angle, and load balance factors.

[0102] For example, in one embodiment, after the material frame is adjusted in attitude on the handling equipment, preliminary parameters show that the front angle is too high, which may cause a shift in the center of gravity. The weighted fusion method generates correction parameters by reducing the front angle and adjusting the lateral position. The corrected attitude parameters can balance the collision risk and the equipment stability requirements. This method ensures the practicality and safety of attitude adjustment.

[0103] Step S45: Based on the corrected attitude parameters, the surrounding obstacle avoidance path is fused to obtain the optimized set of material frame attitude parameters. The surrounding obstacle avoidance path is generated based on the environmental geometry model and describes the safe path for the material frame to bypass obstacles. The corrected attitude parameters are combined with the avoidance path to ensure that the material frame avoids high-risk areas in space by adjusting its motion trajectory.

[0104] For example, in one possible implementation, the material frame moves in a densely shelved area. Correction of attitude parameters adjusts the lateral angle of the frame, and path avoidance further plans a curved trajectory around the shelving. The optimized set of frame attitude parameters includes the final position coordinates, rotation angle, and trajectory information, enabling collision-free movement in complex environments. This fusion approach improves the safety and efficiency of the frame's movement.

[0105] Step S5 involves simulating and calculating the spatial trajectory of the material frame after attitude adjustment using the optimized set of material frame attitude parameters. This trajectory is then compared with dynamic environmental change data to determine if there is any interference and to confirm the validity of the attitude parameter set. The spatial trajectory is generated by simulating the movement path of the material frame after attitude adjustment, describing the area occupied by the material frame in space. Dynamic environmental change data includes information such as obstacle movement and warehouse layout adjustments, and is collected in real time by the sensor network. The comparison process checks the intersection of the spatial trajectory and the dynamic environmental data to determine if there is any interference.

[0106] For example, in one embodiment, after the material frame adjusts its posture, the simulated trajectory shows that the distance between its top and the shelf increases to 0.4 meters. Dynamic environmental data confirms that the shelf position has not changed and the trajectory does not interfere. After confirming the validity of the posture parameter set, the final posture control scheme is generated. This simulation and comparison method ensures the feasibility of the posture adjustment scheme.

[0107] Step S51: Based on the optimized set of material frame posture parameters, simulate and calculate the space occupancy trajectory. The space occupancy trajectory is calculated by performing geometric transformations on the position coordinates and rotation angles of the material frame to generate a three-dimensional motion path of the material frame in space.

[0108] Specifically, the simulation process takes into account the geometry of the material frame;

[0109] For example, for the six faces of a cuboid frame, calculate the position range of each face after adjusting its posture.

[0110] For example, in one possible implementation, the material frame moves along a curved path within the warehouse. A set of attitude parameters adjusts its lateral angle, and the simulated spatial occupancy trajectory shows that the top and sides of the material frame do not overlap with obstacles. The spatial occupancy trajectory is stored as a point cloud or grid, providing data support for subsequent comparisons. This simulation method can intuitively reflect the range of motion of the material frame.

[0111] Step S52: Acquire dynamic environmental change data and compare it with the space occupancy trajectory to determine if there is no interference. The dynamic environmental change data is collected in real time by the sensor network, including obstacle position updates and the movement trajectories of other material frames. The comparison process determines whether there is an overlapping area by calculating the geometric intersection of the space occupancy trajectory and the dynamic environmental data.

[0112] For example, in one embodiment, a new mobile shelf is added to the warehouse. Dynamic environmental data updates the shelf's position, and the space occupancy trajectory shows that the material frame trajectory does not intersect with it, confirming no interference. If there is an intersection, the attitude parameters need to be readjusted. This comparison method can dynamically adapt to environmental changes, ensuring trajectory safety.

[0113] Step S53: If the simulated trajectory has no interference, the attitude parameter set is confirmed to be valid, and the final attitude control scheme is obtained. No interference confirmation is achieved by checking and comparing the results. If the space-occupied trajectory maintains a safe distance from all obstacles, the attitude parameter set is deemed valid. The final attitude control scheme includes optimized position coordinates, rotation angles, and trajectory planning, used to guide the material handling equipment.

[0114] For example, in one possible implementation, the material frame completes attitude adjustment in a dense area, and the simulated trajectory shows that the distance to all obstacles is greater than 0.3 meters, confirming the effectiveness of the scheme. The final attitude control scheme is output as a parameter set for use by the subsequent scheduling system. This confirmation method ensures the reliability of the scheme in practical applications.

[0115] Step S6: Obtain the final attitude control scheme and integrate it into the collaborative scheduling system. Issue control commands to the material handling equipment and confirm that the feedback data after command execution meets the risk assessment requirements. The final attitude control scheme, provided in step S5, includes the optimized position and angle parameters of the material frame. The collaborative scheduling system coordinates the operation of multiple handling devices, integrates the optimized scheme into the scheduling framework, and generates control commands. Feedback data is returned by the handling equipment after executing the commands, reflecting the command execution effect. By analyzing the feedback data, it is determined whether the risk assessment requirements are met, ensuring the safe movement of the material frame.

[0116] For example, in one embodiment, the optimization scheme adjusts the lateral angle of the material frame. After the scheduling system issues the command, the equipment feedback data shows that the material frame trajectory is stable and there is no risk of collision. This integration and feedback mechanism enables precise control of material frame handling.

[0117] Step S61: Obtain the final attitude control scheme, extract the equipment attitude adjustment parameters, and determine the degree of matching with the initial state of the handling equipment. The equipment attitude adjustment parameters are extracted from the final attitude control scheme and include position coordinates and angle values. The initial state is determined by the current attitude and load of the handling equipment, and the degree of matching is calculated by comparing the differences between the adjustment parameters and the initial state.

[0118] For example, in one possible implementation, the initial state of the material handling equipment shows an angular deflection of 2 degrees, while the optimization plan requires adjustment to 5 degrees. The difference is small, indicating a high degree of matching. The degree of matching is determined by a threshold; if the difference is within the allowable range, the parameter is confirmed to be applicable. This matching determination ensures that the optimization plan is compatible with the actual state of the equipment.

[0119] Step S62 involves integrating the final attitude control scheme into the cooperative scheduling system, calculating the compatibility score, and obtaining the integrated scheduling framework. The cooperative scheduling system operates with a distributed control architecture, and the parameters of the optimized scheme are embedded into the scheduling algorithm during the integration process. The compatibility score, calculated by comparing the parameter constraints of the optimized scheme and the scheduling system, reflects the executability of the scheme.

[0120] For example, in one embodiment, the optimization scheme requires the material box to move at a specific angle. The scheduling system verifies that this angle does not conflict with the paths of other devices, achieving a compatibility score of 0.9, thus confirming successful integration. The integrated scheduling framework contains coordination instructions for all devices, ensuring that the material box movement is synchronized with other tasks. This integration method improves the overall efficiency of warehouse handling.

[0121] Step S63: Issue control commands to the material frame handling equipment, obtain real-time feedback data, and adjust the commands to ensure data stability. The control commands are generated based on the integrated scheduling framework and include specific motion parameters. Real-time feedback data is returned by the equipment's sensors, reflecting the actual motion state of the material frame. If the feedback data deviation exceeds a preset threshold:

[0122] For example, if the positional deviation is greater than 0.1 meters, the instruction is regenerated by reducing the offset.

[0123] For example, in one possible implementation, after the material frame executes the instruction, the feedback data shows a slight deviation in trajectory. The instruction is then adjusted to reduce speed and reissued, after which the feedback data stabilizes. This dynamic adjustment mechanism ensures the accuracy of instruction execution.

[0124] Step S64: From the adjusted feedback data, determine whether the data meets the risk assessment requirements and obtain the final verification result. The risk assessment requirements include indicators such as trajectory stability and collision risk, and the feedback data is verified by comparing it with these indicators.

[0125] For example, in one embodiment, the adjusted feedback data shows that the material frame trajectory is consistent with the optimized scheme, and the collision probability is less than 0.01, meeting the risk requirements. The final verification results confirm the effectiveness of the optimized scheme and control commands, providing a reference for subsequent tasks. This verification method can ensure the safety and reliability of material frame handling.

[0126] Step S7: Based on the feedback data after instruction execution meeting the risk assessment requirements, the collision risk records in the historical database are updated. The safety assessment model is strengthened through a cumulative update mechanism, resulting in an enhanced risk assessment system. The feedback data, provided in step S6, includes the actual trajectory and status information of the material frame. The historical database updates collision risk-related information by recording this data. The cumulative update mechanism strengthens the safety assessment model and improves its predictive ability by analyzing patterns in historical and newly added data.

[0127] For example, in one possible implementation, feedback data shows that the collision risk of the material box is low in a specific area. After the database is updated, the safety model adjusts the risk weight of that area. This reinforcement mechanism can continuously optimize the accuracy of risk assessment.

[0128] Step S71: Verify the feedback data to obtain records that meet the risk requirements, resulting in an updated collision risk dataset. The feedback data is then compared with the risk assessment requirements to filter out records that meet safety standards.

[0129] For example, in one embodiment, feedback data shows that the material frame experienced no collisions after its orientation was adjusted, and the relevant records are marked as safe records. The updated collision risk dataset includes these records, reflecting the latest safety status of the material frame handling. This filtering method provides high-quality data for database updates.

[0130] Step S72 involves processing the historical database using a cumulative update mechanism and combining it with records that meet the risk requirements to obtain an enhanced security assessment model. The cumulative update mechanism extracts the cumulative pattern of collision risk through incremental updates to historical data.

[0131] For example, in one possible implementation, database records show that a certain path has experienced multiple minor collisions. After a new safety record is added, the update mechanism reduces the risk weight of that path. The enhanced safety assessment model optimizes the accuracy of risk prediction by integrating this data. This update method can dynamically adapt to changes in the transport environment.

[0132] Step S73: If the parameter deviation of the safety assessment model exceeds a preset threshold, adjust the associated attributes and determine the enhanced risk assessment system. The parameter deviation is calculated by comparing the difference between the model prediction results and the actual feedback data. If the deviation is large, for example, the deviation between the predicted collision probability and the actual probability exceeds 0.1, then adjust the model's weights or threshold.

[0133] For example, in one embodiment, the model predicts a high probability of collision in a certain area, but feedback data shows no collision. The risk weight for that area is then adjusted to be lower. The enhanced risk assessment system includes optimized model parameters and threshold settings, enabling more accurate prediction of risks for subsequent tasks.

[0134] Step S74: Obtain task planning details from the enhanced risk assessment system, integrate system architecture optimization and material handling task planning to obtain an execution feedback loop for subsequent material handling tasks. Task planning details include path planning, attitude adjustment strategies, etc., and incorporate prediction results from the safety assessment model.

[0135] For example, in one possible implementation, the enhanced system predicts a low-risk path, and the task planning details optimize the transport speed along that path. An execution feedback loop continuously collects feedback data and dynamically adjusts the plan to ensure the safety and efficiency of subsequent tasks. This cyclical mechanism enables continuous improvement of transport tasks.

[0136] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A pose control system for pallet handling collaborative scheduling, comprising a control system body, characterized in that, The control system operates by means of the following method: Position coordinates and velocity vector data are collected from the composite motion state of the material frame through a sensor network. The composite motion state includes the superposition of translation and rotation. A particle filter algorithm is used to filter noise from the position coordinate data and velocity vector data. The input of the particle filter algorithm is the position coordinate data and velocity vector data, and the output is the filtered position coordinates and the filtered velocity vector. Based on the filtered position coordinates and the filtered velocity vector, the instantaneous displacement vector of the material frame in space is determined, and the product of the instantaneous displacement vector and the preset time interval is obtained to obtain the displacement increment of the material frame. If the displacement increment of the material frame exceeds a preset threshold, it is determined to be an abnormal motion event. For the abnormal motion event, additional acceleration data is collected from the sensor network to obtain an abnormal acceleration sequence. By integrating the abnormal acceleration sequence with the filtered velocity vector, the integration operation accumulates the abnormal acceleration sequence on the time axis to the filtered velocity vector to calculate the displacement change, determine the predicted trajectory offset of the material frame, and obtain the difference between the predicted trajectory offset and the real-time trajectory change derived from the filtered position coordinates to obtain the trajectory correction factor. Adjust the filtered position coordinates based on the trajectory correction factor, determine the continuity of the adjusted position coordinates, and obtain the real-time spatial trajectory change of the material frame. The real-time spatial trajectory change of the material frame includes: By collecting real-time trajectory changes of the material frame through sensors and combining them with a preset environmental geometric model, the Euclidean distance from the coordinate points of each surface of the material frame to the surface points of surrounding obstacles is calculated, and the distance matrix is ​​obtained. For the distance matrix, if the minimum value is lower than a preset threshold, it is marked as a potential interference region, and the initial interference point in the material frame movement process is obtained; Based on the initial interference point, the relative position of the material frame velocity vector and the obstacle is obtained. By dividing the velocity vector by the relative position distance, the velocity adjustment coefficient is determined, and the adjusted trajectory change is obtained. Using the adjusted trajectory changes, the distance matrix is ​​recalculated and potential interference regions are marked to obtain a set of spatial interference risk points; The set of spatial interference risk points includes: Acquire spatial position data during the movement of the material frame, calculate position coordinate difference and velocity vector from the spatial position data, and extract motion trajectory features; The motion trajectory features are queried from the historical database, and the trajectory similarity is calculated using Euclidean distance. Collision event records under similar trajectories are extracted from trajectories with similarity exceeding a preset threshold to determine the historical collision event sequence. The support vector machine is used to construct feature vectors from the set of spatial interference risk points and the sequence of historical collision events as input. Classification training is performed through an optimization process that maximizes the classification margin to obtain the classification boundary parameters. Based on the classification boundary parameters, the current composite motion state of the material frame is analyzed, and it is determined that if the current state exceeds the classification boundary, the collision probability increases, thus obtaining a preliminary probability value. By fusing motion trajectory features with the initial probability values, a weighted average method is used to calculate probability adjustments and determine the collision probability distribution of the material frame under the current composite motion state. The collision probability distribution includes: Based on the composite motion state of the material frame, the real-time position coordinates and velocity vectors collected by the sensors are obtained, and the collision probability distribution is calculated using the Monte Carlo simulation method to determine the frequency of potential collision events. Based on the collision probability distribution, multiple alternative attitude adjustment sequences are generated, and the influence of ambient lighting factors are incorporated to obtain an extended adjustment sequence; A genetic algorithm is used to iteratively optimize the extended adjustment sequence. The input of the genetic algorithm is the individual of the extended adjustment sequence, and the output is the optimized individual after fitness evaluation, so as to obtain the preliminary optimization parameters. If the initial optimization parameters deviate from the preset load balancing factor, the initial optimization parameters are adjusted, and the corrected attitude parameters are determined by a weighted fusion method. Based on the corrected attitude parameters, the surrounding obstacle avoidance paths are integrated to obtain the optimized set of material frame attitude parameters.

2. The posture control system for pallet handling collaborative scheduling according to claim 1, wherein, The optimized set of material frame attitude parameters includes: The space occupancy trajectory of the material frame after the attitude adjustment is simulated and compared with the dynamic environmental change data. If there is no interference in the simulated trajectory, the attitude parameter set is confirmed to be valid, and the final attitude optimization scheme is obtained.

3. The pose control system for pallet handling collaborative scheduling according to claim 2, wherein, The final attitude optimization scheme includes: The attitude optimization scheme is obtained, the equipment attitude adjustment parameters are extracted from the collaborative scheduling system, the parameters are determined to match the initial state of the material frame handling equipment, and the degree of matching is judged by comparing the difference between the parameter values ​​and the initial state indicators. The attitude optimization scheme is integrated into the collaborative scheduling system. A preset threshold is used to judge the compatibility of the integration process. The compatibility score is calculated from the integration interface data to obtain the integrated scheduling framework. For the integrated scheduling framework, control commands are issued to the material handling equipment, and real-time feedback data after the command execution is obtained; Based on the real-time feedback data, if the data deviation exceeds the preset threshold, the control command is adjusted, the offset is subtracted from the deviation value, and the command is reissued to determine the stability of the adjusted feedback data. From the adjusted feedback data, it is determined that the data meets the risk assessment requirements, and the final verification result is obtained.

4. The pose control system for pallet handling collaborative scheduling according to claim 3, wherein, The final verification results include: By verifying the feedback data, records that meet the risk requirements are obtained, resulting in an updated collision risk dataset. Based on the collision risk dataset, a cumulative update mechanism is used to process the historical database. The mechanism extracts the cumulative collision risk from the historical data maintenance and combines it with the update mechanism to gradually add records that meet the risk requirements, thereby obtaining an enhanced safety assessment model. For the enhanced security assessment model, if the parameter deviation exceeds a preset threshold, the threshold is adjusted to adjust the associated attributes. The deviation is obtained by calculating the difference between the enhanced model parameters and the enhanced security assessment, and the enhanced risk assessment system is determined. Task planning details are obtained from the enhanced risk assessment system. These details are integrated with system architecture optimization and material handling task planning to obtain an execution feedback loop for subsequent material handling tasks.

5. The attitude control system for coordinated scheduling of material frame handling according to claim 4, characterized in that: The control system body includes: The data acquisition and filtering module is used to acquire position coordinates and velocity vector data from the composite motion state of the material frame through a sensor network, and to use a particle filter algorithm to filter the noise in the position coordinate data and velocity vector data, and output the filtered position coordinates and filtered velocity vectors. The displacement increment calculation module is used to determine the instantaneous displacement vector of the material frame in space based on the filtered position coordinates and the filtered velocity vector, and obtain the product of the instantaneous displacement vector and the preset time interval to obtain the displacement increment of the material frame. An abnormal event detection module is used to determine an abnormal motion event if the displacement increment of the material frame exceeds a preset threshold. For the abnormal motion event, additional acceleration data is collected from the sensor network to obtain an abnormal acceleration sequence. The trajectory correction factor calculation module is used to determine the predicted trajectory offset of the material frame by integrating the abnormal acceleration sequence and the filtered velocity vector, and to obtain the difference between the predicted trajectory offset and the real-time trajectory change derived from the filtered position coordinates to obtain the trajectory correction factor. The trajectory adjustment and continuity judgment module is used to adjust the filtered position coordinates according to the trajectory correction factor, judge the continuity of the adjusted position coordinates, and obtain the real-time spatial trajectory change of the material frame.

Images