Self-correction control method and device for hook-removing robot combined with servo analysis

Abstract

The present invention discloses a self-correction control method and device for a hook-removing robot combined with follow-up analysis, which relates to the field of machine control technology, including: obtaining the position data, angle data, force feedback data and environmental data of the hook-removing robot during operation to build a dynamic model of the robot operation; performing real-time analysis on the operation deviation of the hook-removing robot, determining the position deviation and force feedback deviation of the robot in different operation steps, and generating a deviation data set; performing fuzzy processing on the deviation data set according to preset correction rules, matching the preset correction rules and performing fuzzy reasoning calculation to generate an initial correction plan; optimizing the correction strategy based on the initial correction plan to generate an optimized correction plan; and controlling the hook-removing robot to perform self-correction control based on the optimized correction plan. The present invention solves the technical problem of low control accuracy in the prior art when performing self-correction control of the hook-removing robot, and achieves the technical effect of improving the self-correction control accuracy of the hook-removing robot.

Description

Technical Field

[0001] The present invention relates to the technical field of machine control, and in particular to a self-correction control method and device for a hook-removing robot combined with follow-up analysis. Background Art

[0002] In modern industrial automation, hook-removing robots are widely used in a variety of complex operational scenarios. These robots must operate at high speed and with high precision, often facing changing environmental factors and precise operational requirements. During these operations, precise control of parameters such as the robot's position, angle, and force feedback directly impacts both mission success and operational safety.

[0003] However, in actual applications, when performing complex operations, the dehooking robot is often subject to external interference, resulting in reduced operational accuracy or operational failure. In this case, the robot needs to have the ability to self-correct, that is, to be able to detect and correct its operational deviations in real time during the operation to ensure the smooth completion of the task. Traditional correction methods mainly rely on pre-set rules and control algorithms, but because these methods are usually difficult to adapt to the complex and changeable actual operating environment, the correction effect is limited, resulting in low control accuracy when performing self-correction control of the dehooking robot. Summary of the Invention

[0004] The present application provides a self-correction control method and device for a hook-removing robot combined with follow-up analysis, which is used to solve the technical problem of low control accuracy in the existing technology when performing self-correction control of a hook-removing robot.

[0005] In view of the above problems, the present application provides a self-correction control method and device for a hook-removing robot combined with follow-up analysis.

[0006] The first aspect of the present application provides a self-correction control method for a hook-removing robot combined with follow-up analysis, the method comprising:

[0007] The position data, angle data, force feedback data and environmental data of the unhooking robot during operation are acquired in real time to construct a dynamic model of the robot operation, wherein the dynamic model includes operation parameters and environmental status information; based on the dynamic model of the robot operation, the operation deviation of the unhooking robot is analyzed in real time to determine the position deviation and force feedback deviation of the robot in different operation steps, and generate a deviation data set; according to the preset correction rules, the deviation data set is fuzzy processed, the preset correction rules are matched and fuzzy reasoning calculation is performed to generate an initial correction plan; based on the initial correction plan, the correction strategy is optimized to generate an optimized correction plan; based on the optimized correction plan, the unhooking robot is controlled to perform self-correction regulation.

[0008] A second aspect of the present application provides a self-correcting control device for a hook-removing robot combined with servo analysis, the device comprising:

[0009] A dynamic model construction module, which obtains the position data, angle data, force feedback data and environmental data of the unhooking robot in real time during operation, and constructs a dynamic model of the robot operation, wherein the dynamic model includes operation parameters and environmental status information; a deviation data acquisition module, which analyzes the operation deviation of the unhooking robot in real time based on the robot operation dynamic model, determines the position deviation and force feedback deviation of the robot in different operation steps, and generates a deviation data set; a correction scheme acquisition module, which fuzzifies the deviation data set according to preset correction rules, matches the preset correction rules and performs fuzzy reasoning calculation to generate an initial correction scheme; a correction strategy optimization module, which optimizes the correction strategy based on the initial correction scheme and generates an optimized correction scheme; a self-correction control module, which controls the unhooking robot to perform self-correction control based on the optimized correction scheme.

[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0011] The present application obtains the position data, angle data, force feedback data and environmental data of the hook-removing robot in real time during operation, and constructs a dynamic model of the robot operation, wherein the dynamic model includes operation parameters and environmental state information; based on the dynamic model of the robot operation, the operation deviation of the hook-removing robot is analyzed in real time, the position deviation and force feedback deviation of the robot in different operation steps are determined, and a deviation data set is generated; according to a preset correction rule, the deviation data set is fuzzy processed, the preset correction rule is matched and fuzzy reasoning calculation is performed to generate an initial correction plan; based on the initial correction plan, the correction strategy is optimized to generate an optimized correction plan; based on the optimized correction plan, the hook-removing robot is controlled to perform self-correction control. The present invention solves the technical problem of low control accuracy in the prior art when performing self-correction control of the hook-removing robot. By acquiring and analyzing the position, angle, force feedback and environmental data of the hook-removing robot in real time, a dynamic model is constructed, the operation deviation is identified, an initial plan is generated through fuzzy processing and correction rules, and then a strategy is optimized. Finally, self-correction control is performed, thereby achieving the technical effect of improving the self-correction control accuracy of the hook-removing robot. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0013] Figure 1 A flow chart of a method for self-correction and control of a hook-removing robot combined with servo analysis provided in an embodiment of the present application;

[0014] Figure 2 Schematic diagram of the structure of the self-correcting control device of the hook-removing robot combined with follow-up analysis provided in an embodiment of the present application.

[0015] Explanation of the accompanying symbols: dynamic model construction module 11, deviation data acquisition module 12, correction plan acquisition module 13, correction strategy optimization module 14, self-correction control module 15. DETAILED DESCRIPTION

[0016] The present application aims to solve the technical problem of low control accuracy in the existing technology when performing self-correction control of the unhooking robot by providing a self-correction control method and device for the unhooking robot combined with follow-up analysis. By acquiring and analyzing the position, angle, force feedback and environmental data of the unhooking robot in real time, a dynamic model is constructed, operational deviations are identified, and an initial plan is generated through fuzzy processing and correction rules. Then, strategy optimization is performed, and finally self-correction control is performed, thereby achieving the technical effect of improving the self-correction control accuracy of the unhooking robot.

[0017] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only some of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0018] It should be noted that any variations of the terms "include" and "have" are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0019] Example 1, as Figure 1 As shown, the present application provides a self-correction control method for a hook-removing robot combined with follow-up analysis, the method comprising:

[0020] Step S100: acquiring the position data, angle data, force feedback data and environmental data of the unhooking robot in real time during operation, and constructing a dynamic model of the robot operation, wherein the dynamic model includes operation parameters and environmental status information.

[0021] In an embodiment of the present application, position data is acquired by installing displacement sensors on the robot, such as laser rangefinders and ultrasonic sensors, to monitor the position changes of the robot in real time, and the position of the robot in the workspace is acquired through the data output of the sensors. Next, angle data is acquired by installing angle sensors on the joints or other rotating parts of the robot, such as encoders, gyroscopes or accelerometers to acquire angle data. At the same time, force feedback data is acquired by installing force sensors on the end effector or key joints of the robot, such as strain gauge sensors or piezoelectric sensors. These data reflect the mechanical relationship between the robot and the work object, such as the size and change of the grasping or force applied. In addition, environmental data is acquired by collecting environmental data around the robot, such as temperature and humidity sensors and infrared sensors, including temperature, humidity, light intensity, obstacle distance, etc., which can help the system understand the status of the current working environment. Subsequently, through data fusion and dynamic model construction technology, the position, angle, force feedback and environmental data are integrated together to build a real-time dynamic model of the operation.

[0022] Furthermore, in the method provided in the embodiment of the application, the position, angle, force feedback and environmental data of the unhooking robot during operation are obtained in real time to construct a dynamic model of the robot operation, which also includes:

[0023] Based on the position data and in combination with the kinematic equations, a robot kinematic model is constructed; based on the angle data, the kinematic model is geometrically transformed to construct a robot posture channel; based on the force feedback data, a robot mechanical channel is constructed, and the robot mechanical channel is used to describe the mechanical relationship between the unhooking robot and the work object; based on the environmental data, the work scene is restored to construct a robot working environment channel; the robot posture channel, the robot mechanical channel, and the robot working environment channel are fused to construct the robot working dynamic model.

[0024] In an embodiment of the present application, a kinematic model is first constructed based on position data. Specifically, the position data obtained by the displacement sensor is used to calculate the position of each joint of the robot and its relationship with the final position in the workspace by applying kinematic equations, such as forward kinematics. Through these equations, the positional relationship between the various components of the robot in space is determined, and a preliminary kinematic model is constructed. Then, a geometric transformation is performed based on the angle data to construct a posture channel. First, the real-time angle data of each joint of the robot is obtained through the angle sensor. Then, using this angle data, a geometric transformation is performed on the previously constructed kinematic model. Specifically, the angle of the joint is converted into posture information in three-dimensional space, such as the direction and angle of the robotic arm, through a rotation matrix or a homogeneous transformation matrix. This conversion expands the original position information into complete posture information, forming a robot posture channel.

[0025] A mechanical channel is then constructed based on the force feedback data. The force applied by the robot to the work object is monitored in real time through the force sensor installed on the robot's end effector. A mechanical model is constructed using this force feedback data. This model includes mechanical parameters such as contact force and friction force, and analyzes the distribution and influence of these forces through mechanical equations such as Newton's second law, thereby forming a robot mechanical channel. Next, the work scene is restored based on the environmental data, and an environmental channel is constructed. Key parameters in the work environment, such as temperature, humidity, light, and obstacle positions, are obtained through environmental sensors. Using environmental data, the current work scene is restored through image processing technology, and the physical environment in which the robot is located is virtualized, enabling the robot to accurately locate and operate in complex or changing environments, thereby constructing an environmental channel.

[0026] Finally, the posture, force, and environment channels constructed in the previous steps are fused using data fusion techniques such as Kalman filtering or Bayesian estimation. The data from different channels is integrated and calibrated for consistency, ensuring that all data synchronously reflects the robot's state in the workspace. Through these steps, a fused dynamic model of the robot's operation is obtained.

[0027] Step S200: Based on the robot operation dynamic model, the operation deviation of the hook removal robot is analyzed in real time to determine the position deviation and force feedback deviation of the robot in different operation steps, and generate a deviation data set.

[0028] In an embodiment of the present application, the position and posture of the robot during the operation are first monitored in real time through the robot operation dynamic model, and the actual operation is compared with the preset ideal trajectory. The determination of the position deviation depends on the position data in the dynamic model. By calculating the difference between the current position coordinates of the robot and the ideal position, the degree of displacement of the robot in space is accurately identified. At the same time, the mechanical channel in the robot operation dynamic model provides real-time force feedback data. Using these data, compared with the preset ideal force value in the model, the force feedback deviation is determined, that is, the difference between the actual applied force and the expected force. Among them, the preset ideal force value and ideal position in the robot operation dynamic model are pre-set by technical experts based on the working requirements of the unhooking robot.

[0029] Through the above process, the position deviation and force feedback deviation of the robot in different operation steps are determined, and a deviation data set is generated.

[0030] Furthermore, in the method provided in the embodiment of the application, generating the deviation data set also includes:

[0031] Perform real-time trajectory comparison analysis, calculate the deviation from the preset ideal trajectory, and generate position deviation data; perform dynamic correction analysis on angle data, calculate the angular difference between the real-time operation angle and the preset target angle, and generate angle deviation data; perform anomaly detection on force feedback data, monitor the fluctuation amplitude and frequency of force feedback during operation, and compare them with the preset force feedback reference value to generate force feedback deviation data; perform real-time monitoring of environmental data, compare current environmental parameters with preset environmental standard values, and generate environmental deviation data; summarize the position deviation data, the angle deviation data, the force feedback deviation data, and the environmental deviation data to generate the deviation data set.

[0032] In an embodiment of the present application, the actual motion trajectory of the robot is first recorded in a dynamic model using position data obtained in real time. This actual trajectory is then compared with a preset ideal trajectory, and a path planning algorithm, such as the Dijkstra algorithm, is used to calculate the distance deviation between the two, thereby generating position deviation data. Next, angle data is obtained through the dynamic model, and these real-time angle data are geometrically transformed. By comparing these converted data with the preset target angle, the angle difference is corrected using geometric tools such as rotation matrices or quaternions, and finally the difference between the real-time operating angle and the target angle is calculated to generate angle deviation data. Here, the dynamic model provides real-time angle information, while the preset ideal angle is set independently for comparative analysis.

[0033] Anomaly detection is then performed based on the force feedback data from the dynamic model. Force sensors record the mechanical interaction between the robot and the workpiece in real time, and this data is fed into the dynamic model. This real-time force feedback data is then compared with preset ideal force values. Signal processing algorithms, such as fast Fourier transforms or wavelet transforms, are used to identify abnormal fluctuations, generating force feedback deviation data. This process also relies on real-time data from the dynamic model and compares it with independently set ideal force values.

[0034] Furthermore, the operating environment is monitored in real time through the environmental data channel within the dynamic model. Environmental sensors collect current environmental parameters, such as temperature, humidity, and light intensity, and record this data in the dynamic model. This data is then compared with pre-set ideal environmental standards. Through environmental modeling technology, deviations in these parameters are identified and environmental deviation data is generated, reflecting the difference between the actual environment and ideal conditions.

[0035] Finally, the actual deviation information obtained through the robot operation dynamic model is compared with the independently set ideal data, and the position deviation data, angle deviation data, force feedback deviation data and environmental deviation data are summarized to generate a deviation data set.

[0036] Step S300: According to the preset correction rules, the deviation data set is fuzzified, the preset correction rules are matched and fuzzy reasoning calculation is performed to generate an initial correction plan.

[0037] In an embodiment of the present application, the generated deviation data set is processed according to the preset correction rules to formulate an initial correction plan. Specifically, the deviation data set is first fuzzified. The fuzzified deviation data is then matched with the preset correction rules. These rules are set by technical experts based on historical data and describe the correction measures to be taken in different deviation situations. After the rule matching is completed, fuzzy reasoning calculations are performed to generate an initial correction plan, which includes how the robot should adjust operating parameters such as position, angle, and force.

[0038] Furthermore, in the method provided in the embodiment of the application, according to the preset correction rules, the deviation data set is fuzzified, the preset correction rules are matched and fuzzy reasoning calculations are performed to generate an initial correction scheme, and the method also includes:

[0039] Perform multi-level fuzzy processing on the deviation data set, map the initial fuzzy membership, and generate a multi-dimensional membership matrix; based on the multi-dimensional membership matrix, perform hierarchical rule screening to generate a primary correction rule set, perform correlation authentication on the primary correction rule set, and determine the optimized correction rule group; perform cross-activation processing on the optimized correction rule group, assign activation weights and set execution priorities, and output a priority correction rule matrix; based on the priority correction rule matrix, perform multiple defuzzification processing, extract preliminary operation instruction values, and generate a first correction plan; dynamically compare the first correction plan with real-time data to generate the initial correction plan.

[0040] In an embodiment of the present application, the deviation data generated by the unhooking robot during the operation process is first subjected to multi-level fuzzification processing. This process uses fuzzy logic technology to convert precise numerical deviations into fuzzy memberships, enabling the system to better handle uncertainty and complexity. Specifically, membership functions, such as triangular or trapezoidal membership functions, are used to map data such as position deviation, angle deviation, force feedback deviation, etc. into fuzzy sets. For example, the position deviation is fuzzified into "slight deviation" or "moderate deviation" to generate initial fuzzy memberships. Next, based on these initial fuzzy memberships, a multidimensional membership matrix is constructed. Each matrix dimension represents a type of deviation, such as position, angle, and force feedback, and each element represents the membership value of the corresponding deviation in different fuzzy sets. This matrix integrates the membership data of different deviation types and provides a comprehensive view for subsequent rule matching and screening.

[0041] A hierarchical rule screening process is then performed based on the multidimensional membership matrix. Using a rule-matching algorithm, a preset correction rule library is used to filter out the most relevant correction rules for the current deviation, forming a primary correction rule set. The preset correction rule library is developed by technical experts based on historical data. Each rule provides specific corrective measures for different operating scenarios and deviation conditions. For example, "If the position deviation is 'moderate' and the angle deviation is 'slight', then increase the robot arm angle adjustment by 3 degrees." Using rule-matching algorithms, such as conditional or rule-based reasoning, the data in the multidimensional membership matrix is compared to filter out the most relevant rules layer by layer. This screening process is multi-layered, starting with a preliminary screening based on the maximum membership of a single dimension. The membership of other dimensions is then gradually added for cross-screening. For example, if the matrix shows a high membership for position deviation, rules related to position adjustment are selected first, followed by the membership of angle and force feedback, ultimately resulting in a primary correction rule set that is appropriate for the current operating situation. The primary set of correction rules is then verified for relevance. Using association rule mining algorithms, such as the Apriori algorithm, a rule association matrix is constructed to identify dependencies or potential conflicts between the rules. Conflicts are identified using logical reasoning analysis, and conflict resolution algorithms, such as priority sorting, are used to resolve these conflicts, ultimately determining the optimized set of correction rules. Specifically, the Apriori algorithm is first used to mine association rules. By analyzing historical operation data, it identifies which rules are frequently triggered together or have dependencies in past operations. For example, if Rule A and Rule B are applied simultaneously in multiple operation scenarios, they are considered highly correlated. This process constructs a rule association matrix. Next, logical reasoning analysis is used to identify dependencies or potential conflicts between the rules. In this stage, the preconditions and execution results of each rule are analyzed to determine whether there are logical conflicts between them. For example, if Rule A requires increasing the gripping force of the robot arm, while Rule B requires reducing the gripping force under the same conditions, logical reasoning analysis can identify conflicting relationships between these rules. The rule relationships displayed in the association matrix are then used to identify which rules may conflict in actual operations and mark these as conflicting pairs. After identifying potential conflicts, a prioritization algorithm is used to resolve them. Prioritization determines the order in which rules are executed based on their importance, historical success rates, and the needs of the current operational environment. When resolving conflicts, high-priority rules are executed first, while lower-priority rules are adjusted or suppressed to avoid conflicts. Ultimately, these steps generate an optimized set of corrective rules.

[0042] After determining the optimal set of correction rules, a cross-activation process is performed to calculate the activation strength of each rule. Activation weights are assigned based on the needs and importance of the current operation, and execution priorities are set to generate a priority correction rule matrix. Based on the priority correction rule matrix, multiple defuzzification processes are performed, using defuzzification techniques such as the centroid method or the maximum membership method to convert fuzzy memberships into specific operational instruction values, thereby generating the first correction solution.

[0043] Finally, the first correction plan is dynamically compared with real-time operating data to ensure that it is fully consistent with the actual operating environment. This process verifies the consistency of the current operating environment with the pre-set conditions by comparing the real-time data with the assumed conditions of the first correction plan. If the comparison results show no discrepancies between the real-time data and the assumed data in the plan, the first correction plan is directly determined as the initial correction plan.

[0044] Furthermore, in the method provided in the embodiment of the application, cross-activation processing is performed on the optimized correction rule group, activation weights are assigned and execution priorities are set, and a priority correction rule matrix is output, which also includes:

[0045] Based on the multiple memberships of the optimized correction rule group, the multiple initial activation strengths of the optimized correction rule group are calculated to generate an initial activation matrix; the internal relationship of the optimized correction rule group is identified, and the initial activation matrix is adjusted according to the identification results to generate an adjusted activation matrix; based on the adjusted activation matrix, activation weights are assigned to the optimized correction rule group to generate a weight assignment matrix; based on the weight assignment matrix, the execution priority of the optimized correction rule group is set to generate a priority matrix; based on the priority matrix, rules are sorted to generate the priority correction rule matrix.

[0046] In the embodiment of the present application, the activation strength of each rule is first calculated based on the current environmental data and the membership of the optimized correction rule group. Specifically, the activation strength of each rule is calculated using the membership and rule importance factor: activation strength = membership × rule importance factor. The rule importance factor is a weight value pre-set by a technical expert. Finally, the calculated activation strength values are aggregated to form an initial activation matrix, where each row in the matrix represents a rule and each column represents a different environmental variable.

[0047] After generating the initial activation matrix, we further identify the internal relationships between the rules to ensure they can work in harmony. First, we use logical analysis and dependency analysis techniques to evaluate the preconditions and consequences between the rules. For example, if rule A requires the robot to slow down, while rule B requires the robot to adjust its angle after the speed has been reduced, we identify a dependency relationship between rules A and B. Furthermore, we identify potentially conflicting rules, such as one rule requiring an increase in the robot's gripping force while another requires a decrease. After identifying these relationships, we adjust the initial activation matrix. For rules with dependencies, we reduce the activation strength of the dependent rules to ensure that the predecessor rules are executed first. For conflicting rules, we reduce the activation strength of the conflicting rules or suppress one of the rules entirely to avoid conflicts in actual operations. The adjusted activation matrix is called the adjusted activation matrix, which reflects the modified activation strength of each rule after considering the internal relationships within the rules.

[0048] After the activation matrix is adjusted, each rule is assigned a weight to determine its execution priority in actual operations. Specifically, activation weights are first assigned based on the adjusted activation strength. This weighting takes into account factors such as the urgency of the task, the stability of the equipment status, the complexity of the operating environment, and safety requirements. For example, rules related to safety operations are prioritized, while rules related to improving efficiency are assigned lower weights. These weights reflect the importance of each rule in the current operating environment and the priority of its execution order, ultimately summarizing them to form a weight assignment matrix. Next, a priority sorting algorithm is used to convert the information in the weight assignment matrix into specific execution priorities, ensuring that rules with higher weights are executed first. This process generates the priority matrix.

[0049] Finally, all rules are sorted according to the priority matrix. Rules with higher priorities are ranked higher in the order. Once the rules are sorted, an execution sequence is generated based on this order. Finally, a priority correction rule matrix is generated based on the generated execution sequence. Each row of the matrix corresponds to an operation step, or rule, and each column corresponds to a different operation condition or key parameter in the step.

[0050] Step S400: Based on the initial correction plan, the correction strategy is optimized to generate an optimized correction plan.

[0051] Furthermore, in the method provided in the embodiment of the application, based on the initial correction scheme, the correction strategy is optimized to generate an optimized correction scheme, and the method further includes:

[0052] Based on the initial correction scheme, random combinations are performed to generate an initial correction strategy set; based on the robot operation dynamic model, simulated operations are performed according to the initial correction strategy set, and fitness calculation is performed to obtain the initial correction strategy fitness set; based on a preset fitness threshold, the initial correction strategy fitness set is selected to generate a second-generation correction strategy set; a cross operation is performed on the second-generation correction strategy set to exchange the adjustment parameters of position, angle and force feedback in different strategies to generate a third-generation correction strategy set; a mutation operation is performed on the three-generation correction strategy set, and a fitness evaluation is performed to select the strategy with the highest fitness as the optimized correction scheme.

[0053] In the embodiment of the present application, a random combination algorithm is first used to generate multiple possible correction strategies based on the generated initial correction scheme. The random combination algorithm generates a series of strategies by combining different operating parameters, such as the position, angle, and force feedback of the robot arm. These strategies have a variety of parameter configurations, ensuring that the system can explore various possible operation paths. These strategies are combined together to form the initial correction strategy set.

[0054] Next, based on the robot's dynamic operation model, the system simulates the initial set of corrective strategies. Each strategy in the initial set is fed into the dynamic operation model for simulation. During the simulation, the robot's response behavior is tracked in real time, including position changes, angle adjustment accuracy, and force feedback stability. The performance of each strategy is then calculated using a fitness function, which comprehensively evaluates key performance indicators such as operational accuracy, stability, and energy consumption. Finally, these fitness values are aggregated to form the initial set of corrective strategy fitness values.

[0055] On this basis, a fitness screening algorithm, based on a preset fitness threshold, selects the best performing strategies from the initial set of corrective strategies to generate a second-generation set of corrective strategies. Based on the preset fitness threshold, the fitness screening algorithm eliminates strategies with fitness below the threshold and aggregates the selected strategies into a second-generation set of corrective strategies. These strategies have demonstrated promising results in preliminary simulations.

[0056] To increase strategy diversity and explore new possible solutions, a crossover operation is performed on the second-generation correction strategy set, generating new strategy combinations by exchanging parameters between different strategies. Two or more strategies from the second-generation correction strategy set are then randomly selected as parents. These policies' different parameters, such as position parameters, angle parameters, and force feedback parameters, are cross-exchanged to generate new third-generation correction strategies. These crossover strategies may combine the advantages of their parent strategies. This process results in a three-generation correction strategy set.

[0057] Next, the three-generation correction strategy set is mutated to prevent the strategy from prematurely converging to a local optimum. Mutation introduces random changes. Common mutation methods include parameter perturbations and random increases and decreases. A subset of strategies from the three-generation correction strategy set is randomly selected, and parameters are fine-tuned or significantly adjusted. The mutated strategies are then reevaluated for fitness, using the previously defined fitness function to recalculate their performance in simulations. Through this fitness evaluation, the strategy with the highest fitness is ultimately selected as the optimized correction solution.

[0058] Furthermore, in the method provided in the embodiment of the application, based on the robot operation dynamic model, simulated operation is performed according to the initial correction strategy set, and fitness calculation is performed to obtain the initial correction strategy fitness set, which also includes:

[0059] The initial correction strategy set is input into the robot operation dynamic model for simulation operation to obtain the simulation operation results of the robot when executing each correction strategy, including the accuracy, operation stability and energy consumption after correction; based on the simulation operation results, a preset fitness function is used for evaluation to generate an initial correction strategy fitness set.

[0060] In this embodiment, the generated initial set of correction strategies is first input into the robot's dynamic operation model to simulate the actual operation performance of the robot arm under different strategies. During the simulation, each operation step of the robot arm is tracked and recorded in detail to obtain the post-correction operation accuracy, operational stability, and energy consumption.

[0061] After the simulation is complete, the results of each strategy are evaluated using a pre-set fitness function. The fitness function is a mathematical formula used to comprehensively evaluate the effectiveness of a strategy, including the weighted summation of multiple indicators, such as operational accuracy, stability, and energy consumption.

[0062] Finally, the fitness values of all strategies are summarized to generate the initial correction strategy fitness set.

[0063] Among them, the preset fitness function is:

[0064]

[0065] Among them, F represents the fitness value. The higher the fitness value, the better the performance of the strategy in the current simulation environment. p=(\begin{aligned}) represents the operational accuracy error. This is the deviation between the robot arm and the target operation after performing the correction operation. S represents operational stability, which is the robot arm's sensitivity to external environmental influences throughout the operation. A higher S value indicates a more stable robot arm during operation. C represents energy consumption, which is the amount of energy consumed to complete the correction task. The weight factors w1, w2, and w3 are used to indicate the relative importance of each indicator in the fitness calculation and are pre-set: w1 = 0.5, w2 = 0.3, and w3 = 0.2.

[0066] To calculate E p Calculate the difference between the robot's actual operating results and the expected results. Precision error includes position deviation, angular deviation, and other factors. The calculation formula is the square root of the sum of the squares of these deviations, or the Euclidean distance, to measure the overall error.

[0067] To obtain operational stability S, it is expressed by the variance or standard deviation of the vibration sensor data.

[0068] To calculate the energy consumption C, record the energy consumption of the robot arm during the entire operation, that is, the power required for the robot arm to move.

[0069] Step S500: Based on the optimized deviation correction scheme, the unhooking robot is controlled to perform self-correction regulation.

[0070] In an embodiment of the present application, an optimized deviation correction scheme is input into the control system of the hook removal robot. This scheme includes specific operating parameters, such as the position of the robotic arm, angle adjustment, and force feedback settings. A real-time control algorithm ensures that the robot accurately executes each step of the scheme. For example, if the scheme requires adjusting the angle of the robotic arm before performing a grasping operation, these operating instructions are executed sequentially in the established order. This process completes the self-correction control of the hook removal robot.

[0071] In the embodiments of the present application, in summary, the embodiments of the present application have at least the following technical effects:

[0072] The present application obtains the position data, angle data, force feedback data and environmental data of the hook-removing robot in real time during operation, and constructs a dynamic model of the robot operation, wherein the dynamic model includes operation parameters and environmental state information; based on the dynamic model of the robot operation, the operation deviation of the hook-removing robot is analyzed in real time, the position deviation and force feedback deviation of the robot in different operation steps are determined, and a deviation data set is generated; according to a preset correction rule, the deviation data set is fuzzy processed, the preset correction rule is matched and fuzzy reasoning calculation is performed to generate an initial correction plan; based on the initial correction plan, the correction strategy is optimized to generate an optimized correction plan; based on the optimized correction plan, the hook-removing robot is controlled to perform self-correction control. The present invention solves the technical problem of low control accuracy in the prior art when performing self-correction control of the hook-removing robot. By acquiring and analyzing the position, angle, force feedback and environmental data of the hook-removing robot in real time, a dynamic model is constructed, the operation deviation is identified, an initial plan is generated through fuzzy processing and correction rules, and then a strategy is optimized. Finally, self-correction control is performed, thereby achieving the technical effect of improving the self-correction control accuracy of the hook-removing robot.

[0073] Example 2, based on the same inventive concept as the self-correction control method of the hook-removing robot combined with the follow-up analysis in the above embodiment, Figure 2 As shown, the present application provides a self-correction control device for a hook-removing robot combined with servo analysis. The device and method embodiments in the present application are based on the same inventive concept. The device includes:

[0074] A dynamic model construction module 11, the dynamic model construction module 11 acquires the position data, angle data, force feedback data and environmental data of the unhooking robot in real time during the operation process, and constructs a dynamic model of the robot operation, wherein the dynamic model includes operation parameters and environmental status information; a deviation data acquisition module 12, the deviation data acquisition module 12 performs real-time analysis on the operation deviation of the unhooking robot based on the robot operation dynamic model, determines the position deviation and force feedback deviation of the robot in different operation steps, and generates a deviation data set; a correction scheme acquisition module 13, the correction scheme acquisition module 13 performs fuzzy processing on the deviation data set according to preset correction rules, matches the preset correction rules and performs fuzzy reasoning calculation to generate an initial correction scheme; a correction strategy optimization module 14, the correction strategy optimization module 14 optimizes the correction strategy based on the initial correction scheme to generate an optimized correction scheme; a self-correction control module 15, the self-correction control module 15 controls the unhooking robot to perform self-correction control based on the optimized correction scheme.

[0075] Furthermore, the device is also used to implement the following functions:

[0076] Based on the position data and in combination with the kinematic equations, a robot kinematic model is constructed; based on the angle data, the kinematic model is geometrically transformed to construct a robot posture channel; based on the force feedback data, a robot mechanical channel is constructed, and the robot mechanical channel is used to describe the mechanical relationship between the unhooking robot and the work object; based on the environmental data, the work scene is restored to construct a robot working environment channel; the robot posture channel, the robot mechanical channel, and the robot working environment channel are fused to construct the robot working dynamic model.

[0077] Furthermore, the device is also used to implement the following functions:

[0078] Perform real-time trajectory comparison analysis, calculate the deviation from the preset ideal trajectory, and generate position deviation data; perform dynamic correction analysis on angle data, calculate the angular difference between the real-time operation angle and the preset target angle, and generate angle deviation data; perform anomaly detection on force feedback data, monitor the fluctuation amplitude and frequency of force feedback during operation, and compare them with the preset force feedback reference value to generate force feedback deviation data; perform real-time monitoring of environmental data, compare current environmental parameters with preset environmental standard values, and generate environmental deviation data; summarize the position deviation data, the angle deviation data, the force feedback deviation data, and the environmental deviation data to generate the deviation data set.

[0079] Furthermore, the device is also used to implement the following functions:

[0080] Perform multi-level fuzzy processing on the deviation data set, map the initial fuzzy membership, and generate a multi-dimensional membership matrix; based on the multi-dimensional membership matrix, perform hierarchical rule screening to generate a primary correction rule set, perform correlation authentication on the primary correction rule set, and determine the optimized correction rule group; perform cross-activation processing on the optimized correction rule group, assign activation weights and set execution priorities, and output a priority correction rule matrix; based on the priority correction rule matrix, perform multiple defuzzification processing, extract preliminary operation instruction values, and generate a first correction plan; dynamically compare the first correction plan with real-time data to generate the initial correction plan.

[0081] Furthermore, the device is also used to implement the following functions:

[0082] Based on the multiple memberships of the optimized correction rule group, the multiple initial activation strengths of the optimized correction rule group are calculated to generate an initial activation matrix; the internal relationship of the optimized correction rule group is identified, and the initial activation matrix is adjusted according to the identification results to generate an adjusted activation matrix; based on the adjusted activation matrix, activation weights are assigned to the optimized correction rule group to generate a weight assignment matrix; based on the weight assignment matrix, the execution priority of the optimized correction rule group is set to generate a priority matrix; based on the priority matrix, rules are sorted to generate the priority correction rule matrix.

[0083] Furthermore, the device is also used to implement the following functions:

[0084] Based on the initial correction scheme, random combinations are performed to generate an initial correction strategy set; based on the robot operation dynamic model, simulated operations are performed according to the initial correction strategy set, and fitness calculation is performed to obtain the initial correction strategy fitness set; based on a preset fitness threshold, the initial correction strategy fitness set is selected to generate a second-generation correction strategy set; a cross operation is performed on the second-generation correction strategy set to exchange the adjustment parameters of position, angle and force feedback in different strategies to generate a third-generation correction strategy set; a mutation operation is performed on the three-generation correction strategy set, and a fitness evaluation is performed to select the strategy with the highest fitness as the optimized correction scheme.

[0085] Furthermore, the device is also used to implement the following functions:

[0086] The initial correction strategy set is input into the robot operation dynamic model for simulation operation to obtain the simulation operation results of the robot when executing each correction strategy, including the accuracy, operation stability and energy consumption after correction; based on the simulation operation results, a preset fitness function is used for evaluation to generate an initial correction strategy fitness set.

[0087] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0088] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

[0089] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A self-correction control method for a hook-removing robot combined with servo analysis is characterized in that: The method comprises: Acquire the position data, angle data, force feedback data and environmental data of the dehooking robot in real time during operation, and build a dynamic model of the robot operation, wherein the dynamic model includes operation parameters and environmental status information; Based on the robot operation dynamic model, the operation deviation of the hook removal robot is analyzed in real time to determine the position deviation and force feedback deviation of the robot in different operation steps, and generate a deviation data set; According to the preset correction rules, the deviation data set is fuzzified, the preset correction rules are matched and fuzzy reasoning calculations are performed to generate an initial correction plan; Based on the initial correction plan, optimizing the correction strategy and generating an optimized correction plan; Based on the optimized deviation correction scheme, the unhooking robot is controlled to perform self-correction regulation; The position, angle, force feedback, and environmental data of the dehooking robot during operation are acquired in real time to construct a dynamic model of the robot operation. The method includes: Based on the position data and in combination with kinematic equations, a robot kinematic model is constructed; Based on the angle data, geometrically transform the kinematic model to construct a robot posture channel; Based on the force feedback data, a robot mechanical channel is constructed, wherein the robot mechanical channel is used to describe the mechanical relationship between the unhooking robot and the work object; Performing operation scene restoration based on the environmental data to construct a robot operation environment channel; Perform channel fusion on the robot posture channel, the robot mechanics channel, and the robot operation environment channel to construct the robot operation dynamic model; Generate a biased dataset, including: Perform real-time trajectory comparison and analysis, calculate the deviation from the preset ideal trajectory, and generate position deviation data; Perform dynamic correction analysis on angle data, calculate the angle difference between the real-time operating angle and the preset target angle, and generate angle deviation data; Perform anomaly detection on force feedback data by monitoring the fluctuation amplitude and frequency of force feedback during operation, comparing it with the preset force feedback reference value, and generating force feedback deviation data; Monitor environmental data in real time, compare current environmental parameters with preset environmental standard values, and generate environmental deviation data; The position deviation data, the angle deviation data, the force feedback deviation data, and the environment deviation data are aggregated to generate the deviation data set.

2. The self-correction control method for a hook-removing robot combined with servo analysis according to claim 1 is characterized in that: According to the preset correction rules, the deviation data set is fuzzified, the preset correction rules are matched and fuzzy reasoning calculation is performed to generate an initial correction plan, and the method includes: Performing multi-level fuzzification processing on the deviation data set, mapping the initial fuzzy membership, and generating a multi-dimensional membership matrix; Based on the multidimensional membership matrix, hierarchical rule screening is performed to generate a primary correction rule set, and the primary correction rule set is authenticated for relevance to determine an optimized correction rule group; Performing cross-activation processing on the optimized deviation correction rule group, allocating activation weights and setting execution priorities, and outputting a priority deviation correction rule matrix; Based on the priority correction rule matrix, multiple defuzzification processes are performed to extract preliminary operation instruction values and generate the first correction plan; The first correction plan is dynamically compared with the real-time data to generate the initial correction plan.

3. The self-correction control method for a hook-removing robot combined with servo analysis according to claim 2 is characterized in that: Cross-activation processing is performed on the optimized deviation correction rule group, activation weights are assigned and execution priorities are set, and a priority deviation correction rule matrix is output. The method includes: Calculating multiple initial activation intensities of the optimized deviation-correcting rule group based on multiple membership degrees of the optimized deviation-correcting rule group to generate an initial activation matrix; Identifying internal relationships of the optimized deviation-correcting rule group, adjusting the initial activation matrix according to the identification results, and generating an adjusted activation matrix; Based on the adjusted activation matrix, assigning activation weights to the optimized deviation correction rule group to generate a weight assignment matrix; Based on the weight distribution matrix, setting the execution priority of the optimization correction rule group to generate a priority matrix; The rules are sorted based on the priority matrix to generate the priority correction rule matrix.

4. The self-correction control method for a hook-removing robot combined with servo analysis according to claim 1 is characterized in that: Based on the initial correction plan, the correction strategy is optimized to generate an optimized correction plan, the method comprising: Based on the initial correction scheme, random combinations are performed to generate an initial correction strategy set; Based on the robot operation dynamic model, simulate the operation according to the initial correction strategy set, and perform fitness calculation to obtain the initial correction strategy fitness set; Selecting the initial correction strategy fitness set based on a preset fitness threshold to generate a second-generation correction strategy set; Performing a cross operation on the second-generation correction strategy set, exchanging adjustment parameters of position, angle, and force feedback in different strategies, and generating a third-generation correction strategy set; The three generations of correction strategy sets are mutated and evaluated for fitness, and the strategy with the highest fitness is selected as the optimized correction solution.

5. The self-correction control method for a hook-removing robot combined with servo analysis according to claim 4 is characterized in that: Based on the robot operation dynamic model, simulated operation is performed according to the initial correction strategy set, and fitness calculation is performed to obtain the initial correction strategy fitness set. The method includes: Inputting the initial correction strategy set into the robot operation dynamic model for simulation operation, and obtaining the simulation operation results of the robot when executing each correction strategy, including the accuracy, operation stability and energy consumption after correction; According to the simulation operation results, a preset fitness function is used for evaluation to generate an initial correction strategy fitness set.

6. The self-correction control device of the hook-removing robot combined with the follow-up analysis is characterized in that: The device comprises: A dynamic model building module, which acquires the position data, angle data, force feedback data and environmental data of the unhooking robot in real time during operation and builds a dynamic model of the robot operation, wherein the dynamic model includes operation parameters and environmental status information; a deviation data acquisition module, which performs real-time analysis of the operation deviation of the dehooking robot based on the robot operation dynamic model, determines the position deviation and force feedback deviation of the robot in different operation steps, and generates a deviation data set; A correction scheme acquisition module, which performs fuzzy processing on the deviation data set according to preset correction rules, matches the preset correction rules and performs fuzzy reasoning calculation to generate an initial correction scheme; A correction strategy optimization module, which optimizes the correction strategy based on the initial correction plan and generates an optimized correction plan; A self-correction control module, which controls the unhooking robot to perform self-correction control based on the optimized correction scheme; The position, angle, force feedback, and environmental data of the dehooking robot during operation are acquired in real time to construct a dynamic model of the robot operation. The method includes: Based on the position data and in combination with kinematic equations, a robot kinematic model is constructed; Based on the angle data, geometrically transform the kinematic model to construct a robot posture channel; Based on the force feedback data, a robot mechanical channel is constructed, wherein the robot mechanical channel is used to describe the mechanical relationship between the unhooking robot and the work object; Performing operation scene restoration based on the environmental data to construct a robot operation environment channel; Perform channel fusion on the robot posture channel, the robot mechanics channel, and the robot operation environment channel to construct the robot operation dynamic model; Generate a biased dataset, including: Perform real-time trajectory comparison and analysis, calculate the deviation from the preset ideal trajectory, and generate position deviation data; Perform dynamic correction analysis on angle data, calculate the angle difference between the real-time operating angle and the preset target angle, and generate angle deviation data; Perform anomaly detection on force feedback data by monitoring the fluctuation amplitude and frequency of force feedback during operation, comparing it with the preset force feedback reference value, and generating force feedback deviation data; Monitor environmental data in real time, compare current environmental parameters with preset environmental standard values, and generate environmental deviation data; The position deviation data, the angle deviation data, the force feedback deviation data, and the environment deviation data are aggregated to generate the deviation data set.

Images