Robotic cable handling positioning system

Abstract

The present application relates to cable handling positioning technical field, disclose a kind of cable automatic handling positioning system based on manipulator.The system includes data acquisition, state analysis, influence evaluation, synchronous correlation evaluation and position optimization adjustment module.Data acquisition module real-time acquisition position and state data when handling;State analysis module analyzes actual and target position deviation, and generates position abnormal signal beyond preset range;Influence evaluation module monitors moving speed accordingly, calculates moving influence value, and generates influence big signal beyond threshold;Synchronous correlation evaluation module obtains speed change trend and position change similarity value according to the signal, and correlation calculates synchronous correlation value;Position optimization adjustment module combines influence big signal and current target position, and obtains adjusted position using position adjustment coefficient.The system can realize whole process monitoring, accurately adjust position, and adapt to different working conditions.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cable handling and positioning, in particular to an automatic cable handling and positioning system based on a mechanical hand. BACKGROUND

[0002] In the scenarios of cable production, storage and engineering construction, the handling and positioning of cables are important links in daily operations. With the improvement of industrial automation level, mechanical hands gradually replace manual work and become the main execution equipment for cable handling, but they still face many problems to be solved in actual application. In the traditional mechanical hand cable handling mode, action execution mostly depends on preset fixed procedures, and lacks effective collection and analysis of real-time data in the handling process. In the handling process, the actual position of the cable may deviate from the target position due to factors such as uneven weight distribution of the cable itself and external environmental interference.

[0003] In the prior art, some mechanical hand systems have simple position detection functions, but can only send alarm signals when the deviation exceeds a large range, and cannot quantitatively evaluate the influence of the deviation. For example, when the moving speed of the mechanical hand changes, it is impossible to determine the correlation between the speed change and the position deviation, nor can it be determined whether the current speed change will further exacerbate the position deviation, resulting in lack of pertinence in subsequent adjustment. In addition, when adjusting the position, the existing system mostly uses fixed adjustment parameters, without considering the adaptability of the adjustment coefficient under different working conditions, and the problems of over-adjustment or under-adjustment often occur, which not only affects the handling efficiency, but also may cause damage to the cable, increasing the production cost. At the same time, the existing system lacks synchronous correlation analysis of the speed change trend and the similarity of the position change, and it is difficult to grasp the dynamic change law in the handling process as a whole, and it is impossible to predict the possible position abnormalities in advance, resulting in lag in fault handling and further reducing the stability and reliability of the entire handling system. SUMMARY

[0004] The purpose of the present application is to provide an automatic cable handling and positioning system based on a mechanical hand to solve the problems raised in the background art.

[0005] To achieve the above purpose, the present application provides an automatic cable handling and positioning system based on a mechanical hand, which comprises:

[0006] a data acquisition module for acquiring position data and handling state data in real time during the handling of the cable by the mechanical hand;

[0007] a state analysis module for analyzing the deviation of the actual position of the cable from the target position based on the position data, comparing the deviation with a preset position range, and generating a position abnormality signal if the deviation exceeds the preset position range;

[0008] An influence evaluation module is configured to monitor the moving speed data of the robot arm based on the position abnormality signal, acquire the speed data in a monitoring period, calculate a moving influence value, and compare the moving influence value with a moving influence threshold to generate an influence size signal; if the moving influence value is greater than or equal to the moving influence threshold, an influence large signal is generated;

[0009] A synchronous correlation evaluation module is configured to acquire a speed change trend value and a position change similarity value based on the influence large signal, perform correlation calculation on the speed change trend value and the position change similarity value, and obtain a synchronous correlation value.

[0010] A position optimization adjustment module is configured to acquire a position adjustment coefficient based on the influence large signal, combine the current target position with the position adjustment coefficient to obtain an adjusted position, and complete position adjustment.

[0011] Preferably, the data acquisition module acquires the data in the following manner:

[0012] The robot arm carrying path is divided into a plurality of carrying areas, position data in each carrying area is acquired, and the position data of the carrying areas is integrated to determine a comprehensive position value.

[0013] Carrying state data in each carrying area is acquired, and the carrying state data is fused to generate a state comprehensive value.

[0014] Preferably, the state analysis module analyzes in the following manner:

[0015] Based on the comprehensive position value, a difference distance between the actual position and the target position is calculated.

[0016] The difference distance is compared with a maximum boundary value and a minimum boundary value of the preset position range, respectively; if the difference distance is greater than the maximum boundary value or less than the minimum boundary value, it is determined that the deviation exceeds the preset position range.

[0017] When the position abnormality signal is generated, the state comprehensive value is combined.

[0018] Preferably, the influence evaluation module calculates in the following manner:

[0019] In the monitoring period, the fluctuation characteristics of the speed data are extracted.

[0020] Based on the fluctuation characteristics and the position abnormality signal, the moving influence value is derived.

[0021] The moving influence value is compared with an upper limit value and a lower limit value of the moving influence threshold; if the moving influence value exceeds the upper limit value or is lower than the lower limit value, the influence large signal is generated.

[0022] Preferably, the synchronous correlation evaluation module acquires in the following manner:

[0023] Identify a speed change trend value, which reflects the overall change direction of the speed data in the monitoring period;

[0024] Identify a position change similarity value, which reflects the matching degree of the position data and the speed data;

[0025] Correlate the speed change trend value and the position change similarity value to obtain a synchronous correlation value.

[0026] Preferably, the coefficient acquisition mode of the position optimization adjustment module is:

[0027] Summarize the synchronous correlation values corresponding to the large signals;

[0028] Deduce the position adjustment coefficient based on the synchronous correlation value;

[0029] Combine the current target position with the position adjustment coefficient to output the adjusted position.

[0030] Preferably, the system further comprises:

[0031] A path updating module for re-planning the carrying path of the manipulator based on the synchronous correlation value;

[0032] The planning mode of the path updating module is to extract the synchronous correlation value, map it to the key nodes of the carrying path, and generate updated path data.

[0033] Preferably, the mapping mode of the path updating module is:

[0034] Identify the position offset feature in the key node;

[0035] Correct the coordinates of the key node based on the position offset feature and the updated path data;

[0036] Output the corrected carrying path.

[0037] Preferably, the system further comprises:

[0038] A carrying management module for coordinating the carrying operation of the manipulator based on the adjusted position and the corrected carrying path;

[0039] The coordination mode of the carrying management module is to sort the priority sequence of the carrying operation and allocate the action execution time.

[0040] Preferably, the sorting mode of the carrying management module is:

[0041] Extract the action conflict identifier in the priority sequence;

[0042] Re-arrange the priority sequence based on the action conflict identifier and the action execution time;

[0043] generating final handling instructions.

[0044] Compared with the prior art, the present application has the beneficial effects that:

[0045] By setting the data acquisition module, the position data and handling state data during the cable handling process of the manipulator can be collected in real time, realizing full-process dynamic monitoring of the handling process, breaking the limitation of traditional systems that can only perform stage detection, and capturing subtle changes in position and state in a timely manner, providing comprehensive and accurate data support for subsequent analysis and adjustment.

[0046] The state analysis module analyzes the deviation of the actual position of the cable from the target position based on the position data, and compares it with the preset position range to generate a position abnormality signal, so that the position abnormality can be quickly identified, avoiding subsequent serious problems caused by deviation accumulation, allowing the operator or system to know the position deviation in the first time, and creating conditions for timely intervention.

[0047] After receiving the position abnormality signal, the influence evaluation module calculates the movement influence value by monitoring the movement speed data of the manipulator, and compares it with the threshold to generate an influence size signal, realizing quantitative judgment of the influence degree of the position abnormality. Unlike traditional systems that can only determine whether an abnormality exists, the severity of the abnormal influence can also be determined, and when the movement influence value is greater than or equal to the threshold, an influence large signal is generated, prompting the system to prioritize handling of such high-impact abnormal situations, reasonably allocating system resources, and improving the efficiency and pertinence of fault handling.

[0048] The synchronous correlation evaluation module obtains the speed change trend value and the position change similarity value based on the influence large signal, and performs correlation calculation to obtain the synchronous correlation value, which can deeply explore the internal relationship between the speed change and the position change. Through this correlation analysis, the dynamic correlation law of the two can be clearly grasped, the influence of speed change on position change and the synchronization of the two changes can be understood, and a more in-depth analysis basis is provided for subsequent position adjustment, which helps to find the key factors causing the position abnormality from the root, rather than just staying on the surface of deviation correction.

[0049] The position optimization adjustment module adjusts the position based on the influence large signal acquisition position adjustment coefficient and in combination with the current target position to achieve individualization and precision of position adjustment. Unlike the traditional system which adopts a fixed adjustment parameter mode, the module can dynamically determine the adjustment coefficient according to the actual influence degree, so that the adjusted position is more in line with the current working condition requirements, effectively avoiding the problems of over-adjustment or under-adjustment, and ensuring the position accuracy during the cable carrying process. At the same time, the precise position adjustment reduces the cable collision, wear and tear and other situations caused by position deviation, prolongs the service life of the cable, and reduces the production cost. In addition, the whole system forms a complete closed loop from data acquisition, abnormality identification, influence evaluation, correlation analysis to precise adjustment through the cooperative work of each module, significantly improves the stability and reliability of the manipulator cable carrying positioning, ensures the smooth progress of the carrying process, improves the overall carrying efficiency, is suitable for the needs of different specifications of cables and different carrying scenes, and has strong practicality and adaptability. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 a timing diagram of the manipulator-based automatic cable carrying positioning system of the present application;

[0051] Figure 2 a flowchart for the data acquisition module;

[0052] Figure 3 a flowchart for the influence evaluation module calculation;

[0053] Figure 4 a flowchart for the synchronous correlation evaluation module acquisition. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0055] Please refer to Figure 1 The present application provides a manipulator-based automatic cable carrying positioning system, which comprises a data acquisition module, a state analysis module, an influence evaluation module, a synchronous correlation evaluation module and a position optimization adjustment module. The specific implementation is as follows:

[0056] The data acquisition module collects position data and handling state data in real time during the handling process of the manipulator, the state analysis module calculates the deviation between the actual position of the cable and the target position based on the position data, and generates a position abnormal signal if the deviation exceeds the preset range. The influence evaluation module monitors the moving speed data of the manipulator according to the position abnormal signal, calculates the moving influence value and generates an influence size signal. The synchronous correlation evaluation module obtains the speed change trend value and the position change similarity value based on the influence size signal, and obtains the synchronous correlation value through correlation calculation. The position optimization adjustment module combines the position adjustment coefficient and the current target position, and outputs the adjusted position to complete the position optimization.

[0057] Embodiment 1: see Figure 2 In the manipulator-based automatic cable handling positioning system, the data acquisition module first divides the entire handling path of the manipulator into multiple logical regions, and the division of each region is based on the motion characteristics of the manipulator, the physical state of the cable, and the specific conditions of the handling environment. For example, in a long-distance handling scenario, the path may be divided into a start-up acceleration zone, a uniform handling zone, and an end deceleration zone; while in a complex path handling, the regions may be divided according to the turning angle or obstacle distribution. Within each handling region, the position data of the manipulator is collected by the encoders, laser range finders, or vision sensors installed on the joints and the end of the manipulator. The encoders record the angle changes of each joint, the laser range finders measure the distance between the end of the manipulator and the reference point, and the vision sensors capture the real-time position of the cable through image recognition technology. After coordinate transformation and unit unification, these raw data are transmitted to the data processing unit.

[0058] The force sensor is installed on the end effector of the manipulator, which monitors the load changes in real time when grabbing the cable. When the cable swings or slips due to external forces, the load data will fluctuate accordingly. The inertial measurement unit records the acceleration and angular velocity of the manipulator during motion, reflecting the stability of the handling process. The environmental sensor detects external interference factors such as wind speed and vibration, and these data are correlated with the motion state of the manipulator. For example, when the wind speed increases, the swing amplitude of the cable may also increase, causing abnormal fluctuations in the load data. After aligning by time stamp, all state data are sent to the data fusion algorithm for processing.

[0059] The comprehensive position value is calculated by weighted average of all position data in the region, and the weight is dynamically adjusted according to the reliability and real-time nature of the data. For example, the vision sensor has a higher weight when the lighting conditions are good, while in insufficient light, it relies on the data of the encoder and laser range finder. The state comprehensive value is generated by feature extraction and fusion of multi-source state data through a neural network model, which learns the correlation rules between different state parameters in different handling scenarios during the training phase.

[0060] The target position is pre-set by the handling task planning system, including three-dimensional spatial coordinates and attitude information. When the manipulator is disturbed during handling, the actual position will deviate from the target position, forming a difference distance. The pre-set position range is determined according to the diameter of the cable, flexibility and handling accuracy requirements. For example, for large diameter cables, the position range may be set wider to allow moderate swinging; while for fine cables in precision applications, the range is correspondingly reduced. When the difference distance exceeds the pre-set boundary, the state analysis module will combine the load change, swing amplitude and other information in the state comprehensive value to judge the nature of the deviation. For example, if the load suddenly decreases accompanied by large amplitude swing, it may indicate that the cable has slipped; while sustained load fluctuations may be due to transmission abnormalities of the manipulator joints.

[0061] The generation of the position abnormality signal not only depends on the numerical size of the deviation, but also on the duration and change trend of the deviation. Transient and short-term deviations may be caused by random disturbances, which will be marked by the system as low-priority abnormalities; while continuously increasing deviations will be judged as high-priority abnormalities, triggering the rapid response of the subsequent modules. The environmental data in the state comprehensive value is used to assist in judging the cause of the abnormality, for example, when strong wind records are detected during the occurrence of an abnormality, the system will attribute the deviation to environmental factors rather than mechanical failure. This multi-factor comprehensive analysis mechanism improves the accuracy of abnormality judgment and avoids false triggering of adjustment actions.

[0062] The system records the frequency and severity of deviations in each region during each handling process, and identifies problem-prone areas through cluster analysis. For regions where abnormalities frequently occur, the system will automatically reduce the region range or increase the sensor sampling frequency, achieving dynamic adjustment of monitoring density. The collection period of handling state data is also not fixed, when the state analysis module detects abnormal signs, it will temporarily increase the data collection frequency to obtain more detailed state change process. This adaptive collection mechanism ensures the response speed of the system while effectively reducing the data processing load.

[0063] For the straight carrying section, the system uses linear interpolation to complete the data between sampling points; while in the curved motion section, high-order polynomial fitting is used to track the trajectory. This differentiated data processing method is more in line with the actual motion law of the manipulator, improving the accuracy of the comprehensive position value. The fusion algorithm of state data is continuously updated, and the system regularly compares the actual carrying result with the predicted state to adjust the parameters of the neural network model, so that the state comprehensive value can more accurately reflect the true carrying condition. The preset position range of the state analysis module is not globally uniform, but sets differentiated standards for different carrying stages. In the acceleration and deceleration stages, due to the effect of inertia, a larger position fluctuation range is allowed; while in the uniform speed carrying stage, a more stringent position tolerance standard is implemented. This dynamic threshold management method conforms to the physical motion law and avoids unnecessary adjustment actions. The deviation judgment algorithm also introduces fuzzy logic processing of boundary conditions, when the difference distance approaches the boundary value, the system will comprehensively consider historical data and adjacent region state to make more reasonable abnormal judgment.

[0064] The system records the processing results of each abnormal signal, establishes an association model between signal characteristics and final adjustment effect. When similar characteristics appear again, the system can predict the possible adjustment requirements and prepare optimization schemes in advance. This predictive analysis shortens the system response time, especially when dealing with repetitive carrying tasks. Abnormal signals also have a confidence index reflecting the reliability of the judgment result for subsequent module reference. The actual effect after position adjustment will be recorded and transmitted back to the data acquisition module for sensor parameter calibration and optimization of region division strategy. The state analysis module will also continuously optimize the deviation judgment algorithm based on historical adjustment data, forming a continuously improving intelligent system. This closed-loop operation mode enables the system to gradually adapt to different carrying environments and cable types, improving overall carrying precision and stability.

[0065] The control instructions contain expected position and state targets, and the data acquisition module compares these theoretical values with actual measurement values to quickly identify abnormal conditions. The operation results of the state analysis module are transmitted to the subsequent processing unit through a high-speed bus to ensure the real-time response of the system. All collected data and analysis results have accurate time stamps to ensure data synchronization and accurate cause-and-effect relationship between modules. The system also has a data caching mechanism that can still maintain basic function operation when communication is temporarily interrupted, and automatically synchronizes the latest state after communication is restored.

[0066] The hardware configuration of the data acquisition module has a redundant design, with multiple backup channels for critical sensors. When an abnormal signal is detected on one channel, the system automatically switches to the backup channel. The algorithm implementation of the state analysis module uses multi-thread parallel computing to ensure real-time processing capability under high data throughput. The system performs regular self-diagnosis to detect the working status of each sensor and the running effect of the data analysis algorithm, and promptly discovers and fixes potential problems. This high-reliability design ensures stable operation of the system in industrial environments.

[0067] Example 2: see Figure 3 In the cable automatic handling positioning system, the influence evaluation module is responsible for analyzing the degree of influence of the manipulator's moving speed on the position deviation. This module starts a monitoring period immediately after receiving a position abnormality signal, and the duration of this period is dynamically adjusted according to the handling stage. For example, during the precise positioning stage of the cable, the monitoring period is shortened to the millisecond level to capture transient changes; while in the long-distance uniform speed handling stage, the period is extended to the second level to observe the trend influence. During the monitoring period, the encoders installed on each joint of the manipulator record angular velocity data at a fixed sampling frequency, while the laser Doppler velocimeter directly measures the linear velocity of the end effector. These raw velocity data are converted into three-dimensional velocity vectors in the manipulator's working coordinate system after coordinate transformation.

[0068] The system first performs noise reduction on the original velocity sequence, eliminating high-frequency interference signals through a sliding window filter. Then it calculates the standard deviation and coefficient of variation of the velocity in the time domain, reflecting the overall amplitude of the velocity fluctuations; in the frequency domain, it identifies periodic fluctuation components through spectral analysis, such as fixed frequency vibrations caused by mechanical transmission systems. For non-periodic fluctuations, the system uses entropy analysis to quantify their randomness. After aligning all fluctuation characteristic parameters by timestamp, a complete fluctuation feature vector is formed.

[0069] The system constructs a three-dimensional state space, with the coordinate axes representing the position deviation, the velocity fluctuation amplitude, and the fluctuation frequency characteristics. The distribution of the state point cloud formed by historical data reveals the mapping rules between the three: when the position deviation increases, if accompanied by high-frequency large-amplitude velocity fluctuations, it is determined that the speed influence is significant. In specific calculations, the system retrieves the neighboring historical points in the current state space and determines the movement influence value through a weighted voting mechanism. The weight allocation considers a time decay factor, with recent data having higher weights than early data, allowing the model to adapt to slow changes in system characteristics.

[0070] The base threshold is determined according to the dynamic parameters of the robot, such as the maximum allowed acceleration, joint torque limit, and other physical constraints. The dynamic threshold is adjusted through an online learning mechanism, and the system continuously records the speed fluctuation range under normal handling conditions. When the fluctuation characteristics continue to deviate from the reference level, the threshold sensitivity is automatically increased. The threshold interval is divided into three levels: safe zone, warning zone, and abnormal zone. When the movement impact value is in the safe zone, the system only records data and does not trigger an alarm; when it enters the warning zone, it starts an enhanced monitoring mode; if it breaks through the boundary of the abnormal zone, it immediately generates a major impact signal.

[0071] The system checks whether the signal appears continuously in a short period of time to avoid false triggering caused by occasional interference. At the same time, it correlates environmental sensor data to exclude false positives caused by external factors such as sudden strong winds. Signal strength is graded to represent the degree of impact, with level one indicating slight impact and level three indicating severe impact. The signal is accompanied by detailed feature fingerprints, including the spectral feature map of speed fluctuation, time domain waveform slices, and the evolution curve of related position deviation. These data provide a complete analysis background for subsequent modules.

[0072] When a major impact signal is detected, the system automatically traces back two monitoring periods of data to analyze the accumulation process of abnormal phenomena. For example, before a cable unhooking accident, there are usually signs of gradually increasing speed fluctuation amplitude. This retrospective analysis helps to distinguish between sudden abnormalities and gradual failures, providing a basis for maintenance decisions. At the same time, the system extends the monitoring for one period after the adjustment measure is implemented to observe the response, forming a complete abnormal event closed-loop record.

[0073] The speed data collection network uses a heterogeneous sensor fusion strategy. In addition to joint encoders, vibration accelerometers are deployed at key positions of the robot to capture abnormal vibrations of the transmission system; high-speed cameras are installed at the cable suspension points to indirectly calculate the swing speed through image analysis; current sensors are installed in the control cabinet to infer speed changes through motor current fluctuations. Multiple data sources are fused through a Kalman filter to eliminate measurement errors of a single sensor. When a type of sensor fails, the system automatically switches to a backup data source to ensure continuous monitoring.

[0074] The system periodically evaluates the effectiveness of feature extraction, and when it finds that a newly emerging fluctuation pattern cannot be described by existing features, it automatically starts the feature expansion process. The new feature dimension is added to the feature vector after correlation verification, such as the discovery of a unique low-frequency fluctuation pattern in a wet environment. The system adds the energy proportion feature in the 0.5-2Hz frequency band. This dynamically evolving feature system enables the system to adapt to changes in complex working conditions.

[0075] After each handling task, the system compares the actual result with the predicted impact value, calculates the prediction deviation, and automatically adjusts the model parameters when the same-direction deviation occurs continuously. Major abnormal events trigger the model reconstruction process, and the system re-trains the state space mapping relationship using recent data. The model version management function records all change history and allows rollback to a stable version if necessary. This continuous optimization mechanism improves the accuracy of impact value assessment over time. A transition interval is set near the threshold boundary, and the impact value within this interval does not immediately trigger a signal, but starts a verification measurement. The system temporarily increases the sensor sampling frequency and performs multiple rounds of review measurements. Only when the review result exceeds the threshold for three consecutive times, the final impact signal is confirmed. This design effectively filters transient interference and improves the reliability of the signal. The threshold itself is also dynamically adjusted according to seasonal factors, such as appropriately relaxing the threshold in winter low-temperature environment to consider the impact of increased lubricant viscosity.

[0076] The first-level signal triggers the local controller to fine-tune the speed parameters; the second-level signal notifies the host computer to optimize the handling path; and the third-level signal starts the safety protocol, including reducing the running speed, issuing sound and light warnings, or even temporarily stopping. The response strategy considers the importance of the handling stage, such as when the cable passes through the area of precision equipment, even the first-level signal will trigger path re-planning. The signal lifecycle management function ensures that each signal has a clear start, confirmation, processing, and closing record, forming a traceable event chain.

[0077] Cases are automatically classified according to feature fingerprints, forming a failure mode knowledge base. When the matching degree of new signal features with historical cases exceeds a certain value, the system automatically retrieves the corresponding processing scheme, shortening the response time. For new abnormalities with unknown characteristics, the system starts a special recording mode to collect high-density data for subsequent in-depth analysis. The case library is periodically compressed and archived, retaining feature fingerprints and key parameters for long-term trend analysis. During implementation, the impact assessment module and the robot control system maintain two-way communication. The expected speed curve in the control command is transmitted to the assessment module in real time as the theoretical baseline value. The assessment result is fed back to the control system for adjusting the speed planning parameters. This closed-loop interaction enables the system to maintain handling efficiency while dynamically controlling the impact range of speed fluctuations. All data exchanges are strictly synchronized by timestamp to ensure the timing consistency of analysis results and action execution.

[0078] Feature extraction tasks are assigned to multiple dedicated processors for parallel execution, mobile impact value calculation is handled by a high-performance computing unit, and threshold comparison and signal generation are completed by a real-time controller. The task scheduler dynamically allocates computing resources based on system load, prioritizing real-time performance of core functions during peak periods. The data pipeline uses a ring buffer design to avoid processing delays caused by data accumulation.

[0079] Example 3: refer toFigure 4 The synchronous correlation evaluation module is responsible for establishing the dynamic relationship model between the manipulator speed variation and the position deviation in the cable automatic handling positioning system. The module is activated immediately after receiving the impact of the large signal, first extracts the complete data set in the current monitoring period from the system cache, including the speed sampling value and the position sampling value in the form of time series. The speed data comes from the angular velocity sensor array of each joint of the manipulator, and these raw data are converted into the linear velocity vector of the end effector in the three-dimensional workspace through kinematics forward solution. The position data comes from the fusion output of the laser tracker and the visual positioning system, reflecting the actual motion trajectory of the cable grabbing point.

[0080] The calculation of the speed trend value uses the trend decomposition algorithm to separate the speed time series into trend items, periodic items, and random items. The trend item is extracted by the weighted moving average method, and the window width is adjusted dynamically according to the handling stage: a narrower window is used to capture rapid changes in the acceleration segment, and a wider window is used to smooth noise in the uniform speed segment. After normalization processing, the extracted trend curve is converted into a dimensionless trend intensity index, which quantifies the persistence and directionality of the speed change. For example, when the manipulator performs a deceleration action, the trend intensity index presents a negative value, and its absolute value reflects the deceleration amplitude.

[0081] The system first constructs a position deviation vector sequence, each vector containing three-dimensional position offset and time derivative information. The time axes of the speed change curve and the position deviation curve are aligned through the dynamic time warping algorithm, eliminating the phase difference between them. The similarity measure considers two dimensions: shape similarity and change amplitude matching degree. The shape similarity is evaluated by the correlation of the curve derivative, and the amplitude matching degree compares the normalized fluctuation intensity. The final position change similarity value is the weighted harmonic average of these two components, and the weight coefficient is configured according to the accuracy requirement of the handling task.

[0082] The calculation of the synchronous correlation value uses the following formula:

[0083]

[0084] Where: represents the synchronous correlation value, is the speed trend intensity of the th sampling point, is the position change similarity component at the corresponding time, is the time decay weight factor, is the total number of sampling points in the monitoring period. The time decay weight factor in the formula gives higher weight to recent data, and its value decreases exponentially over time. The value of the synchronous correlation value ranges from -1 to 1, a positive value indicates that the speed and position change are coordinated, and a negative value indicates that there is a contradiction between the two.

[0085] The system maintains a historical case library, each case stores the sequence of synchronization correlation values under a specific working condition and its corresponding successful adjustment parameters. When new synchronization correlation values are input, the retrieval module searches for the most similar correlation pattern in the case library and extracts the corresponding adjustment parameters as initial suggested values. The suggested values are then corrected by the real-time optimizer, which takes into account real-time factors such as the dynamics state of the robot, cable tension, and environmental disturbances. The final output position adjustment coefficient is a three-dimensional vector, each component corresponds to the adjustment proportion of an axis in the work coordinate system.

[0086] During implementation, the system checks whether the adjusted position meets the robot workspace constraints, and automatically introduces a softening factor when approaching joint limits to avoid mechanical structure damage. The position adjustment command is sent to the motion controller through the real-time communication bus, accompanied by the confidence index of the synchronization correlation value. The controller determines the execution strength of the adjustment action based on the confidence, using direct positioning for high confidence and gradual approximation for low confidence.

[0087] The system runs multiple trend extraction threads of different time scales simultaneously, from second-level trends to minute-level trends. The trend fusion algorithm selects the dominant scale according to the current handling stage, for example, short-term trends are preferred during precise positioning, while long-term trends are emphasized during long-distance handling. This multi-scale analysis can capture both transient anomalies and slow drifts, improving the comprehensiveness of trend judgment. The trend strength index also includes a reliability score, reflecting the support of current data quality for trend judgment, for reference by subsequent modules.

[0088] The system uses a deep neural network to build a feature extractor for position changes, which is trained offline through a large amount of historical data to learn typical motion patterns during cable handling. In online operation, the network outputs the latent space representation of the position sequence in real time, and similarity calculation is performed in this low-dimensional space, which improves computational efficiency and enhances feature expression ability. Network parameters are updated regularly through incremental learning to absorb new motion patterns.

[0089] The system continuously monitors the consistency of correlation values and actual adjustment effects, and when systematic deviations are found, it starts the model parameter calibration process. The calibration process uses a sliding window least squares method, with the window width automatically adjusted according to the characteristics of the deviation: a wide window is used for slow-changing systematic deviations, while a narrow window is used for sudden anomalies for fast response. The calibrated parameters are tested for stability before being put into formal use, avoiding overfitting to noise data.

[0090] The adjustment coefficients of adjacent time points need to satisfy the smooth transition condition to avoid sudden changes in the motion of the manipulator. The system establishes a three-dimensional coefficient space, and each newly generated coefficient must be located within the feasible region determined by the previous coefficients. When the coefficient suggested by the optimization algorithm exceeds the feasible region, the constraint processor will project it onto the boundary of the region while recording the violation event for subsequent analysis. This constraint mechanism ensures the continuity of the motion during the transfer process and reduces the impact on the cable.

[0091] The system establishes a time delay model based on the dynamic response characteristics of the manipulator to predict the actual effective time of the adjustment command. When sending the position adjustment command, the predicted effective time is sent synchronously, and the motion controller starts the compensation action accordingly. This feedforward control method effectively offsets the influence of system response delay, making the actual position accurately track the adjustment target. The time delay model parameters are updated periodically through closed-loop identification to adapt to the slow changes in the characteristics of the mechanical system.

[0092] The addition of new cases requires multiple verifications, including adjustment effect evaluation, feature uniqueness test, and conflict resolution. The case retrieval uses the approximate nearest neighbor algorithm, and the index structure is periodically reorganized to maintain query efficiency. The elimination of expired cases is based on the dual indicators of usage frequency and timeliness, while retaining their statistical features for trend analysis. The case library implements version control, supporting rollback to historical stable versions if necessary.

[0093] Through the above implementation, the synchronous correlation evaluation module builds an accurate correlation model between speed and position change. Multi-scale trend analysis comprehensively captures dynamic characteristics, intelligent similarity evaluation accurately quantifies the degree of motion coordination, and rigorous mathematical formulas establish a standardized correlation calculation framework. The generation of position adjustment coefficients integrates historical experience and real-time optimization, and the execution process ensures accuracy through advanced prediction and compensation. The systematic integration of these technical elements enables the module to effectively coordinate the relationship between manipulator speed and cable position, providing key dynamic adjustment capabilities for the automatic transfer system.

[0094] Example 4: The path update module in the cable automatic transfer positioning system is responsible for dynamically optimizing the motion trajectory of the manipulator based on real-time monitoring data. The module receives processed synchronous correlation values from the synchronous correlation evaluation module, which reflect the coordination degree between manipulator speed change and cable position deviation. The module has a path planning engine inside, which divides the entire transfer path into several key nodes, each containing spatial coordinates, motion parameters, and allowable error range attributes. The system defaults to linear interpolation to connect adjacent nodes to form an initial path, but starts the path re-planning process when it detects abnormal synchronous correlation values.

[0095] The second derivative of the path is calculated by the system, and the positions where the curvature exceeds a set threshold are marked as key nodes. For cable handling tasks, typical nodes include the start point, end point, obstacle avoidance point, and load direction mutation point. Each node is provided with an influence radius parameter, and when a synchronization correlation anomaly occurs within the influence radius of a node, the node is marked as a node to be optimized. The node optimization priority is evaluated comprehensively according to the anomaly degree and the importance of the position, as shown in Table 1.

[0096] Table 1: Example of node optimization queue.

[0097] Node number Spatial coordinates (x, y, z) Synchronization correlation value Influence radius (mm) Optimization priority N203 (1250,780,320) -0.47 150 Emergency N156 (980,1020,280) -0.32 200 High N341 (1530,650,350) -0.18 120 Medium N422 (1420,880,310) -0.25 180 High

[0098] Within the influence radius of the node to be optimized, the system activates a high-density monitoring mode, and the position sampling frequency is increased to five times the regular value. By comparing the deviation sequence of the theoretical coordinates and the actual coordinates, the main direction and the change gradient of the offset are calculated. The offset direction is divided into axial offset and radial offset, the axial offset is along the motion direction of the manipulator, and is usually caused by improper speed control; the radial offset is perpendicular to the motion direction, and is mostly caused by cable swing or external interference. The system establishes an offset feature vector, which includes dimensions such as offset type, duration, and spatial distribution pattern.

[0099] For nodes dominated by axial offset, the system adjusts the node spacing while keeping the path curvature unchanged, and corrects the position by changing the motion speed distribution. Radial offset is compensated by introducing new transition nodes, and the position of the transition node is calculated according to the offset feature to ensure that the corrected path meets the joint angle limit of the manipulator. Each update operation is verified by dynamic simulation to check whether it will cause torque overrun or resonance risk. The update data that passes the verification is packaged as a path segment descriptor, which includes the start node, the end node, the array of intermediate path points, and the motion constraint conditions.

[0100] The system first calculates a preliminary correction vector according to the offset feature, and applies the vector to the original node coordinates to obtain the candidate position. The candidate position needs to pass the reachability test, including inverse kinematics solution existence verification and workspace boundary check. For complex offset patterns, the system uses a step-by-step correction strategy, first solving the main offset component and then processing the secondary component. The synchronization correlation value is re-evaluated after each correction until it returns to the normal range or reaches the maximum number of iterations. The final node coordinate update scheme is recorded as a versioned revision, supporting rollback operations if necessary.

[0101] The system inserts multiple auxiliary points between adjacent key nodes, and uses a cubic spline curve fitting to generate a continuous and smooth trajectory. The second derivative of the curve is strictly controlled during the interpolation process to avoid sudden changes in acceleration. For load-sensitive areas, the density of path points is additionally increased to ensure that the cable swing amplitude is within the allowed range. The smoothed path is converted into a mechanical hand control instruction sequence, including axis position, speed and acceleration parameters, which is sent to each joint controller through a real-time communication bus.

[0102] The system initiates special monitoring after path update, recording the degree of agreement between the actual motion trajectory and the expected path. Evaluation indicators include position tracking error, speed following accuracy and synchronization correlation value improvement. These data feedback to the path planning engine for adjusting node influence radius parameters and optimizing strategy weights. For path sections that repeatedly appear problems, the system will automatically mark them as sensitive areas and implement preventive optimization in subsequent tasks.

[0103] Each node entry contains initial coordinates, previous correction records, related synchronization correlation value change curves, and optimization effect scores. The database supports multi-dimensional query functions, allowing retrieval of related nodes by spatial area, time window, or abnormal type. Periodic database compression operations merge redundant records, but retain all key revision metadata for audit tracking. Any coordinate correction operation requires electronic signature verification to ensure the credibility of the instruction source. Update instruction transmission uses redundancy check encoding to prevent misoperation caused by communication errors. Before implementing path changes, the system temporarily reduces the speed of the mechanical hand to reserve a safety time window for emergency stop. Major path adjustments also require a human confirmation link, and operators can review the change details through the visual interface before authorizing execution.

[0104] When detecting that the synchronization correlation value continues to deteriorate after path update, the system automatically reverts to the last stable version. For sudden and serious abnormalities, an emergency recovery path is provided, which is simplified to only retain the most critical safety nodes, ensuring that the mechanical hand can stop or return in the safest way. The state data during the recovery process is recorded in detail for post-analysis and system improvement. The path update module realizes the dynamic optimization of the cable carrying path. Key node identification accurately captures path characteristics, offset feature analysis reveals the root cause of the problem, iterative correction algorithm ensures adjustment accuracy, and closed-loop evaluation mechanism continuously improves optimization effect. The entire path update process is carried out under strict safety constraints, ensuring both flexibility and reliability. This dynamic path optimization capability enables the mechanical hand to adapt to complex changes in the carrying environment, ensuring the accuracy and stability of cable positioning.

[0105] The system maintains global optimization objectives, avoiding local optimization that can degrade performance in other sections. When modifying a single node, the system assesses the impact on upstream and downstream nodes and implements associated adjustments as necessary. For long-distance transport tasks, the system uses a segmented optimization strategy, dividing the path into logical sections and processing them in parallel, followed by section alignment optimization. This holistic approach prevents secondary problems during optimization.

[0106] The system provides a three-dimensional path display interface, using different colors to represent node status: red for abnormal nodes, yellow for nodes to be optimized, and green for normal nodes. Operators can view analysis results such as path curvature cloud maps and synchronous correlation value heat maps through the interface. The visualization tool also supports path comparison functions, allowing simultaneous display of differences between the original path and the modified path, helping to understand the optimization effect.

[0107] The system continuously monitors environmental parameter changes, such as environmental temperature affecting the dimensional stability of the robot arm and air flow changing the cable swing characteristics. These environmental factors are converted into optimization constraints, allowing the generated path to automatically adapt to environmental changes. For periodic environmental disturbances, the system can also learn their change patterns and adjust path parameters in advance to compensate.

[0108] Example 5: The transport management module plays a core role in action coordination and task scheduling in the cable automatic transport positioning system. This module receives adjusted position data from the position optimization adjustment module and modified transport paths from the path update module, and converts this information into executable robot action sequences. The action planning engine first analyzes the target requirements of the transport task and decomposes them into basic action units, each containing action type, target parameters, allowed error range, and execution time limit. The system establishes an action resource model, quantifying the motion capabilities of each joint of the robot, the load state of the end effector, and the physical constraints of the cable, providing physical basis for action scheduling.

[0109] The priority sequence is generated using a dynamic sorting algorithm. The system calculates an urgency score for each action unit, which considers task progress requirements, position deviation risks, and device safety factors. For example, when the cable passes through a narrow channel, anti-collision actions have the highest priority; while in the stable transport phase, position accuracy control actions are prioritized over speed optimization actions. Conflict detection is implemented during the sorting process to identify action combinations that may compete for the same mechanical resources. When potential conflicts are detected, the system automatically inserts buffer actions or adjusts action timing to eliminate resource competition risks. The sorting result forms an initial priority queue, which is continuously updated during task execution.

[0110] The system maintains a conflict pattern library that stores the characteristics of typical conflict scenarios that have occurred in the past. Real-time monitoring of the action parameter combinations in the action queue, when detecting similar feature combinations to historical conflict cases, immediately mark the conflict identifier. For new conflict patterns, the system analyzes through the rule inference engine, evaluates the joint angle limit, torque distribution conflict and cable tension safety boundary constraints, and determines the conflict nature. Each conflict identifier is accompanied by a detailed description, including conflict type, impact range and resolution suggestion level.

[0111] The system calculates the theoretical shortest execution time for each action unit according to the dynamic response characteristics of the manipulator. When actually allocating, considering the execution margin of the previous action and the buffer demand of the subsequent action, set a reasonable time margin. Time window allocation algorithm ensures that high-priority actions get enough time resources, while avoiding excessive delay of low-priority actions. For critical precision actions, the system automatically extends the execution time window to ensure positioning accuracy; while for non-critical transition actions, the time window is compressed to improve overall efficiency.

[0112] When detecting action conflicts or receiving new adjustment instructions, the system first tries local adjustment strategy, only modifying the queue order near the conflict action. When local adjustment cannot solve the problem, global rearrangement is started. The history record of completed actions is preserved during the rearrangement process to ensure the coherence of the new sequence with the executed actions. Virtual simulation is performed after each rearrangement to verify the feasibility of the new sequence and prevent secondary conflicts. The rearrangement decision is accompanied by a change impact evaluation report for the operator's reference.

[0113] Before the system converts the scheduling results into device-level control instructions, it performs inverse kinematics verification to ensure that all target positions are within the manipulator's workspace. Dynamics verification checks whether the joint torque requirement is within the motor output range to avoid overload risk. Cable state verification evaluates the bending radius and tension change of the cable during handling to prevent damage to the cable insulation layer. After all verifications pass, the instruction compiler converts the high-level action description into low-level control signals, including the target angle of each joint, motion curve parameters and safety monitoring thresholds.

[0114] Control commands are encapsulated as timestamp-synchronized data packets and transmitted to joint controllers through deterministic networks. Each data packet contains target states, allowed errors, and emergency stop conditions. The transmission system implements a redundancy check mechanism that automatically triggers retransmission when data anomalies are detected. The receiving controllers strictly execute commands in timestamp order to ensure the precise synchronization of multi-axis motion. The system reserves a command buffer to maintain basic motion control during short network interruptions. Joint sensors provide actual position, velocity, and torque data, which are compared with expected states. When deviations exceed thresholds, the monitoring system immediately sends an exception report to the handling management module. The system distinguishes between temporary deviations and trend deviations, initiating fine-tuning compensation for temporary deviations and triggering scheduling reevaluation for trend deviations. Monitoring data are also used to update the robot performance model, making subsequent scheduling decisions more consistent with the actual state of the device.

[0115] When receiving an emergency stop signal or detecting a major anomaly, the system immediately freezes the current action queue and starts an interruption handling program. The program first controls the robot to enter a safe holding state to avoid cable falling or mechanical damage. Then it analyzes the interruption cause and develops a recovery plan. For recoverable interruptions, the system generates an optimized sequence to continue execution from the breakpoint; for non-recoverable interruptions, it plans a safe return path. All operations of interruption events are recorded in detail to form a complete accident report.

[0116] The scheduling decision support system accumulates historical operation data. After each task is completed, the system analyzes the differences between the actual execution trajectory and the planned sequence to evaluate the effectiveness of the scheduling strategy. The scheduling parameter self-learning module adjusts key parameters such as sorting weights and time window allocation proportions based on the evaluation results. For repetitive tasks, the system establishes optimal scheduling templates, significantly improving the planning efficiency of subsequent tasks. Long-term operation data are also used to identify performance degradation trends of the robot, providing a basis for preventive maintenance.

[0117] Operators can view the current action queue, conflict identification distribution, and execution progress through a graphical interface. The interface supports manual adjustment of priorities, modification of time allocation, or insertion of custom actions. All manual intervention operations require double confirmation and are automatically recorded in operation logs. The system provides decision-making assistance information for operators, including conflict resolution suggestions and parameter adjustment impact predictions, supporting efficient human-machine collaborative decision-making.

[0118] Through the above embodiments, the carrying management module builds an efficient and reliable action coordination system. The dynamic priority ranking mechanism flexibly responds to changes in task demand, the intelligent conflict detection timely resolves resource competition risks, and the self-adaptive time allocation optimizes the overall execution efficiency. The multi-level verification ensures the safety and feasibility of the instructions, the strict monitoring ensures the controlled execution process, and the perfect interruption processing maintains the system robustness. These functions work together to enable the manipulator to accurately coordinate various carrying actions and achieve smooth and efficient operation of cable automation.

[0119] The system periodically saves the current action queue state to non-volatile memory, and after unexpected power failure or system restart, the task execution can be restored from the latest saved point. The stored data includes the complete queue structure, executed action marks and to-be-executed action parameters, ensuring the consistency of the task state after recovery. The storage interval is dynamically adjusted according to the task criticality, and the storage frequency is increased during important stages. The carrying management module receives task instructions from the factory scheduling system and feeds back the task status after completing cable carrying. The interface uses standard industrial communication protocols to support data exchange with material management systems and quality monitoring systems. When abnormal incoming materials or incomplete downstream process preparation are detected, the system automatically adjusts the carrying pace to achieve cross-process coordination.

[0120] The system continuously monitors environmental parameters such as temperature and humidity, and automatically adjusts action parameters when environmental changes affect the performance of the manipulator. For example, in a high-temperature environment, the cooling action time window is extended, and in a humid environment, the cable swing threshold is reduced. The association model between environmental parameters and scheduling decisions is continuously optimized through machine learning to improve the system's adaptability. For common carrying scenarios such as straight-line carrying, curved routing, and precise positioning, the system preloads optimized action sequence templates. When similar scenarios are identified, the corresponding template is automatically called as the basis for scheduling, greatly reducing planning time. The template supports parameterized customization, allowing action parameters to be adjusted according to specific task requirements. The library content is updated regularly to absorb excellent scheduling cases from actual operation.

[0121] It should be noted that, in this document, the terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device.

[0122] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and adaptions can be effected without departing from the spirit and scope of the present application, which is not limited to the exact construction and arrangement described. It is intended, therefore, to cover all modifications and adaptions that fall within the scope of the claims and their equivalents.

Claims

1. A robotic cable handling positioning system based on a mechanical hand, characterized in that, The method comprises the following steps: a data acquisition module is used to acquire position data and carrying state data in real time during the process of carrying the cable by the manipulator; a state analysis module is used to analyze the deviation of the actual position of the cable from the target position based on the position data, compare the deviation with a preset position range, and generate a position abnormal signal if the deviation exceeds the preset position range; an influence evaluation module is used to monitor the moving speed data of the manipulator based on the position abnormal signal, acquire the speed data within a monitoring period, calculate a moving influence value, compare the moving influence value with a moving influence threshold, and generate an influence size signal; if the moving influence value is greater than or equal to the moving influence threshold, an influence large signal is generated; a synchronous correlation evaluation module is used to acquire a speed change trend value and a position change similarity value based on the influence large signal, perform correlation calculation on the speed change trend value and the position change similarity value, and obtain a synchronous correlation value; a position optimization adjustment module is used to acquire a position adjustment coefficient based on the influence large signal, combine the current target position with the position adjustment coefficient, obtain an adjusted position, and complete position adjustment; the synchronous correlation evaluation module is acquired in the following manner: a speed change trend value is identified, which reflects the overall change direction of the speed data within the monitoring period; a position change similarity value is identified, which reflects the matching degree of the position data and the speed data; the speed change trend value and the position change similarity value are correlated to obtain the synchronous correlation value.

2. The robotic cable handling positioning system of claim 1, wherein, The data acquisition module is acquired in the following manner: the carrying path of the manipulator is divided into a plurality of carrying regions, position data in each carrying region is acquired, and comprehensive position values are determined by integrating the position data of the carrying regions; carrying state data in each carrying region is acquired, and a state comprehensive value is generated by fusing the carrying state data.

3. The robotic cable handling positioning system of claim 2, wherein, The state analysis module is analyzed in the following manner: based on the comprehensive position values, the difference distance between the actual position and the target position is calculated; the difference distance is compared with the maximum boundary value and the minimum boundary value of the preset position range, respectively; if the difference distance is greater than the maximum boundary value or less than the minimum boundary value, it is determined that the deviation exceeds the preset position range; when the position abnormal signal is generated, the state comprehensive value is combined.

4. The robotic cable handling positioning system of claim 3, wherein, The influence evaluation module is calculated in the following manner: within the monitoring period, the fluctuation characteristics of the speed data are extracted; based on the fluctuation characteristics and the position abnormal signal, the moving influence value is derived; the moving influence value is compared with the upper limit value and the lower limit value of the moving influence threshold; if the moving influence value exceeds the upper limit value or is lower than the lower limit value, the influence large signal is generated.

5. The robot-based cable-handling positioning system of claim 1, wherein, The coefficient of the position optimization adjustment module is acquired in the following manner: the synchronous correlation values corresponding to the influence large signals are summarized; based on the synchronous correlation values, the position adjustment coefficient is derived; the current target position is combined with the position adjustment coefficient, and the adjusted position is output.

6. The robotic cable handling positioning system of claim 5, wherein, Further comprising: a path updating module is used to re-plan the carrying path of the manipulator based on the synchronous correlation value; the planning manner of the path updating module is to extract the synchronous correlation value, map it to the key nodes of the carrying path, and generate update path data.

7. The robotic cable handling positioning system of claim 6, wherein, The mapping manner of the path updating module is: identify the position offset characteristics in the key nodes; based on the position offset characteristics and the update path data, correct the coordinates of the key nodes; Output the corrected carrying path.

8. The robotic cable handling positioning system of claim 7, wherein, Further comprising: A carrying management module, configured to coordinate the carrying operation of the robot hand based on the adjusted position and the corrected carrying path; The coordination manner of the carrying management module is to sort a priority sequence of the carrying operation and assign an action execution time.

9. The robotic cable handling positioning system of claim 8, wherein, The sorting manner of the carrying management module is: Extract an action conflict identifier in the priority sequence; Re-arrange the priority sequence based on the action conflict identifier and the action execution time; Generate a final carrying instruction.

Images