File management robot control system and method

Abstract

This invention discloses a control system and method for a document management robot, belonging to the field of robot operation and control technology. It includes acquiring RFID scan data and visual grayscale image data of the surface area of ​​a document shelf to generate a gap location dataset; updating a preset priority dynamic matrix based on the gap location dataset and a preset document borrowing request queue; adjusting the robot's trajectory parameters based on the priority dynamic matrix and a preset multi-layer potential field model to generate a composite path instruction set; and outputting the composite path instruction set to the robot drive unit for path tracking. This invention employs multi-modal sensor data fusion, dynamic priority matrix updating, and multi-layer potential field model path planning technology, enabling high-precision spatial perception, adaptive task scheduling, and intelligent path planning, significantly improving the operating efficiency and system reliability of the document management robot.

Description

Technical Field

[0001] This invention relates to the field of robot operation and control technology, and in particular to a control system and method for a document management robot. Background Technology

[0002] An archive management robot is an automated device used for intelligent archive management, whose main functions include the storage, retrieval, and distribution of archives. Its core technologies include path planning, task scheduling, and system control. Traditional archive management robots typically use fixed path planning methods for inspecting and operating archive shelves.

[0003] Currently, document management robots typically use a single sensor, such as RFID or a vision sensor, to obtain information about gaps in the surface area of ​​the document shelf, and then combine this information with pre-defined rules and logic to prioritize tasks and plan paths. Task priorities are usually assigned fixed weights based on the submission time of the document borrowing request or the document category.

[0004] However, the measurement results of a single sensor are greatly affected by environmental interference, resulting in low accuracy in gap location identification, which in turn affects the accuracy of path planning. Secondly, the fixed-weight task priority allocation cannot effectively adapt to rapidly changing task requirements and changes in the robot's own state, resulting in low task scheduling efficiency. In addition, traditional path planning methods lack the ability to perceive the health of the file rack structure in real time, and cannot predict the risk of potential gap generation in advance, which significantly reduces the system's adaptability and reliability. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a control system and method for an archive management robot. Employing multimodal sensor data fusion, dynamic priority matrix updating, and multi-layer potential field model path planning techniques, it achieves high-precision spatial perception, adaptive task scheduling, and intelligent path planning, significantly improving the operational efficiency and system reliability of the archive management robot.

[0006] The above objectives can be achieved through the following approach:

[0007] A file management robot control system and method includes acquiring RFID scanning data and visual grayscale image data of the surface area of ​​the file shelf to generate a gap location dataset; updating a preset priority dynamic matrix based on the gap location dataset and a preset file borrowing request queue; adjusting the robot's trajectory parameters based on the priority dynamic matrix and a preset multi-layer potential field model to generate a composite path instruction set; and outputting the composite path instruction set to the robot drive unit to perform path tracking.

[0008] Optionally, the step of acquiring RFID scan data and visual grayscale image data of the surface area of ​​the file shelf to generate a gap location dataset includes: scanning the radio frequency signal of the file shelf node with a dual-band RFID reader to obtain a high-precision gap coordinate sequence; performing edge gradient detection on the visual grayscale image to generate gap contour confidence data; and spatiotemporally aligning the high-precision gap coordinate sequence and the gap contour confidence data to obtain a gap location dataset, wherein the gap location dataset includes a coordinate confidence calibration field.

[0009] Optionally, updating the preset priority dynamic matrix based on the gap location dataset and the preset archive borrowing request queue includes: calculating a primary priority parameter based on a preset timeliness coefficient function; calculating a secondary adjustment coefficient by combining the gap location dataset and the preset archive value weight table; calculating a final priority parameter based on the secondary adjustment coefficient; and mapping the final priority parameter to the third-order index field of the priority dynamic matrix.

[0010] Optionally, the calculation of the secondary adjustment coefficient further includes: obtaining the number of gaps and gap location information in each region based on the gap location dataset; calculating the distance between the gap and the robot's current position based on the gap location information to obtain distance data; and when the number of gaps in the same region exceeds a preset gap threshold, correcting the secondary adjustment coefficient in real time based on the distance data.

[0011] Optionally, generating the composite path instruction set includes: parsing the high-priority task node information in the priority dynamic matrix to generate gradient gravitational field parameters; generating the repulsive force field coverage area based on the coordinate data in the gap location dataset to construct a hybrid potential field space; superimposing a dynamic balance calculation of the gravitational field strength and the repulsive force field strength in the hybrid potential field space to generate composite path parameters containing velocity and acceleration constraints; and generating the composite path instruction set based on the composite path parameters.

[0012] Optionally, the method further includes: acquiring real-time coordinate information of the robot, calculating the distance deviation based on the composite path parameters; acquiring the robot's battery data; when the distance deviation is greater than a preset safety margin, calculating a correction activation judgment value by combining the battery data and the distance deviation; and when the correction activation judgment value is less than a preset correction threshold, correcting the final priority parameter.

[0013] Optionally, the method further includes: real-time acquisition of vibration spectrum data during robot path tracking to generate a frame structure health index; obtaining historical distribution characteristics of the gap location dataset; predicting future gap generation probability parameters based on the frame structure health index and the historical distribution characteristics; and writing the future gap generation probability parameters into the priority dynamic matrix to form a decision pre-optimization link.

[0014] Optionally, the real-time acquisition of vibration spectrum data during robot path tracking to generate a frame structure health index includes: real-time acquisition of vibration spectrum data during robot path tracking, extracting the vibration main frequency and comparing it with a preset reference frequency to generate a first anomaly coefficient; performing sliding window dispersion analysis on the torque fluctuation data during robot movement to generate a second anomaly coefficient; and combining the historical coordinate migration data in the gap location dataset with the first and second anomaly coefficients through weighted fusion to output the frame structure health index.

[0015] Optionally, the method further includes: periodically collecting task response delay data and gap density data for each file shelf area to generate a spatiotemporal hotspot distribution map; generating shelf layout optimization suggestion parameters based on the spatiotemporal hotspot distribution map; and updating the priority dynamic matrix based on the shelf layout optimization suggestion parameters.

[0016] Based on the same inventive concept, the present invention also provides a file management robot control system, the system comprising: a gap data acquisition module, used to acquire RFID scanning data and visual grayscale image data of the surface area of ​​the file shelf, and generate a gap location dataset; a priority generation module, used to update a preset priority dynamic matrix according to the gap location dataset and a preset file borrowing request queue; a path generation module, used to adjust the robot's trajectory parameters based on the priority dynamic matrix and a preset multi-layer potential field model, and generate a composite path instruction set; and an instruction sending module, used to output the composite path instruction set to the robot drive unit for path tracking.

[0017] Compared with the prior art, the present invention has the following advantages:

[0018] 1. This invention integrates dual-band RFID scanning and visual grayscale image data to construct a gap location dataset, achieving high-precision spatial perception, effectively solving the problem of inaccurate identification of hidden gaps by file management robots, and significantly improving the reliability and accuracy of path planning;

[0019] 2. Based on the dynamic priority matrix and multi-factor weight adjustment method, the value of the archives, the urgency of the task, the robot's status and other multi-dimensional information are dynamically coupled to realize the adaptive adjustment of task priority. This effectively solves the problem of task conflict and low efficiency caused by the fixed path planning of traditional archive management robots, and significantly improves the task scheduling efficiency in complex scenarios.

[0020] 3. By adopting a multi-layer potential field model combined with a dynamic weight adjustment algorithm, composite path parameters containing velocity and acceleration constraints are generated to achieve the balance calculation of gravitational and repulsive fields. This effectively solves the local minima trap problem of traditional potential field methods in complex environments and significantly improves the robot's path planning ability and motion control smoothness in dense gap regions.

[0021] 4. By monitoring robot vibration spectrum data and torque fluctuation data in real time, a health index of the frame structure is generated. Combined with the historical distribution characteristics of the gap location dataset, the probability of future gap generation is predicted. A decision pre-optimization link is constructed to realize the transformation from passive response to active prevention, which significantly improves the system's early warning capability for the degradation of the file rack structure and its long-term operational stability.

[0022] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating the file management robot control method according to an embodiment of the present invention.

[0025] Figure 2 This is a schematic diagram of multi-layer potential field path planning according to an embodiment of the present invention.

[0026] Figure 3 This is a schematic diagram of the structure of the file management robot control system according to an embodiment of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Reference Figure 1 One embodiment of the present invention proposes a control method for an archive management robot, which employs multimodal sensor data fusion, dynamic priority matrix update and multi-layer potential field model path planning technology to achieve high-precision spatial perception, adaptive task scheduling and intelligent path planning, thereby significantly improving the operating efficiency and system reliability of the archive management robot.

[0029] The method described in this embodiment specifically includes:

[0030] Acquire RFID scan data and visual grayscale image data of the surface area of ​​the file shelf to generate a dataset of gap locations;

[0031] Based on the gap location dataset and the preset file borrowing request queue, update the preset priority dynamic matrix;

[0032] Based on the priority dynamic matrix and the preset multi-layer potential field model, the robot's trajectory parameters are adjusted to generate a composite path instruction set.

[0033] The composite path instruction set is output to the robot drive unit to perform path tracking.

[0034] Specifically, on the surface area of ​​the file shelf, a dual-band RFID reader scans the radio frequency signals of the file shelf nodes to obtain a high-precision gap coordinate sequence. Simultaneously, a vision system acquires grayscale images of the file shelf and generates confidence data for the gap contour through edge gradient detection. The RFID scan data and visual image data are spatiotemporally aligned and combined with coordinate confidence calibration to generate a gap location dataset. This multimodal data fusion method effectively improves the accuracy of gap location identification on the file shelf, providing reliable spatial foundation data for subsequent path planning. Based on the gap location dataset and combined with the file borrowing request queue, a priority dynamic matrix is ​​dynamically updated through nonlinear calculations of multiple factors (including remaining task deadline, file value weight, gap density, and robot body state). This method considers the urgency of the file task, the value of the file, and the robot's own energy and motion state, ensuring more intelligent and flexible task priority allocation through dynamic weight adjustment. The task node information in the priority dynamic matrix is ​​parsed into gradient gravitational field parameters, combined with the gap location dataset to generate a repulsive force field coverage area, constructing a hybrid potential field space. In this space, the strengths of the gravitational and repulsive fields are dynamically balanced to generate composite path parameters that include velocity and acceleration constraints. This method effectively overcomes the local minima problem of traditional potential field methods in complex environments, ensuring the smoothness of path planning and obstacle avoidance. The generated composite path instruction set is sent to the robot drive unit to control the robot's motion trajectory, enabling it to perform file storage, retrieval, and delivery operations according to the optimized path. Through multimodal data fusion, intelligent task scheduling, and optimized path planning, the task execution efficiency, path planning reliability, and system stability of the file management robot in complex file shelf environments are significantly improved, providing strong technical support for the efficient operation of intelligent file management systems.

[0035] Optionally, the step of acquiring RFID scan data and visual grayscale image data of the surface area of ​​the file shelf to generate a dataset of notch locations includes:

[0036] High-precision gap coordinate sequence is obtained by scanning the radio frequency signal of the file rack node with a dual-band RFID reader;

[0037] Specifically, a dual-band RFID reader is first used to scan the radio frequency signal at the node of the file rack. The low-frequency band is used to locate the physical coordinates of the rack, and the high-frequency band captures the azimuth data of the tag. A high-precision gap coordinate sequence is obtained through a triangulation algorithm. The triangulation algorithm calculates the geometric relationship between the radio frequency signal strength of three adjacent nodes and their coordinate values, and outputs a sequence of data containing X-axis, Y-axis and Z-axis coordinate values.

[0038] Edge gradient detection is performed on the visual grayscale image to generate notch contour confidence data;

[0039] Specifically, edge gradient detection is performed on the visual grayscale image data. Specifically, the set of edge points of the notch contour is obtained by setting a gradient direction threshold filtering method, the matching degree between the tangent direction of each edge point and the preset standard notch shape is calculated, and notch contour confidence data containing coordinate position and confidence value is generated.

[0040] The high-precision gap coordinate sequence and the gap contour confidence data are spatiotemporally aligned and fused to obtain a gap location dataset, which includes a coordinate confidence calibration field.

[0041] Specifically, the high-precision notch coordinate sequence and the notch contour confidence data are spatiotemporally aligned. Spatiotemporal alignment requires that the timestamp deviation between the RFID scanning signal and the image acquisition time be less than 50 milliseconds. After unifying the reference systems of the two types of data using a three-dimensional coordinate system transformation matrix, a weighted optimization function is used to fuse the coordinate data. The weighted optimization function is as follows:

[0042] ,

[0043] in This represents the final fused 3D coordinate values. The RFID coordinate confidence factor is derived from the signal-to-noise ratio (SNR) of the radio frequency signal. The visual coordinate confidence factor is derived from the gradient detection matching score. The original coordinate values ​​representing RFID coordinates The original coordinate values ​​representing visual coordinates are calculated and stored in the gap location dataset along with... and The coordinate confidence calibration field of the product result.

[0044] For example, when the robot performs a notch scan on the file rack numbered A03, the dual-band RFID reader measures signal strengths of -45dBm, -48dBm, and -43dBm at three directional nodes, respectively. Based on a preset RF attenuation rate model, the notch coordinates are calculated to be (2.15m, 3.70m, 1.32m). After the vision system captures a grayscale image, edge gradient detection shows a match of 87% with the standard notch shape, yielding coordinates of (2.18m, 3.72m, 1.30m). After spatiotemporal alignment and weighted calculation, the RFID coordinate confidence factor is 0.8 (signal-to-noise ratio greater than the threshold), and the visual coordinate confidence factor is 0.87 (match degree conversion), obtaining the three-dimensional notch coordinates (2.16m, 3.71m, 1.31m), with a coordinate confidence scaling value of 0.696. The beneficial effects of this validation example are reflected in the effective correction of data errors from dual-band RFID and visual sensors. A confidence calibration mechanism ensures the accuracy of the coordinate system, preventing measurement deviations from a single sensor from interfering with subsequent path planning and providing reliable spatial foundation data for the multi-layer potential field model. By fusing multimodal sensor data to construct 3D gap location information, the complementary characteristics of dual-band RFID positioning and image gradient analysis are utilized to improve spatial perception accuracy. Combined with a confidence calibration mechanism, dynamic error correction is achieved, solving the key problem of inaccurate identification of concealed gaps in existing file management robots. This method not only ensures that the subsequent path planning module obtains accurate obstacle location information but also provides high-quality data support for updating the priority dynamic matrix, effectively improving the robot's navigation efficiency and task execution reliability on complex file shelves.

[0045] Optionally, updating the preset priority dynamic matrix based on the gap location dataset and the preset file borrowing request queue includes:

[0046] The primary priority parameters are calculated based on the preset timeliness coefficient function;

[0047] Specifically, the remaining deadline compression ratio, i.e., the primary priority parameter, is first calculated using the timeliness coefficient function. The timeliness coefficient formula is as follows:

[0048] ,

[0049] in This refers to the compression ratio of the remaining task deadline, i.e., the primary priority parameter. This represents the remaining time from the current time until the task deadline. The total processing time is set for this task, where k is a preset curve adjustment index with a value of 0.5. The primary priority parameter is used to characterize the non-linear relationship between task urgency and time.

[0050] By combining the gap location dataset with a preset archive value weight table, a secondary adjustment coefficient is calculated;

[0051] Specifically, gap density data is extracted from the gap location dataset. Gap density data refers to the statistical value of the number of gaps per unit area. This data is then multiplied by a pre-defined archive value weighting table to obtain the secondary adjustment coefficient. :

[0052] ,

[0053] In the formula, For the weight table weight, To measure the density of gaps, the archive value weight table stores priority coefficients for different archive categories. The archive categories are divided into three classes, A, B, and C, based on the archive's security classification and borrowing frequency. Class A has a weight value of 1.8, Class B 1.2, and Class C 0.6.

[0054] The final priority parameter is calculated based on the secondary adjustment coefficient.

[0055] The final priority parameter is mapped to the third-order index field of the priority dynamic matrix to complete the dynamic matrix update.

[0056] Specifically, the robot's own state parameters are introduced for coefficient correction. The robot's current battery level parameter is processed using a piecewise reduction algorithm. When the battery level is above 80%, the original coefficient value is retained, and a linear reduction factor is applied in the 40%-80% range. :

[0057] ,

[0058] in This displays the real-time battery percentage; a protective reduction is triggered when the battery level drops below 40%. The inertia index is generated by calculating the root mean square value of the three most recent acceleration changes and comparing it with a preset safety threshold to produce an inertia compensation factor. :

[0059] ,

[0060] in The preset inertial safety threshold, This represents the root mean square value of acceleration. The secondary adjustment coefficient includes a linear reduction factor. and inertia compensation factor Finally, regarding the final priority parameter... ,have:

[0061] ,

[0062] The final priority parameter will be mapped to the third-order index field of the priority dynamic matrix. The storage location of the third-order index field is determined by three parts: task type code, rack area number, and time window identifier.

[0063] For example, for a Class A archive task with borrowing request number JY2024-001, the total processing time is... Set to 120 minutes, current remaining time With a time limit of 30 minutes, the task compression ratio, calculated as the remaining task deadline compression ratio (i.e., the primary priority parameter), is 0.5. The target area gap density is 2.5 gaps / m², and the archive value weight is 1.8. Therefore, the secondary adjustment coefficient base is 2.5 × 1.8 = 4.5. The reduction factor calculated when the robot's current battery level is 75% is 0.875, and the measured motion inertia... Assuming a safety threshold The inertia compensation factor is calculated to be 0.8. Therefore, the final priority parameter is 1.575, which, after normalization, becomes 0.68 and is stored in the index field of shelf area B in the dynamic matrix for the second time period. The system can dynamically calculate priorities by comprehensively considering multiple factors such as archival value, environmental state, and robot operating conditions. It ensures timely response to emergency tasks through nonlinear compression ratios and automatically matches the optimal task sequence based on environmental state parameters, avoiding the decision-making blind spot of relying solely on time parameters. By establishing a multi-dimensional priority calculation model, archival value parameters, environmental feature data, and robot body state are dynamically coupled. A nonlinear compression ratio algorithm is used to enhance the responsiveness of emergency tasks, and a dual reduction factor is combined to achieve the optimal balance between energy efficiency and motion safety. This method effectively solves the task conflict problem caused by the fixed path planning of traditional archival robots, realizing adaptive task scheduling based on real-time state, significantly improving the timeliness of retrieving important archives and the stability of system operation, especially demonstrating excellent task coordination capabilities in multi-task concurrency and high-load work scenarios.

[0064] Optionally, the calculation of the secondary adjustment coefficient further includes:

[0065] Based on the gap location dataset, obtain the number and location information of gaps in each region;

[0066] Based on the gap location information, the distance between the gap and the robot's current position is calculated to obtain distance data;

[0067] When the number of gaps in the same area exceeds the preset gap threshold, the secondary adjustment coefficient is adjusted in real time based on the distance data.

[0068] Specifically, the preset gap threshold is 5 gaps per unit area (m²). When the statistical value in the gap location dataset exceeds this threshold, the compensation algorithm is activated to calculate the attenuation rate parameter. The initial value of the attenuation rate parameter is set as follows: The distance is dynamically adjusted based on the real-time distance between the gap and the robot's current position. The distance calculation uses the Euclidean coordinate difference algorithm, and the adjusted attenuation rate parameter is used for this purpose. ,have:

[0069] ,

[0070] in, This represents the real-time distance of the gap relative to the robot's current position. The preset distance attenuation reference value, for example, 10m. The distance sensitivity coefficient, which can be set to 1.5, is used in this formula to reduce the task priority decay rate when the robot approaches dense gap areas and automatically accelerates decay when it moves away. The compensation algorithm reads the latest coordinate information from the gap location dataset every 30 seconds and recalculates the distance matrix. The updated decay rate parameter will be applied to the correction process of the secondary adjustment coefficient. ,have:

[0071] ,

[0072] in 0.1 is the primary and secondary adjustment coefficient, and 0.1 is the time step conversion coefficient.

[0073] For example, when the system detects that the current gap density in area F of the rack is 6.2 gaps / m², exceeding the threshold, a compensation algorithm is triggered. At this time, the robot is located at coordinates (15.5m, 8.2m), the initial value of the attenuation rate parameter is set to 0.3, and the distances of the three nearest gaps are 3.2m, 4.1m, and 2.9m, respectively. The minimum distance of 2.9m is used to calculate the adjusted attenuation rate parameter. The original secondary adjustment coefficient was 3.8, and the new coefficient was obtained after time step conversion. When the robot approaches a high-density gap area, it automatically slows down the priority decay rate to prevent critical tasks from failing prematurely due to excessive environmental complexity. Simultaneously, dynamic distance parameters ensure precise control of the compensation amount, avoiding the adjustment lag problem caused by mechanically fixed compensation values. By introducing a dynamic decay compensation mechanism, the robot can perceive the distribution of gaps around it in real time and finely adjust the priority parameters. A distance-sensitive algorithm is used to establish a correlation model between environmental complexity and task validity, solving the problem of rigid task priority in traditional methods when facing dense gap areas. This method can adaptively balance the relationship between file retrieval efficiency and path planning safety, significantly reducing the number of repetitive path planning attempts caused by sudden changes in the local environment. Especially under high-load operating scenarios, it can significantly improve system stability and task completion rate.

[0074] Optionally, the method for generating a composite path instruction set includes:

[0075] The task node information with preset priority in the priority dynamic matrix is ​​analyzed to generate gradient gravity field parameters;

[0076] Based on the coordinate data in the gap location dataset, the repulsive force field coverage area is generated, and a hybrid potential field space is constructed.

[0077] Dynamic equilibrium calculations of gravitational and repulsive field strengths are superimposed in the mixed potential field space to generate composite path parameters that include velocity and acceleration constraints.

[0078] A composite path instruction set is generated based on the composite path parameters.

[0079] Specifically, the path diagram is as follows: Figure 2 As shown, firstly, the information of the highest priority task node in the third-order index field of the priority dynamic matrix is ​​parsed, and its coordinate position is extracted as the center point of the gradient gravitational field. For the gradient gravitational field strength... ,have:

[0080] ,

[0081] in, This represents the highest priority parameter value in the current dynamic matrix. Let Euclidean distance be the distance from the robot's current position to the target node. To prevent the extremely small value when the distance is zero, it is set to 0.001m. Then, the repulsive force field generated by each gap is calculated based on the three-dimensional coordinate data in the gap location dataset. The formula for calculating the repulsive force field is:

[0082] ,

[0083] in This represents the coordinate confidence calibration field value for the i-th gap. The distance from the gap to the path measurement point. The repulsive stability coefficient can be taken as 0.2 m². When establishing the mixed potential field space, the gravitational field and the repulsive field are vector-summed and superimposed every 50 ms planning period to generate a direction vector field. The potential field equilibrium point is obtained through dynamic equilibrium calculation, and the gravitational field strength weight can be set. Repulsive field weight Establish constraint equations:

[0084] ,

[0085] Solve for the equilibrium path point sequence. Simultaneously, set an upper limit on the robot's velocity parameters. ,in The maximum acceleration of the robot is 1.2 m / s², where s is the length of the dynamic path segment. The acceleration constraint uses a gradient rate of change threshold. To ensure the smoothness of movement, among which It is obtained by adding the gradient components of the gravitational field and the repulsive field.

[0086] For example, when the task node with a priority parameter of 0.93 is located at coordinates (18.6m, 5.3m), and the current robot is located at (15.2m, 4.8m), the gravitational field strength is calculated. The confidence values ​​of 0.68, 0.72, and 0.65 for the three surrounding gaps correspond to a calculated repulsive force of 0.138. (Setting...) hour Total gap distance hour The equilibrium equations yield a path point offset of 0.75m and an azimuth angle of 42 degrees, generating a composite parameter of a velocity command of 1.2m / s and an acceleration of 0.95m / s². A dynamic weighting mechanism actively enhances the gravitational weight as the robot approaches the target to improve endpoint approach efficiency. Combined with the gap confidence metric, the repulsive force strength is quantified, achieving an optimal balance between the important file retrieval task and obstacle avoidance safety, avoiding path oscillations caused by traditional fixed weighting strategies. By establishing a multi-dimensional potential field coupled path planning model, task priorities are quantified into computable gravitational field parameters. A refined repulsive force field is constructed by combining the confidence index of gap data, and the dynamic weighting algorithm adjusts the interaction strength of the two physical fields in real time. This method effectively overcomes the local minima trap problem of traditional potential field methods in complex file shelf environments. The dual constraint mechanism of velocity and acceleration ensures smooth motion control, significantly improving the robot's ability to pass through narrow, densely gaped areas while ensuring timely priority access to high-value files. Especially in multi-objective task scheduling scenarios, it enables collaborative adaptation between path planning and task priority changes, enhancing the overall robustness of the system.

[0087] Optionally, the method further includes:

[0088] Obtain the robot's real-time coordinate information and calculate the distance deviation based on the composite path parameters;

[0089] Obtain the robot's battery level data;

[0090] When the distance deviation is greater than the preset safety margin, a correction activation judgment value is calculated by combining the power data and the distance deviation;

[0091] When the correction activation judgment value is less than the preset correction threshold, the final priority parameter is corrected.

[0092] Specifically, the safety margin is set as The robot's real-time coordinates are obtained through a laser positioning system. Theoretical coordinates in the calculation and composite path parameters Distance deviation :

[0093] .

[0094] when A low-power mode priority reallocation judgment mechanism is activated. Power mode switching is determined based on a combination of battery status and deviation level, calculating a corrected activation judgment value. (The corrected activation judgment value is then used.) ,have:

[0095] ,

[0096] Low-power mode is activated when the corrected activation judgment value is less than the corrected threshold. During priority reallocation, the final priority parameter of the current task node is corrected according to the exponential decay rule as follows:

[0097] ,

[0098] in, This is the final priority parameter after correction. This is a dynamic attenuation coefficient, initially set to 0.1, and iteratively updated according to a preset algorithm as the number of consecutive triggers increases. The iterative function of the preset algorithm can be... Iterative updates are performed, where m represents the number of consecutive out-of-limit occurrences. The adjusted priority parameters are fed back into the priority weight field of the gradient gravity field parameters via the feedback link, updating the gravity field strength and achieving real-time self-correction of path parameters.

[0099] For example, when the robot experiences positioning drift while executing path PATH_0112, the real-time coordinate detection value is (5.32m, 3.18m), and the theoretical coordinate is (5.50m, 3.30m). The calculated distance deviation is 0.216m, exceeding the safety margin of 0.2. At this time, the power supply deteriorates to 55%, and the corrected activation judgment value is calculated. If the correction threshold is 60, and the correction activation judgment value is less than the correction threshold, the low-power mode is activated. The final priority parameter is 0.82, and the final priority parameter after correction by the attenuation formula is 0.833. The priority is increased in reverse to compensate for the positioning deviation. The updated gravitational field parameters drive the path planning module to generate a corrected trajectory. The new path angle offset compensation value is 2.7 degrees, and the speed is adjusted to 85% of the original value. The beneficial effect of this verification example is that the dynamic weight compensation mechanism can intelligently identify the combined working condition of accidental positioning deviation and systemic power shortage. By increasing the priority in reverse, it enhances the gravity weight of the current path, avoiding task interruption caused by traditional direct speed reduction or pause strategies, while optimizing energy allocation to maintain the system's continuous operation capability. By constructing a deviation-power consumption joint decision model, the positioning accuracy and energy status are innovatively coupled for analysis. A dynamic attenuation coefficient is used to realize the elastic adjustment of the priority parameter, forming an adaptive repair mechanism for the path parameters. This method effectively solves the efficiency loss problem caused by the forced termination of tasks or emergency braking of traditional robots under abnormal working conditions. It maintains the continuity of path planning through feedback parameter updates, significantly improving the system's fault tolerance and task completion reliability in complex operating environments. In particular, it exhibits superior resilience and recovery characteristics in scenarios with power supply fluctuations or positioning signal interference.

[0100] Optionally, the method further includes:

[0101] Real-time acquisition of vibration spectrum data during robot path tracking generates structural health indicators;

[0102] Obtain the historical distribution characteristics of the gap location dataset;

[0103] Based on the structural health index and the historical distribution characteristics, predict the probability parameters for future gap generation;

[0104] The probability parameter of future gap generation is written into the extended field of the priority dynamic matrix to form a decision pre-optimization link.

[0105] Specifically, by statistically analyzing the historical distribution characteristics of the gap location dataset, including gap generation density and temporal correlation, a prediction model for the probability of future gap generation is constructed. This prediction model is then combined with current health indicators to calculate the probability parameters of future gap generation in each region over a future time period. :

[0106] ,

[0107] in, for The probability parameter for gap generation at time t. Historical gap density, The normalization coefficient is preset. The future gap generation probability parameter will be used as a dynamic weight to adjust the task priority of each region. The future gap generation probability parameter is written as a new weighting factor into the extended field of the priority dynamic matrix, constructing the following decision pre-optimization link: the future gap generation probability parameter is assigned to the corresponding extended field of the dynamic matrix according to the region, and each field records the gap generation probability in the future time period for that region; combined with the current task priority parameter and the gap generation probability, the weight of each task node is dynamically adjusted to optimize the path planning parameters and achieve preventative path adjustment. For example, if the future gap generation probability parameter for region D is 0.0512, after being written into the extended field of the dynamic matrix, the priority parameter adjustment module adjusts the task weight for the next cycle based on this value, so that the robot prioritizes avoiding regions with a high probability of gap generation when planning its path, thereby reducing potential obstacles and the need for path adjustment.

[0108] By integrating vibration spectrum data and torque fluctuation data from the robot's movement, a structural health index is established. This index, combined with historical gap distribution characteristics, predicts the probability of future gap formation. The predicted parameters are then incorporated into a dynamic task priority matrix to dynamically adjust the robot's path planning and task scheduling strategies. The beneficial effect lies in shifting from passive response to proactive prevention, significantly improving the path planning efficiency and task reliability of the document management robot in complex dynamic environments, enhancing the system's early warning capability for document shelf structural degradation, and thus ensuring the robot's long-term stable operation.

[0109] Optionally, the real-time acquisition of vibration spectrum data during robot path tracking to generate a structural health index includes:

[0110] Vibration spectrum data of the robot during path tracking is collected in real time, the vibration main frequency is extracted and compared with the preset reference frequency to generate the first anomaly coefficient;

[0111] A sliding window dispersion analysis was performed on the torque fluctuation data during robot walking to generate a second anomaly coefficient.

[0112] By combining the historical coordinate migration data in the gap location dataset, and by weighted fusion of the first anomaly coefficient and the second anomaly coefficient, a structural health index is output.

[0113] Specifically, a triaxial vibration sensor pre-installed in the robot's locomotion mechanism is used to collect vibration spectrum data in real time during robot movement via an RS-485 communication interface. The sensor monitors vibration acceleration signals along three axes (X, Y, Z), with a sampling rate set to 1000Hz. After processing the signals through a low-pass filter, the dominant vibration frequency signal is extracted. The collected dominant vibration frequency signal is compared with a preset reference frequency (the vibration frequency of the filing cabinet under normal conditions, with a frequency value of 50Hz), and the frequency deviation is calculated. :

[0114] ,

[0115] in, This is the measured frequency. The reference frequency is used. The first anomaly coefficient is calculated based on the frequency deviation. :

[0116] .

[0117] Torque fluctuation data during robot movement is collected, and the torque dispersion (standard deviation) is calculated using a sliding window technique (window size set to 3 seconds). :

[0118] ,

[0119] Where N is the number of data points within the window. For the i-th torque value, The average torque. The second anomaly coefficient is calculated based on the dispersion. :

[0120] .

[0121] By combining historical coordinate migration data (the historical changes in the gap location over time), a health index is obtained by weighted fusion of two anomaly coefficients. :

[0122] ,

[0123] in, and The weighting coefficients for vibration anomalies and torque anomalies are respectively, for example, they could be... , Based on historical data, this method utilizes multi-dimensional analysis of vibration and torque data to achieve real-time assessment of the structural health of the filing cabinet, providing a reliable basis for gap prediction. Compared with single-factor analysis methods, this method can comprehensively capture early signs of structural degradation, providing a scientific basis for optimizing robot path planning, thereby improving task execution efficiency and overall system reliability.

[0124] Optionally, the method further includes:

[0125] Periodically collect task response delay data and gap density data for each file shelf area to generate a spatiotemporal hotspot distribution map;

[0126] Specifically, task response latency is defined as the time interval from the submission of a file borrowing request to the actual retrieval and delivery of the file to the user by the robot. This process requires precise measurement, and timestamp data can be extracted from the user's borrowing request records. Gap density refers to the number of gaps on the file shelves per unit area. The location of the gaps can be detected by RFID scanning and vision systems, and their coordinate information can be recorded. The unit area can be set to 1 square meter. The data collection frequency should be determined based on the size of the archive and the workload; it is recommended to collect data once per hour, or multiple times within specific time periods each day, to ensure the timeliness and comprehensiveness of the data. The collected task response latency and gap density data need to undergo preliminary processing, including data cleaning (removing outliers), standardization (unifying data formats), and storage (managing using a database or data warehouse). When storing, the timestamp of the collection should be recorded for subsequent analysis. Through statistical analysis of the task response latency and gap density data, it is possible to identify which areas have higher task response latency and higher gap density during which time periods. Statistical methods (such as mean and standard deviation) and machine learning algorithms (such as cluster analysis) can be used to identify hotspot areas. Use professional data visualization tools such as Tableau or Python's matplotlib library to generate spatiotemporal hotspot distribution maps. Different colors or sizes can be used on the map to represent task response latency and gap density in different areas.

[0127] Based on the spatiotemporal hotspot distribution map, generate suggested parameters for rack layout optimization;

[0128] The priority dynamic matrix is ​​updated based on the proposed rack layout optimization parameters.

[0129] Specifically, the submission time of each borrowing request and the time it takes for the robot to complete the task are recorded, and the delay time is calculated. A dual-band RFID reader is used to scan the radio frequency signals of the file shelf nodes to obtain a high-precision gap coordinate sequence. Edge gradient detection is performed on the visual grayscale image to generate gap contour confidence data; after spatiotemporal alignment, the data is fused to generate a gap location dataset. Next, outliers, such as excessively high delay times or unreasonable gap data, are cleaned up, and the data is standardized to unify the units of measurement. Using GIS or visualization tools, the delay time and gap density of different areas are reflected by color or size, and the time dimension can be segmented by hour, day, or week. An optimization algorithm is selected, for example, a genetic algorithm, defining the fitness function as a weighted sum of task response delay and gap density, and setting the objective function to minimize the value of the fitness function. Initial shelf layout parameters are input, and the optimization algorithm is run to obtain the optimal or near-optimal shelf layout scheme. The shelf layout parameters in the optimization suggestions are written into the extended field of the priority dynamic matrix, and the priority parameters of each task node are adjusted according to the new layout to ensure reasonable resource allocation.

[0130] For example, suppose an archive has three areas. After data collection and optimization, it was found that area B has excessively high task response latency and a high density of missing shelves. The optimization suggestion is to move one shelf from area B to area A, generate a new layout, and update the priority dynamic matrix. The optimized shelf layout can reduce task response latency and improve the robot's work efficiency. Through scientific layout adjustments, the number of missing shelves is reduced, avoiding potential safety hazards. The dynamic feedback mechanism enables the system to adapt to changes in the archive and maintain efficient operation.

[0131] Based on the same inventive concept, such as Figure 3 As shown, the present invention also provides a file management robot control system, the system comprising:

[0132] The gap data acquisition module is used to acquire RFID scanning data and visual grayscale image data of the surface area of ​​the file rack, and generate a gap location dataset;

[0133] The priority generation module is used to update the preset priority dynamic matrix based on the gap location dataset and the preset file borrowing request queue.

[0134] The path generation module is used to adjust the robot's trajectory parameters and generate a composite path instruction set based on the priority dynamic matrix and the preset multi-layer potential field model.

[0135] The instruction sending module is used to output the composite path instruction set to the robot drive unit for path tracking.

[0136] It should be noted that the formulas described above, through the principle of dimensional consistency and mathematical standardization methods (such as normalization, dimensionless parameter conversion, or unit system unification), can translate physical quantities with different properties into unitless standard values ​​or parameters that can be superimposed in the same dimension. This eliminates the interference of different dimensions on the computational logic, allowing the formulas to retain the original data distribution characteristics while possessing mathematical rationality and adaptability to objective laws. These are conventional technical methods and will not be elaborated further. The electrical connections between the various units described above do not necessarily represent direct or indirect connections; any indirect connection method is applicable to the embodiments of this invention as long as it achieves the purpose of this invention. The above descriptions are merely exemplary embodiments of this invention and should not be construed as limiting the scope of this invention.

[0137] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A control method for a document management robot, characterized in that, The method includes: The process involves acquiring RFID scan data and visual grayscale image data of the surface area of ​​a filing shelf to generate a gap location dataset. This includes: scanning the radio frequency signals of the filing shelf nodes with a dual-band RFID reader to obtain a high-precision gap coordinate sequence; performing edge gradient detection on the visual grayscale image to generate gap contour confidence data; and aligning the high-precision gap coordinate sequence and the gap contour confidence data spatiotemporally to obtain a gap location dataset, wherein the gap location dataset includes a coordinate confidence calibration field. Based on the gap location dataset and the preset file borrowing request queue, update the preset priority dynamic matrix; including: calculating the primary priority parameter according to the preset timeliness coefficient function; calculating the secondary adjustment coefficient by combining the gap location dataset and the preset file value weight table; calculating the final priority parameter according to the secondary adjustment coefficient; and mapping the final priority parameter to the third-order index field of the priority dynamic matrix. Based on the aforementioned priority dynamic matrix and a preset multi-layer potential field model, the robot's trajectory parameters are adjusted to generate a composite path instruction set. This includes: parsing the high-priority task node information in the priority dynamic matrix to generate gradient gravitational field parameters; generating the repulsive force field coverage area based on the coordinate data in the gap location dataset to construct a hybrid potential field space; superimposing a dynamic balance calculation of the gravitational field strength and the repulsive force field strength in the hybrid potential field space to generate composite path parameters containing velocity and acceleration constraints; and generating a composite path instruction set based on the composite path parameters. The composite path instruction set is output to the robot drive unit to perform path tracking.

2. The document management robot control method according to claim 1, characterized in that, The calculation of the secondary adjustment coefficient also includes: Based on the gap location dataset, obtain the number and location information of gaps in each region; Based on the gap location information, the distance between the gap and the robot's current position is calculated to obtain distance data; When the number of gaps in the same area exceeds the preset gap threshold, the secondary adjustment coefficient is adjusted in real time based on the distance data.

3. The document management robot control method according to claim 1, characterized in that, The method further includes: Obtain the robot's real-time coordinate information and calculate the distance deviation based on the composite path parameters; Obtain the robot's battery level data; When the distance deviation is greater than the preset safety margin, a correction activation judgment value is calculated by combining the power data and the distance deviation; When the correction activation judgment value is less than the preset correction threshold, the final priority parameter is corrected.

4. The document management robot control method according to claim 1, characterized in that, The method further includes: Real-time acquisition of vibration spectrum data during robot path tracking generates structural health indicators; Obtain the historical distribution characteristics of the gap location dataset; Based on the structural health index and the historical distribution characteristics, predict the probability parameters for future gap generation; The probability parameter of future gap generation is written into the priority dynamic matrix to form a decision pre-optimization link.

5. The document management robot control method according to claim 4, characterized in that, The vibration spectrum data collected in real time during robot path tracking is used to generate structural health indicators, including: Vibration spectrum data of the robot during path tracking is collected in real time, the vibration main frequency is extracted and compared with the preset reference frequency to generate the first anomaly coefficient; A sliding window dispersion analysis was performed on the torque fluctuation data during robot walking to generate a second anomaly coefficient. By combining the historical coordinate migration data in the gap location dataset, and by weighted fusion of the first anomaly coefficient and the second anomaly coefficient, a structural health index is output.

6. The document management robot control method according to claim 1, characterized in that, The method further includes: Periodically collect task response delay data and gap density data for each file shelf area to generate a spatiotemporal hotspot distribution map; Based on the spatiotemporal hotspot distribution map, generate suggested parameters for rack layout optimization; The priority dynamic matrix is ​​updated based on the proposed rack layout optimization parameters.

7. A record management robot control system, applied to the record management robot control method as described in any one of claims 1-6, characterized in that, The system includes: The gap data acquisition module is used to acquire RFID scan data and visual grayscale image data of the surface area of ​​the file shelf to generate a gap location dataset. This includes: scanning the radio frequency signals of the file shelf nodes with a dual-band RFID reader to obtain a high-precision gap coordinate sequence; performing edge gradient detection on the visual grayscale image to generate gap contour confidence data; and spatiotemporally aligning the high-precision gap coordinate sequence and the gap contour confidence data to obtain a gap location dataset, which includes a coordinate confidence calibration field. The priority generation module is used to update a preset priority dynamic matrix based on the gap location dataset and a preset archive borrowing request queue. This includes: calculating a primary priority parameter based on a preset timeliness coefficient function; calculating a secondary adjustment coefficient by combining the gap location dataset and a preset archive value weight table; calculating a final priority parameter based on the secondary adjustment coefficient; and mapping the final priority parameter to a third-order index field of the priority dynamic matrix. The path generation module is used to adjust the robot's trajectory parameters and generate a composite path instruction set based on the priority dynamic matrix and a preset multi-layer potential field model. This includes: parsing high-priority task node information in the priority dynamic matrix to generate gradient gravitational field parameters; generating a repulsive force field coverage area based on coordinate data in the gap location dataset to construct a hybrid potential field space; performing dynamic balance calculations of gravitational and repulsive force field strengths in the hybrid potential field space to generate composite path parameters including velocity and acceleration constraints; and generating a composite path instruction set based on the composite path parameters. The instruction sending module is used to output the composite path instruction set to the robot drive unit for path tracking.

Images