AGV motion trail optimization method for lens processing

Abstract

The invention belongs to the technical field of vehicle motion control, and particularly discloses an AGV motion trail optimization method for lens processing, which comprises the following steps: acquiring real-time operation data of an AGV and process rule data of lens processing; stability, timeliness and path compliance scores are calculated based on real-time operation and process rule data, and a comprehensive trajectory quality score is obtained through weighted fusion; each score is compared with a preset score threshold value, a single performance qualification and trajectory quality grade judgment result is output, and if it is judged that the trajectory is excellent, the process is ended; carrying out multi-objective decision making based on the qualification and grade judgment result, and generating a trajectory optimization strategy; and according to the trajectory optimization strategy, performing cooperative adjustment and control on the trajectory parameters of the AGV at different levels. According to the invention, multi-dimensional performance perception and intelligent strategy arbitration are realized, and the stability, timeliness and path compliance of lens transportation are guaranteed.

Description

Technical Field

[0001] This invention relates to the field of vehicle motion control technology, and more specifically, to a method for optimizing the motion trajectory of an AGV used in lens processing. Background Technology

[0002] In the field of lens processing and manufacturing, the quality of the transport trajectory of automated guided vehicles (AGVs) directly affects the yield and processing accuracy of lenses. Due to the characteristics of lens materials, such as high brittleness and high precision requirements, vibration and shock during AGV operation, timing accuracy, and path cleanliness all have a significant impact on lens quality.

[0003] In existing technologies, AGV trajectory optimization mainly adopts two technical approaches: The first approach is an optimization method based on a single performance metric. This method independently monitors a key operating parameter of the AGV, and when the parameter exceeds a preset threshold, corresponding single-point adjustment measures are taken. For example, when the vibration sensor detects excessive vibration, the system reduces the AGV's operating speed to decrease vibration; when the running time exceeds expectations, compensation is made by locally adjusting the path or increasing the operating speed. This approach primarily achieves control by establishing a direct mapping relationship between a single performance parameter and its corresponding execution action.

[0004] The second method is a multi-index weighted evaluation method. This method simultaneously monitors multiple operational indicators of the AGV, including vibration, efficiency, and path tracking accuracy. It calculates a comprehensive score by weighting each indicator using preset fixed weighting coefficients. Then, based on the range of this comprehensive score, a unified adjustment strategy is used to optimize the AGV trajectory parameters. This method considers multiple dimensions at the performance evaluation level but employs a unified response mechanism at the decision-making level.

[0005] However, it still has some drawbacks in practical use: 1. Lack of system-level coordination mechanism for multi-objective conflicts makes it difficult to achieve optimal overall performance: Due to the physical coupling and constraints between indicators such as vibration, timeliness and path compliance, existing methods lack the ability to make collaborative decisions when multiple objectives conflict, often leading to a situation of "optimizing one dimension at the expense of others", and failing to achieve the global optimality of system-level performance. 2. The decision-making mechanism lacks dynamic adaptability and the control response efficiency is low: Existing methods mostly adopt fixed weights and uniform response strategies, which cannot dynamically adjust the optimization intensity and execution mode according to the real-time status of the system. This may lead to over-optimization in the case of minor anomalies, while under-optimization in the case of severe anomalies, resulting in waste of control resources or delayed response. 3. Insufficient coordination between optimization decision-making and execution, and lack of dynamic priority transmission mechanism: Existing methods have a gap between optimization decision-making and underlying execution. The execution processes such as motion planning and path planning lack a priority-based dynamic response mechanism and often use fixed constraints, which cannot ensure that key process indicators are given priority in control actions when multiple performance indicators conflict. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, the present invention provides an AGV motion trajectory optimization method for lens processing, which solves the problems mentioned in the background art through the following scheme.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for optimizing the motion trajectory of an AGV used in lens processing, comprising: S1. Multi-dimensional information acquisition: Simultaneously acquire the real-time operation data of the AGV and the process rule data of lens processing. The real-time operation data includes vibration data, time sequence data and path data. The process rule data includes vibration tolerance rules, time sequence constraint rules and cleanliness path rules. S2. Trajectory performance score calculation: Calculate the stability score based on the vibration data and the vibration tolerance rule, calculate the timeliness score based on the time series data and the time series constraint rule, calculate the path compliance score based on the path data and the cleanliness path rule, and then weight and fuse the stability score, timeliness score and path compliance score to obtain a comprehensive trajectory quality score. S3. Trajectory Performance Determination: The stability score, timeliness score, and path compliance score are compared with their preset individual performance score thresholds, and the pass / fail determination results for each individual performance are output. The comprehensive trajectory quality score is also compared with the preset comprehensive trajectory quality score threshold, and the comprehensive trajectory quality level determination result is output. If the comprehensive trajectory quality level determination result is the highest level and all individual performance pass / fail determination results are qualified, the process ends; otherwise, proceed to step S4. S4. Trajectory Strategy Generation: Based on the individual qualification judgment results and the comprehensive trajectory quality level judgment results, multi-objective decision-making is performed to generate a trajectory optimization strategy that includes optimization objectives, execution modes, and priority sequences. S5. Trajectory Strategy Execution: Based on the generated trajectory optimization strategy, the trajectory parameters of the AGV are coordinated and controlled at the servo control, motion planning, and global path planning levels.

[0008] The technical effects and advantages of this invention are as follows: 1. Achieved system-level collaborative optimization of multi-dimensional performance and solved the problem of multi-objective conflicts: This invention achieves system-level collaborative optimization of multi-dimensional performance such as vibration, timeliness, and path compliance by constructing an optimization strategy arbitration mechanism under multi-objective constraints. The system calculates scores for each dimension in parallel based on real-time data and obtains a comprehensive quality evaluation through weighted fusion. When multiple objectives conflict, it identifies the dimensions to be optimized based on the individual qualification, and combines the comprehensive quality level to intelligently decide the optimization mode and priority, technically solving the dilemma of existing methods "paying attention to one thing but losing another" and ensuring optimal overall system performance. 2. A closed-loop control mechanism based on levels and priorities was established, realizing multi-level precise control of trajectory parameters: This invention constructs a closed-loop control architecture for motion trajectories by dynamically associating the comprehensive trajectory quality level with the execution mode, parameter adjustment range, and priority sequence of the optimization strategy. The system decides the optimization mode based on the quality level and generates a priority sequence. At the execution level, each control level dynamically adjusts its control parameters and rule consideration order according to this sequence, realizing precise and on-demand control of trajectory parameters. 3. A priority guarantee system based on process rules was constructed to ensure the rigid constraints of key performance indicators: This invention transforms the process requirements of lens processing into technical decision-making basis by pre-setting priority rules. In the serial optimization mode, the system strictly follows the priority order of "stability > path compliance > timeliness" to ensure that key indicators such as vibration control are given priority. At the same time, the implementation of priority rules is further strengthened by dynamically configuring constraint weights at the execution layer, which technically solves the problem of insufficient guarantee of key indicators. Attached Figure Description

[0009] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0010] Figure 1 The method shown is for optimizing the motion trajectory of an AGV used in lens processing, comprising: S1. Multi-dimensional information acquisition: Simultaneously acquire the real-time operation data of the AGV and the process rule data of lens processing. The real-time operation data includes vibration data, time sequence data and path data. The process rule data includes vibration tolerance rules, time sequence constraint rules and cleanliness path rules. S2. Trajectory performance score calculation: Calculate the stability score based on the vibration data and the vibration tolerance rule, calculate the timeliness score based on the time series data and the time series constraint rule, calculate the path compliance score based on the path data and the cleanliness path rule, and then weight and fuse the stability score, timeliness score and path compliance score to obtain a comprehensive trajectory quality score. S3. Trajectory Performance Determination: The stability score, timeliness score, and path compliance score are compared with their preset individual performance score thresholds, and the pass / fail determination results for each individual performance are output. The comprehensive trajectory quality score is also compared with the preset comprehensive trajectory quality score threshold, and the comprehensive trajectory quality level determination result is output. If the comprehensive trajectory quality level determination result is the highest level and all individual performance pass / fail determination results are qualified, the process ends; otherwise, proceed to step S4. S4. Trajectory Strategy Generation: Based on the individual qualification judgment results and the comprehensive trajectory quality level judgment results, multi-objective decision-making is performed to generate a trajectory optimization strategy that includes optimization objectives, execution modes, and priority sequences. S5. Trajectory Strategy Execution: Based on the generated trajectory optimization strategy, the trajectory parameters of the AGV are coordinated and controlled at the servo control, motion planning, and global path planning levels.

[0011] The system for implementing the method of the present invention includes a central data management and processing module, a multi-dimensional trajectory data acquisition module, a trajectory performance scoring calculation module, a trajectory performance determination module, a trajectory strategy generation module, and a trajectory strategy execution module.

[0012] The system modules are connected as follows: the central data management and processing module is connected to the multi-dimensional trajectory data acquisition module; the multi-dimensional trajectory data acquisition module is connected to the trajectory performance scoring calculation module; the trajectory performance scoring calculation module is connected to the trajectory performance judgment module; the trajectory performance judgment module is connected to the trajectory strategy generation module; and the trajectory strategy generation module is connected to the trajectory strategy execution module.

[0013] The S1, multi-dimensional information acquisition: synchronously acquire the real-time operation data of the AGV and the process rule data of lens processing. The real-time operation data includes vibration data, time sequence data and path data. The process rule data includes vibration tolerance rules, time sequence constraint rules and cleanliness path rules.

[0014] This step, "Multi-dimensional Information Acquisition," serves as the data foundation for the entire trajectory optimization method. The core execution unit is the multi-dimensional trajectory data acquisition module deployed on the AGV, which runs continuously after the AGV starts and throughout the entire task execution. In the actual working environment of the lens processing workshop, this module collects real-time operational data through an integrated sensor network. Simultaneously, it establishes communication with the central data management and processing module, which acts as the system's data hub, and retrieves lens processing process rule data from the module's process rule library. Specifically, this includes: S101. Real-time data acquisition: The multi-dimensional trajectory data acquisition module deployed on the AGV body is responsible for execution. It runs continuously after the AGV starts and throughout the entire task execution. It completes the raw data acquisition through three independent sensor networks: vibration data acquisition uses a three-axis MEMS inertial measurement unit integrated into the AGV body to continuously output the raw values ​​of three-axis acceleration in the carrier coordinate system at a sampling frequency of not less than 100Hz; time-series data acquisition is achieved through the AGV's local clock system and RFID readers deployed at key nodes of the path. When the AGV passes through a preset path point, it automatically records the actual arrival timestamp accurate to milliseconds; path data acquisition is completed by a lidar SLAM system and an encoder. The SLAM system continuously outputs the global path coordinate point sequence, while the encoder outputs wheel speed pulse signals in real time.

[0015] S102. Acquisition of process rule data: The system is centrally managed and controlled by a data management and processing module within its process rule base. During the initialization phase, the system fully loads all rule data from the lens processing technology knowledge base and distributes it to the vehicle-mounted system on demand via an industrial wireless network. Specifically, this includes: vibration tolerance rules stored as numerical pairs of maximum allowable acceleration thresholds and jerk thresholds, which are strictly set based on the brittle characteristics of the lens material; timing constraint rules defined as the planned arrival time window for each path point, stored in a central database table as a time interval data structure; and cleanliness path rules digitally represented by sets of physical area boundary coordinates for different cleanliness levels, organized and managed as a linked list of polygonal vertex coordinates.

[0016] S103, Data Acquisition and Processing: After completing data acquisition, the multi-dimensional trajectory data acquisition module initiates a dedicated data processing flow. For real-time operational data: raw vibration data undergoes noise reduction preprocessing via a Butterworth low-pass digital filter and is then cached in a time-domain sequence buffer; time-series data is verified and normalized using a timestamp alignment algorithm, and the actual running time of the task segment between adjacent path points is calculated based on consecutive timestamps; path data undergoes accuracy optimization by fusing the coordinate sequence output by SLAM and the encoder pulse signal using Kalman filtering technology to obtain the AGV's high-precision position coordinates in the global coordinate system. For process rule data, digital signature verification is used during transmission to ensure integrity, and the received data is loaded into the memory mapping area. Finally, all processed real-time operational data and process rule data are spatiotemporally synchronized and aligned within the module, establishing a complete input dataset for subsequent trajectory performance evaluation.

[0017] S2, trajectory performance score calculation: Based on the vibration data and the vibration tolerance rules, a stability score is calculated; based on the time series data and the time series constraint rules, a timeliness score is calculated; based on the path data and the cleanliness path rules, a path compliance score is calculated; the stability score, timeliness score, and path compliance score are weighted and fused to obtain a comprehensive trajectory quality score.

[0018] The "Track Performance Score Calculation" step is executed by the trajectory performance score calculation module in the system. This module receives real-time operational data and process rule data that have been processed and spatiotemporally synchronized from stage S1. Based on this data, it quantitatively calculates the performance satisfaction level of the current motion trajectory from three dimensions: stability, timeliness, and path compliance. Finally, a comprehensive trajectory quality evaluation index is obtained through weighted fusion. Specifically, this includes: S201. Calculation of stationarity score: The stability score is used to quantify the degree to which the vibration level during AGV operation meets the requirements of the lens processing technology. The calculation process is as follows: 1. Data Input: Obtain the three-axis acceleration time-domain sequence after Butterworth low-pass filtering from S103. And the vibration tolerance rules obtained from S102, including the maximum permissible acceleration threshold. and maximum permissible accelerometer threshold .

[0019] 2. Index Calculation: First, calculate the root mean square value of the triaxial composite acceleration representing the overall vibration intensity within the evaluation period T. : , The evaluation period T is defined as the total time for the AGV to execute a single task segment (i.e., the journey between two adjacent waypoints).

[0020] Secondly, calculate the absolute value of the jerk. And find its maximum value within the same evaluation period T. It is used to assess the intensity of the impact.

[0021] 3. Scoring Quantification: Stability Scoring This is determined by comparing the calculated value with the rule threshold. To ensure the score is within the [0,1] interval, the following formula is used for calculation: , Where a (0≤a≤1) and b are the weighting coefficients of acceleration and jerk in the stability evaluation, respectively. .

[0022] S202, Timeliness score calculation: The timeliness score is used to evaluate the accuracy of the match between the actual AGV running time and the planned process time requirements. The calculation process is as follows: 1. Data Input: Obtain the normalized actual arrival time of each waypoint from S103. Simultaneously, the timing constraint rules, i.e., the planned arrival time windows of each path point, are obtained from S102. .

[0023] 2. Quantitative scoring: Timeliness scoring The time deviations of all N path points are combined. First, the time deviation score of the i-th (i=1,2,...,N) path point is calculated. The calculation is based on the ideal midpoint of the planned time window. Based on this, a piecewise function form is adopted: .

[0024] Then, the overall timeliness score. Take the average score of all waypoints: .

[0025] S203, Calculation of Path Compliance Score: The path compliance score is used to determine whether the AGV's actual travel path strictly adheres to the boundary constraints of the cleanliness zone. The calculation process is as follows: 1. Data Input: Obtain the high-precision position coordinate sequence (X(k), Y(k)) of the AGV after Kalman filtering fusion from S103, where k=1,2,..., and M is the total number of position sampling points within this task segment. Simultaneously, obtain the cleanliness path rules from S102, i.e., the linked list of polygon boundary vertex coordinates for areas of different cleanliness levels.

[0026] 2. Compliance Judgment: Traverse each coordinate point (X(k), Y(k)) in the AGV position coordinate sequence, and use the ray casting method or the point-in-polygon algorithm to determine whether the point is located within its permitted cleanliness area. For example, lenses in transport are only allowed in areas with a specific cleanliness level.

[0027] 3. Quantitative Scoring: Path Compliance Scoring Defined as the number of compliant location points The ratio of the total number of locations M to the total number of locations: .

[0028] S204. Calculation of Comprehensive Trajectory Quality Score: To obtain an overall evaluation of trajectory quality, the three individual scores mentioned above are weighted and fused to obtain a comprehensive trajectory quality score. .

[0029] 1. Weighted fusion: The calculation formula is as follows: , in, These are the weighting coefficients corresponding to the stability score, timeliness score, and path compliance score, respectively. , 2. Weighting: The allocation of weighting coefficients is strictly based on the lens manufacturing process requirements. Given the high brittleness and sensitivity to vibration of the lens material, stability is given a weighting. Typically set to the highest level; path compliance Secondly, to ensure the product is not contaminated; timeliness is a key factor. The efficiency is relatively low, but overall production efficiency still needs to be ensured.

[0030] At this point, step S2 has completed the transformation from multidimensional data to quantitative performance indicators, providing accurate data for subsequent trajectory performance assessment. All calculated scoring indicators... , It will be sent to the trajectory performance assessment module in real time.

[0031] S3, Trajectory Performance Determination: The stability score, timeliness score, and path compliance score are compared with their preset individual performance score thresholds, and the pass / fail determination results for each individual performance are output. The comprehensive trajectory quality score is also compared with the preset comprehensive trajectory quality score threshold, and the comprehensive trajectory quality level determination result is output. If the comprehensive trajectory quality level determination result is the highest level and all individual performance pass / fail determination results are qualified, the process ends; otherwise, proceed to step S4.

[0032] This step, "Track Performance Determination," is performed by the trajectory performance determination module in the system. This module receives all scoring metrics calculated from stage S2—including the stationarity score. Timeliness score Path compliance score and comprehensive trajectory quality score Based on a preset scoring threshold system, the system accurately judges the passability and quality of various aspects of the current trajectory, and its output results are the key basis for determining the subsequent workflow of the system.

[0033] S301. Data Preparation and Input: The trajectory performance assessment module receives four core inputs in real time from the trajectory performance scoring calculation module: , , , Simultaneously, the module internally stores or retrieves preset judgment criteria from the central data management and processing module, including: Individual performance scoring threshold: stationarity threshold Timeliness threshold Path compliance threshold These thresholds are rigid standards for judging whether an AGV is "passing" in a single performance dimension. Their setting is directly derived from the lens processing rules obtained in S102, which is a manifestation of transforming process requirements into calculable criteria. Comprehensive trajectory quality score threshold range: Multiple comprehensive trajectory quality score thresholds used to classify quality levels. These thresholds are used to classify the overall quality level of the trajectory, and their setting is based on a comprehensive consideration of historical operation data statistical analysis and system optimization goals.

[0034] S302, Single-item performance qualification judgment: This process compares the performance score for each dimension with the corresponding individual performance score threshold, outputting a binary pass / fail conclusion, which provides a basis for step S4 to identify specific optimization directions: Stability compliance determination: If If the stability is satisfactory, then the stationarity is considered "qualified"; otherwise, it is considered "unqualified". Timeliness and compliance determination: If If the timeliness is satisfactory, the timeliness is deemed "acceptable"; otherwise, it is deemed "unacceptable". Path compliance and qualification determination: If If the path is deemed compliant, it is considered "compliant"; otherwise, it is considered "incompatible".

[0035] S303, Comprehensive Trajectory Quality Level Determination: This process is based on a comprehensive trajectory quality score. The system combines the overall trajectory quality score threshold range and outputs a discrete quality level to macroscopically reflect the overall quality level of the trajectory, providing a basis for the S4 step to determine the strength of the optimization mode. The decision logic is as follows: like If the overall trajectory quality level is determined to be "excellent"; like If the overall trajectory quality level is determined to be "good"; like If the overall trajectory quality level is determined to be "medium"; like If the overall trajectory quality level is poor, then the overall trajectory quality level is determined to be "poor".

[0036] S304, Process Decisions: If the overall trajectory quality level is "excellent" and all individual performance qualification results are "qualified", then the current trajectory is deemed to meet the requirements and the current control loop ends; otherwise, step S4 is immediately triggered.

[0037] At this point, step S3 completes the transformation from quantified performance indicators to process decisions. All judgment results and decision instructions will be sent to the trajectory strategy generation module in real time as direct input for its strategy arbitration.

[0038] S4, Trajectory Strategy Generation: Based on the individual qualification judgment results and the comprehensive trajectory quality level judgment results, multi-objective decision-making is performed to generate a trajectory optimization strategy that includes optimization objectives, execution modes, and priority sequences.

[0039] The "Track Strategy Generation" step is executed by the trajectory strategy generation module in the system. This module receives all outputs from the S3 trajectory performance evaluation module—namely, the pass / fail results for each individual performance metric and the overall trajectory quality level evaluation result. Based on these inputs, the module arbitrates optimization strategies under multi-objective constraints according to preset rules. Its core tasks are to identify the dimension to be optimized, select the optimization mode, and determine the allocation priority of optimization resources, ultimately generating a specific and executable trajectory optimization strategy, which is then output to the trajectory strategy execution module.

[0040] S401. Data Input and Identification of Dimensions to be Optimized: The trajectory strategy generation module receives the judgment results from S3 in real time. First, based on the individual pass / fail judgment results, the module accurately identifies the set of performance dimensions that need optimization. The identification logic is as follows: If the stationarity is deemed "unqualified", then stationarity will be marked as a dimension to be optimized. If the timeliness is deemed "unqualified", then timeliness will be marked as a dimension to be optimized. If the path compliance is deemed "unqualified", then path compliance will be marked as a dimension to be optimized.

[0041] The output of this step is a clear list of dimensions to be optimized, providing specific optimization targets for subsequent strategy decisions.

[0042] S402, Arbitration of Optimization Strategies under Multi-Objective Constraints: This process is the core of step S4, which takes into account both the list of dimensions to be optimized and the overall trajectory quality level, and makes two-level decisions: first, it determines the macro mode of optimization execution, and then it specifies strict priorities for the serialization mode.

[0043] 1. Optimization Execution Mode Decision: Based on the overall trajectory quality level, the following optimization modes are adopted: Parallel Optimization Mode: When the overall trajectory quality level is "Good," it indicates that the overall system state is acceptable, and the existing problems are relatively independent or minor. In this mode, the system employs parallel optimization for all identified dimensions to be optimized. In this mode, optimization strategies for each dimension are generated and distributed simultaneously, and optimization actions are executed concurrently, aiming to quickly and synchronously repair multiple minor defects and maximize efficiency.

[0044] Serialization fusion optimization mode: When the overall trajectory quality level is "medium" or "poor," it indicates severe performance degradation or strong conflicts across multiple dimensions. To avoid resource dispersion and interference between strategies, the system adopts a serialization fusion optimization mode based on preset priorities. In this mode, system control is centralized, and each dimension to be optimized is processed sequentially according to a clear and unchangeable priority order.

[0045] 2. Priority Arbitration Optimization: When the decision enters the "Serial Fusion Optimization Mode," the system strictly follows preset priority rules to sort the list of dimensions to be optimized. These rules directly reflect the highest technological principles of lens manufacturing: Stability Optimization > Path Compliance Optimization > Timeliness Optimization. This means that when stability needs optimization, it always has the highest priority, and the system will concentrate all resources to prioritize resolving vibration issues to ensure lens safety. Path compliance issues will only be addressed after stability is met or no optimization is required. Timeliness optimization is performed last. This ensures that the most critical performance indicators are absolutely guaranteed when resources are limited.

[0046] S403, Trajectory Optimization Strategy Generation: Based on the arbitration and decision-making results described above, the module generates the final trajectory optimization strategy. This strategy is a structured instruction set, containing the following core elements: Optimization goal: A clearly defined list of dimensions to be optimized; Execution mode: Specifies whether execution is "parallel" or "serialized"; Priority sequence: If serialization mode is used, an optimized order after arbitration is provided; Optimization intensity parameter: This parameter is dynamically correlated with the overall trajectory quality level determined in S3. The worse the level, the larger the optimization intensity parameter value, which means that the parameter adjustment range and the constraints are more stringent in the S5 execution phase, in order to achieve a more aggressive and thorough optimization effect.

[0047] S5, Trajectory Strategy Execution: Based on the generated trajectory optimization strategy, the trajectory parameters of the AGV are coordinated and controlled at the servo control, motion planning, and global path planning levels.

[0048] The "Track Strategy Execution" step is executed by the Track Strategy Execution module within the system. This module receives the structured trajectory optimization strategy output from the S4 Track Strategy Generation module. Its core task is to parse and map the optimization strategy to specific levels of the AGV motion control system. By coordinating the actuators at different control layers, it jointly completes the dynamic optimization of the trajectory. This process ensures that the optimization commands are accurately implemented, achieving closed-loop control and real-time correction of the AGV's motion trajectory.

[0049] S501, Strategy Resolution and Mapping: Upon receiving the trajectory optimization strategy, the trajectory strategy execution module first parses it, extracting key instruction elements, including: optimization objective, execution mode, priority sequence, and optimization intensity parameters. Then, based on the inherent attributes of the optimization objective, the module intelligently maps it to the most suitable control level within the AGV motion control system. Stability optimization is mainly mapped to the servo control level and is executed by directly suppressing vibration by adjusting the underlying drive parameters. Path compliance optimization is mainly mapped to the global path level for execution, ensuring regional compliance by replanning macro-paths; Timeliness optimization is mainly mapped to the motion planning level and executed, and time deviations are compensated by adjusting local trajectories.

[0050] This mapping ensures a precise correspondence between "what problem, at what level, and what method to use to solve it".

[0051] S502, Servo Control Level Optimized Execution: When the optimization strategy includes a stability optimization objective, optimization is performed at the servo control level based on the mapping results of S501. This level achieves vibration suppression by adjusting the underlying servo control parameters: according to the optimization intensity parameters in the strategy, the proportional gain and derivative gain of the servo loop are adjusted according to a preset level relationship. The parameter adjustment range is related to the overall trajectory quality level; when the level is "poor," a larger adjustment range is used to strongly suppress mechanical resonance. Simultaneously, jerk constraints are applied to the current loop or velocity loop to strictly limit the rate of acceleration change. The constraint threshold is negatively correlated with the optimization intensity parameters, thereby reducing the impact at its source. The optimization command is issued from the trajectory strategy execution module to the motion controller, which adjusts the parameters in real time and monitors the motor feedback, forming a high-frequency vibration suppression closed loop.

[0052] S503, Motion Planning Level Optimization Execution: When the optimization strategy includes timeliness optimization or path compliance optimization requiring local obstacle avoidance, optimization is performed at the motion planning level. This level achieves online adjustment of local trajectories through a real-time rolling trajectory adjustment method: This method dynamically sets the strength of various constraints for trajectory adjustment based on the optimization priorities provided in the S4 strategy. For example, when stability is the highest priority, acceleration and jerk constraints are given higher weights to avoid introducing excessive vibrations in pursuit of timeliness. Based on this, a smooth, collision-free local trajectory sequence that conforms to the current dynamic constraints is calculated in real time within a limited look-ahead period based on the AGV's current position, speed status, and local environmental information, and the corresponding speed and acceleration control commands are directly generated and sent to the underlying servo system for execution.

[0053] S504, Global Path-Level Optimization Execution: When the optimization strategy includes macro-level path compliance optimization, optimization is performed at the global path level. This level achieves global path replanning by implementing a process rule-oriented path replanning method: First, based on the optimization priority in the S4 strategy, the priority of each process rule in path selection is directly determined. When path compliance has a high priority, its cost weight is significantly increased, making the algorithm tend to generate fully compliant paths. Subsequently, based on the latest cleanliness path rules and map information, a new path from the starting point to the ending point that meets the priority requirements is quickly generated globally, and this path is sent to the motion planning level as a new global trajectory reference.

[0054] S505, Multi-level Collaboration and Status Feedback: In parallel optimization mode, the optimization actions involved in S502, S503, and S504 are executed concurrently at their respective levels. In serialized fusion optimization mode, the system strictly follows the priority sequence provided by S4, initiating the optimization of the next dimension only after the optimization objective of the previous dimension is achieved. After completing the optimization action, all executors at each level report their execution status back to the trajectory strategy execution module. This module monitors and coordinates the entire execution process and summarizes the execution results of this optimization cycle, feeding them back to the central data management and processing module for system logging and learning.

[0055] Thus, step S5 successfully decomposed and executed the upper-level optimization strategy through the servo control level, motion planning level, and global path level of the AGV motion control system, realizing the complete transformation from intelligent decision-making to physical motion, and forming the final closed loop of the entire trajectory optimization method.

[0056] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other. In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An AGV motion trajectory optimization method for lens processing, characterized in that, The method comprises the following steps: S1, multi-dimensional information acquisition: synchronously acquiring real-time running data of the AGV and process rule data of lens processing, the real-time running data comprising vibration data, timing data and path data, the process rule data comprising vibration allowance rule, timing constraint rule and cleanliness path rule; S2, trajectory performance score calculation: calculating smoothness score based on the vibration data and the vibration allowance rule, calculating timeliness score based on the timing data and the timing constraint rule, and calculating path compliance score based on the path data and the cleanliness path rule, and performing weighted fusion on the smoothness score, the timeliness score and the path compliance score to obtain comprehensive trajectory quality score; S3, trajectory performance determination: comparing the smoothness score, the timeliness score and the path compliance score with preset single performance score thresholds respectively, outputting single performance qualification determination results, and comparing the comprehensive trajectory quality score with a preset comprehensive trajectory quality score threshold, outputting comprehensive trajectory quality level determination result, if the comprehensive trajectory quality level determination result is the highest level and all single performance qualification determination results are qualified, ending the process, otherwise, entering step S4; S4, trajectory strategy generation: performing multi-objective decision based on the single performance qualification determination results and the comprehensive trajectory quality level determination result to generate trajectory optimization strategy comprising optimization target, execution mode and priority sequence; S5, trajectory strategy execution: performing coordinated adjustment and control on the trajectory parameters of the AGV according to the generated trajectory optimization strategy at the servo control, motion planning and global path planning levels.

2. The AGV motion trajectory optimization method for lens processing according to claim 1, wherein, The vibration data comprises three-axis acceleration values; the timing data comprises path point actual arrival time stamp and task segment actual running time; and the path data comprises coordinate point sequence of the global path and real-time position coordinates of the AGV.

3. The AGV motion trajectory optimization method for lens processing according to claim 1, wherein, The vibration allowance rule is defined by maximum allowable acceleration threshold and maximum allowable jerk threshold; the timing constraint rule is defined by a planned arrival time window of the path point; and the cleanliness path rule is defined by boundary coordinate set of different cleanliness regions.

4. The AGV motion trajectory optimization method for lens processing according to claim 1, wherein, The calculation formula of the comprehensive trajectory quality score is: ， wherein, is a comprehensive trajectory quality score, are a smoothness score, a timeliness score, and a path compliance score, respectively, are weight coefficients corresponding to the smoothness score, the timeliness score, and the path compliance score, respectively, and .

5. The AGV motion trajectory optimization method for lens processing according to claim 1, wherein, The qualification determination result is a binary conclusion of "qualified" or "unqualified", and the comprehensive trajectory quality level determination result is divided into four discrete levels of "excellent", "good", "medium" and "poor".

6. The AGV motion trajectory optimization method for lens processing according to claim 1, wherein, The arbitration of the optimization strategy under multi-objective constraints in S4 comprises: identifying performance dimensions to be optimized according to the single performance qualification determination results, and determining execution mode and priority of the optimization strategy according to the comprehensive trajectory quality level determination result.

7. The AGV motion trajectory optimization method for lens processing according to claim 6, wherein, The determination of execution mode and priority of the optimization strategy according to the comprehensive trajectory quality level determination result comprises: when the level is "good", adopting parallel optimization mode for the identified performance dimensions to be optimized; and when the level is "medium" or "poor", adopting a sequenced fusion optimization mode based on preset priority.

8. The AGV motion trajectory optimization method for lens processing according to claim 7, wherein, The preset priority rule for the sequenced fusion optimization mode is that smoothness optimization is prior to path compliance optimization, and path compliance optimization is prior to timeliness optimization.

9. The AGV motion trajectory optimization method for lens processing according to claim 1, wherein, In the S5, the trajectory parameters of the AGV are adjusted and controlled in the servo control, motion planning and global path planning levels, specifically including: in the servo control level, the optimization is performed by adjusting the bottom servo control parameters; in the motion planning level, the optimization is performed by adjusting the local trajectory in real time; in the global path level, the optimization is performed by executing the path re-planning guided by the process rules.

10. The AGV motion trajectory optimization method for lens processing according to claim 9, wherein, In the trajectory strategy execution, the adjustment of the bottom servo control parameters includes adjusting the servo gain and applying the jerk constraint, and the parameter adjustment range is associated with the comprehensive trajectory quality level; in the motion planning level, the process of adjusting the local trajectory in real time dynamically sets the trajectory constraint strength according to the optimization priority; In the global path level, the process of path re-planning guided by the process rules determines the consideration order of each process rule when selecting the path according to the optimization priority.

Images