Control parameter reasoning method of laser digital galvanometer scanning system

Abstract

The invention relates to the technical field of laser galvanometers, in particular to a control parameter reasoning method of a laser digital galvanometer scanning system, which comprises the following steps: a preparation stage: acquiring operation parameters of an operation task of a laser galvanometer, creating a fact library, and storing the acquired operation parameters of the operation task of the laser galvanometer based on the fact library, according to the method, the operation parameters of the laser galvanometer are acquired through the operation parameter acquisition equipment of the laser galvanometer, and then the fact library and the knowledge library are created to store the acquired operation parameters of the laser galvanometer and ideal operation parameters of the laser galvanometer; further, based on comprehensive analysis and evaluation of stored parameters, the effectiveness of the operation task of the laser galvanometer is judged, and finally, judgment on the effectiveness of the operation task of the laser galvanometer in the incomplete operation parameter state is achieved on the basis that a judgment result serves as prior data and is matched with the method in the reasoning application stage.

Description

Technical Field

[0001] The invention relates to the technical field of laser galvanometers, and in particular to a control parameter reasoning method for a laser digital galvanometer scanning system. Background Art

[0002] The laser galvanometer scanning system is a high-precision laser control device. It uses a high-speed swinging reflector to accurately control the direction of the laser beam to achieve fast and accurate scanning. It is widely used in laser marking, engraving, cutting and other fields, with the advantages of fast scanning speed, high precision and good stability.

[0003] The invention patent with application number 202110096027.5 discloses a drive control method for a digitally controlled galvanometer motor system for laser marking. The drive control method is implemented in a drive controller of the galvanometer motor system. The drive controller is based on an FPGA processor and adopts an H-bridge inverter topology to drive the galvanometer motor. The drive controller includes a digital control circuit, a power drive circuit, a signal sampling and conversion circuit, a power supply circuit, and a communication circuit. The digital control circuit is based on an FPGA processor and is composed of an FPGA chip and its peripheral circuits to complete the reception of system name signals and the implementation of control algorithms, and output command signals to control the drive circuit and control the action of the galvanometer motor. The power drive circuit includes a drive chip circuit and An H-bridge inverter circuit composed of MOSFET devices, a driving chip circuit receives a PWM control signal output by a digital control circuit, and outputs the PWM control signal to the H-bridge inverter circuit after isolation and power amplification, so as to control the on and off of the MOSFET devices in the H-bridge inverter circuit: the H-bridge inverter circuit is connected to the winding of the motor to realize the power supply control of the motor: the signal sampling and conversion circuit includes a current detection circuit and a position signal measurement circuit, which are respectively composed of corresponding current or position sensors and AD conversion chip circuits, and are responsible for collecting motor current and rotor position information, and performing analog-to-digital conversion so that the digital control circuit completes the calculation; the power supply circuit includes a DC main power supply, various isolated power supply chips and their peripheral circuits, and supplies power to various digital chips and the H-bridge drive circuit.

[0004] The application is about the problem that "domestic galvanometer motor system control implementation mostly relies on analog circuits. Compared with digital control methods, analog circuit control is difficult to apply advanced intelligent algorithms or the cost of implementation is extremely high. Therefore, simple control methods are mostly used in analog circuits, and the control effect is far behind the advanced level abroad. Moreover, the control system of analog circuits has poor portability and requires manual parameter debugging of each set of equipment."

[0005] However, when the laser galvanometer scanning system is used in the fields of laser marking, burning, and 3D printing technology, the operating accuracy of the laser galvanometer determines the quality of the output product. At present, the detection technology for this type of usage scenario often requires complete operating data as a reference, and the detection results have a large delay;

[0006] To this end, we propose a control parameter reasoning method for a laser digital galvanometer scanning system. Summary of the invention

[0007] In view of the above-mentioned shortcomings of the prior art, the present invention provides a control parameter reasoning method for a laser digital galvanometer scanning system, which solves the technical problems raised in the above-mentioned background technology.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0009] A control parameter reasoning method for a laser digital galvanometer scanning system, comprising:

[0010] Preparation stage: Collect the operating parameters of the laser galvanometer operation task, create a fact database, store the collected operating parameters of the laser galvanometer operation task based on the fact database, determine the effectiveness of the laser galvanometer operation task according to the laser galvanometer operation task parameters stored in the fact database, and simultaneously input the laser galvanometer operation task effectiveness determination result into the fact database and bind it with the corresponding laser galvanometer operation task parameters for storage;

[0011] A knowledge base is created, and the user end uploads the ideal operating parameters of the laser galvanometer operating task based on the knowledge base, and the ideal operating parameters of the laser galvanometer operating task are stored in the knowledge base;

[0012] Continuously collect the operating parameters of the laser galvanometer operation task and determine the effectiveness of the corresponding operation task;

[0013] Reasoning application stage: upload the operating parameters of the laser galvanometer under the running state, identify the integrity of the uploaded operating parameters of the laser galvanometer under the running state, and infer the validity of the current laser galvanometer operation task based on the uploaded operating parameters of the laser galvanometer under the running state.

[0014] Furthermore, the operating parameters of the laser galvanometer operation task include: laser wavelength, spot diameter, focal depth, galvanometer angle range, scanning speed, galvanometer acceleration, motor voltage, motor current, and output product image of the operation task;

[0015] The laser wavelength, spot diameter, and focal depth are collected based on a spectrum analyzer, a beam analyzer, and a focusing detector. The galvanometer angle range, scanning speed, and galvanometer acceleration are collected based on distance and speed sensors. The motor voltage and motor current are fed back in real time by the laser galvanometer control motor. The output product image of the running task is collected by an industrial camera.

[0016] When determining the validity of the laser galvanometer operation task, the laser galvanometer operation task operating parameters are retrieved from the fact database, and the determination logic is set. The determination logic is applied in combination with the laser galvanometer operation task operating parameters to determine whether the laser galvanometer operation task is valid.

[0017] Furthermore, in the stage of retrieving the operating parameters of the laser galvanometer operation task, the target time domain for parameter retrieving is set, and the operating parameters of the laser galvanometer operation task of the corresponding time domain are retrieved from the fact database based on the target time domain for parameter retrieving, and the retrieved operating parameters of the laser galvanometer operation task are applied to the determination logic to determine whether the laser galvanometer operation task is valid;

[0018] The validity determination logic of the laser galvanometer operation task is expressed as:

[0019]

[0020] Where: f is the validity judgment value of the laser galvanometer operation task; sim(p finish , p0) is the similarity between the output product image of the running task and the ideal product image in the ideal running parameters of the laser galvanometer running task; L max , L min are the maximum and minimum laser wavelengths in the laser galvanometer operation task parameters; L0 is the ideal laser wavelength; d max d min is the maximum and minimum spot diameter in the laser galvanometer operation task operating parameters; d0 is the ideal spot diameter; l max , l min is the maximum and minimum focal depth in the laser galvanometer operation task operating parameters; l0 is the ideal focal depth; A is the span of the actual galvanometer angle range in the laser galvanometer operation task operating parameters; A0 is the span of the ideal galvanometer angle range; n is the total amount of scanning speed acquisition results in the laser galvanometer operation task operating parameters; V i is the scanning speed corresponding to the i-th group of acquisition results; V0 is the ideal scanning speed; v max is the extreme value of the galvanometer acceleration; v0 is the ideal galvanometer acceleration; It is the maximum and minimum voltage-current ratio in the laser galvanometer operation task operating parameters; is the ideal voltage-to-current ratio; ω1, ω2, ω3, ω4, ω5, ω6, ω7 are weights;

[0021] in, Table pair The sum of weights ω1, ω2, ω3, ω4, ω5, ω6, and ω7 is 1. The value of each weight is customized by the user. After the laser galvanometer operation task validity judgment value f is obtained, the laser galvanometer operation task validity judgment threshold is further set, and the judgment value f is compared with the judgment threshold to determine whether the laser galvanometer operation task is valid.

[0022] Furthermore, the output product image of the running task and the ideal product image are both contour images of the product. In the phase of collecting the output product image of the running task, the contour image of the output product image of the running task is synchronously output based on the Canny edge detection algorithm. The similarity sim(p finish ,p0) is obtained by the following formula:

[0023]

[0024] Where: Output the feature vector of the directional gradient histogram corresponding to the product image and the ideal product image for the running task; is the modulus of the eigenvector; K is the number of low-frequency Fourier coefficients retained; F(p finish ) j 、F(p0) j is the i-th element in the Fourier descriptor feature vector of the running task output product image and the ideal product image; γ is the correction factor;

[0025] Among them, the correction factor γ∈(0,1], and the more regular the image texture in the ideal product image is, the larger the value of the correction factor γ is, and vice versa, the smaller the value of the correction factor γ is.

[0026] Furthermore, when the effectiveness of the laser galvanometer operation task is judged, the output product image of the operation task applied by the logic formula and the ideal product image are both contour images, and the ideal operation parameters of the laser galvanometer operation task applied in the logic formula are all derived from the knowledge base.

[0027] Furthermore, the regularity of the image texture in the ideal product image is evaluated by the following logic:

[0028] The ideal product image is divided into several groups of sub-ideal product image blocks of equal size and specification, and the comprehensive similarity between each group of sub-ideal product image blocks is identified. The higher the similarity, the more regular the texture of the ideal product image is, and vice versa, the less regular the texture of the ideal product image is.

[0029] Furthermore, in the reasoning application stage, the operating parameters of the laser galvanometer operating state uploaded are complete or incomplete operating parameters. When the operating parameters are identified as complete operating parameters, the validity of the current laser galvanometer operating task is determined based on the laser galvanometer operating task validity judgment logic in the preparation stage. When the operating parameters are identified as incomplete operating parameters, the reasoning and judgment of the validity of the laser galvanometer operating task are completed based on the laser galvanometer operating task validity reasoning logic.

[0030] Furthermore, the operating parameter integrity identification logic of the laser galvanometer in the operating state is:

[0031] Traverse the operating parameters of the laser galvanometer in the running state uploaded in the inference application stage, and identify whether the operating parameters include laser wavelength, spot diameter, focal depth, galvanometer angle range, scanning speed, galvanometer acceleration, motor voltage, motor current, and running task output product image. The integrity of the operating parameters of the laser galvanometer in the running state is:

[0032]

[0033] Where: W is the percentage of the integrity of the operating parameters under the operating state of the laser galvanometer; f(·) is the judgment function;

[0034] Among them, the weights ω1, ω2, ω3, ω4, ω5, ω6, and ω7 are consistent with the weight values ​​used when obtaining the laser galvanometer operation task validity judgment value f, and the judgment function f(·) takes a value of 1 or 0. When the parameter in the brackets of the judgment function f(·) is one of the operating parameters of the laser galvanometer under the operating state uploaded in the inference application stage, the judgment function f(·) = 1, otherwise, the judgment function f(·) = 0.

[0035] Furthermore, the reasoning logic of the effectiveness of the laser galvanometer operation task is:

[0036] Identify the type of operating parameters when uploading the laser galvanometer operating state, and the corresponding application weight when calculating the effectiveness of the laser galvanometer operating task in the preparation phase:

[0037]

[0038] Where: f reason is the reasoning result of the effectiveness of the laser galvanometer operation task; u is the total amount of the operating parameters uploaded under the laser galvanometer operation state; α is the operating parameter α; ω α is the corresponding weight of the operating parameter α; is a set of laser galvanometer operation task operation parameters from the laser galvanometer operation task operation parameters stored in the fact base in the preparation stage and the operation parameters corresponding to α;

[0039] in, It represents the value of the minimum absolute value of a set of laser galvanometer operation task operation parameters among the laser galvanometer operation task operation parameters stored in the fact base and corresponding to α.

[0040] Furthermore, the effectiveness reasoning result of the laser galvanometer operation task f reason After the calculation, based on the effective judgment threshold of the running task set in the method preparation stage and f reason Compare and infer to determine the effectiveness of the current laser galvanometer operation task.

[0041] Compared with the known public technology, the technical solution provided by the present invention has the following beneficial effects:

[0042] The present invention provides a control parameter reasoning method for a laser digital galvanometer scanning system. During the execution of the method, the laser galvanometer operating parameters are collected by a laser galvanometer operating parameter collection device, and then a fact base and a knowledge base are created to store the collected laser galvanometer operating parameters and the laser galvanometer ideal operating parameters. Further, based on a comprehensive analysis and evaluation of the stored parameters, the effectiveness of the laser galvanometer operating task is judged. Finally, the judgment result is used as a priori data to cooperate with the reasoning application stage of the method to realize the effectiveness judgment of the laser galvanometer operating task under the incomplete operating parameter state. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 A schematic diagram of the logical structure of a control parameter reasoning method for a laser digital galvanometer scanning system;

[0045] Figure 2 A schematic diagram of a system architecture for expanding the application of the technical solution based on the method of the present invention;

[0046] Figure 3 A schematic diagram of the structure of an inference engine based on the application of the method and technical solution of the present invention;

[0047] Figure 4 A schematic diagram of a RETE network based on the application of the technical solution expanded by the method of the present invention;

[0048] Figure 5 A schematic diagram of a RETE algorithm based on the application of the method expansion technical solution in the present invention;

[0049] Figure 6A schematic diagram of a reverse reasoning tree process for applying the method to expand the technical solution of the present invention;

[0050] Figure 7 This is a schematic diagram of the appearance of a motor drive device and a circuit board of a laser galvanometer scanning system of an expert system digital PID based on the application of the technical solution expanded by the method of the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] The present invention will be further described below in conjunction with the embodiments.

[0053] Embodiment 1:

[0054] A control parameter reasoning method for a laser digital galvanometer scanning system of this embodiment includes:

[0055] Preparation stage:

[0056] Collect the operating parameters of the laser galvanometer operation task, create a fact library, store the collected operating parameters of the laser galvanometer operation task based on the fact library, determine the effectiveness of the laser galvanometer operation task according to the laser galvanometer operation task parameters stored in the fact library, and simultaneously input the results of the determination of the effectiveness of the laser galvanometer operation task into the fact library and bind them to the corresponding laser galvanometer operation task parameters for storage;

[0057] A knowledge base is created, and the user end uploads the ideal operating parameters of the laser galvanometer operating task based on the knowledge base, and the ideal operating parameters of the laser galvanometer operating task are stored in the knowledge base;

[0058] Continuously collect the operating parameters of the laser galvanometer operation task and determine the effectiveness of the corresponding operation task;

[0059] Reasoning application stage:

[0060] Upload the operating parameters of the laser galvanometer under the running state, identify the integrity of the uploaded operating parameters of the laser galvanometer under the running state, and infer the validity of the current laser galvanometer operation task based on the uploaded operating parameters of the laser galvanometer under the running state;

[0061] The regularity of image texture in the ideal product image is evaluated by the following logic:

[0062] The ideal product image is divided into several groups of sub-ideal product image blocks of equal size and specification, and the comprehensive similarity between each group of sub-ideal product image blocks is identified. The higher the similarity, the more regular the texture of the ideal product image is, and vice versa, the less regular the texture of the ideal product image is.

[0063] In the inference application stage, the operating parameters of the laser galvanometer operating state uploaded are complete or incomplete operating parameters. When the operating parameters are identified as complete operating parameters, the validity of the current laser galvanometer operating task is determined based on the laser galvanometer operating task validity determination logic in the preparation stage. When the operating parameters are identified as incomplete operating parameters, the validity of the laser galvanometer operating task is inferred and determined based on the laser galvanometer operating task validity inference logic.

[0064] The logic for identifying the integrity of operating parameters when the laser galvanometer is in operation is as follows:

[0065] Traverse the operating parameters of the laser galvanometer in the running state uploaded in the inference application stage, and identify whether the operating parameters include laser wavelength, spot diameter, focal depth, galvanometer angle range, scanning speed, galvanometer acceleration, motor voltage, motor current, and running task output product image. The integrity of the operating parameters of the laser galvanometer in the running state is:

[0066]

[0067] Where: W is the percentage of the integrity of the operating parameters under the operating state of the laser galvanometer; f(·) is the judgment function;

[0068] Among them, the weights ω1, ω2, ω3, ω4, ω5, ω6, and ω7 are consistent with the weight values ​​used when obtaining the effectiveness judgment value f of the laser galvanometer operation task, and the judgment function f(·) takes a value of 1 or 0. When the parameter in the brackets of the judgment function f(·) is one of the operating parameters of the laser galvanometer under the running state uploaded in the reasoning application stage, the judgment function f(·)=1, otherwise, the judgment function f(·)=0;

[0069] The reasoning logic of the effectiveness of the laser galvanometer operation task is:

[0070] Identify the type of operating parameters when uploading the laser galvanometer operating state, and the corresponding application weight when calculating the effectiveness of the laser galvanometer operating task in the preparation phase:

[0071]

[0072] Where: f reason is the reasoning result of the effectiveness of the laser galvanometer operation task; u is the total amount of the operating parameters uploaded under the laser galvanometer operation state; α is the operating parameter α; ω α is the corresponding weight of the operating parameter α; is a set of laser galvanometer operation task operation parameters from the laser galvanometer operation task operation parameters stored in the fact base in the preparation stage and the operation parameters corresponding to α;

[0073] in, Indicates the value of the minimum absolute value of a set of laser galvanometer operation task operation parameters in the laser galvanometer operation task operation parameters stored in the fact base and corresponding to α;

[0074] The reasoning results of the effectiveness of the laser galvanometer operation task f reason After the calculation, based on the effective judgment threshold of the running task set in the method preparation stage and f reason Compare and infer to determine the effectiveness of the current laser galvanometer operation task.

[0075] In this embodiment, through the execution of the method in the above embodiment, a new effective detection method for operation tasks is provided for the laser galvanometer scanning system during operation, ensuring that the laser galvanometer scanning system can perform production tasks more stably when applied in technical fields such as laser marking, burning, 3D printing, etc.

[0076] In addition, based on the expansion technical solutions and concepts of the system in the above embodiments, see Figure 1-7 As shown:

[0077] Figure 1 It includes six parts: human-computer interaction interface, knowledge base, knowledge acquisition mechanism, comprehensive database, reasoning engine and interpretation mechanism;

[0078] (1) Human-computer interaction interface

[0079] The human-computer interaction interface is the port for users to exchange information with the expert system. The user interaction interface is mainly divided into two categories: one is for ordinary users, who ask questions to the system or set reasoning goals. The interface will display the conclusion after the reasoning is completed and the text description of the conclusion; the other is for domain experts or professional engineers. This type of user can use this port to interact with the system with professional knowledge and increase the professional domain knowledge in the system.

[0080] (2) Interpretation body

[0081] The interpretation mechanism will provide a textual explanation of the system's reasoning process and final reasoning conclusion. The interpretation mechanism is mainly divided into two categories: one is to track the conclusion formation process and preset the interpretation text. This can make the expert system's reasoning process more transparent and increase the credibility of the expert system's information. At the same time, problems in the knowledge base or reasoning engine can be discovered from the explanation of the reasoning results, which is more convenient for the daily maintenance of the expert system.

[0082] (3) Comprehensive database

[0083] The comprehensive database is the memory of the computer when it works with the expert system. It is used to store the relevant data content of the reasoning process, including the initial state of the data, the intermediate process and the reasoning conclusion.

[0084] (4) Inference Engine

[0085] The inference engine is the core module for knowledge reasoning in the system. It matches data with the rule-related content in the knowledge base under the control of the relevant reasoning strategy algorithm based on the current problem conditions or known information. Generally speaking, the inference engine and the knowledge base are independent of each other.

[0086] Reasoning methods can be divided into forward reasoning, reverse reasoning and bidirectional reasoning according to the direction of reasoning; they can be divided into logical reasoning and computational reasoning according to the calculation method in reasoning; they can be divided into monotonic reasoning and non-monotonic reasoning according to the time relationship between the new knowledge content generated in the reasoning process and the original knowledge content; they can be divided into precise reasoning and inprecise reasoning according to the logic of using the rules in the knowledge base in the reasoning process.

[0087] (5) Knowledge Base

[0088] The knowledge base stores the knowledge that has been organized and standardized. It is a module in the expert system used to store relevant knowledge and theoretical knowledge in professional fields. The premise of carrying a knowledge base is to determine a fixed knowledge model, that is, the storage structure of knowledge. In the knowledge base, the standardized form of knowledge can be diverse, including production rules and semantic networks. According to different system adaptation conditions, the most appropriate method is selected for the design requirements of professional knowledge content in different fields, and it is saved in the knowledge base. The quantity and quality of knowledge in the system must determine the performance of the expert system. Generally speaking, the inference engine and the knowledge base are independent of each other.

[0089] (6) Knowledge Acquisition Institutions

[0090] The knowledge acquisition mechanism is the main structure of the expert system to acquire knowledge. In the system, the way to acquire knowledge can be manual, semi-automatic or automatic. The function of the knowledge acquisition mechanism is mainly to operate the content in the knowledge base, including adding, modifying and deleting.

[0091] System architecture design:

[0092] See also Figure 2 As shown in the figure, the system design mainly includes system interface, controller, fact base, rule base and rule engine. The system interface is the main interactive interface between the user and the system, which is used for input and output of information, including various processing of this information. The rule base can also be modified through the system interface;

[0093] The fact base is used to store all known information in the system, including the information content of the initial state, reasoning process and reasoning conclusion. The rule base is used to store the fitting knowledge theory of the control signal and feedback signal of the galvanometer scanning system with different digital PID parameters (KP, Ki and Kd);

[0094] Overview of the inference engine:

[0095] The inference engine is the core module of the expert system. It is the module used to complete the inference in the system and is also called the inference engine. The inference engine generally consists of three parts: the scheduler, the executor, and the consistency coordinator. Depending on the characteristics of the engine, the usage scenarios are also different, and the learning cost is high.

[0096] 3.1 Inference Engine

[0097] See also Figure 3 As shown in the figure, the structure of the inference engine is the knowledge base rules and the fact base (total database) facts. The right part is the structure of the inference engine. The structure on the right mainly consists of three parts:

[0098] (1) Pattern Mather is a scheduler, also known as a pattern matcher. It matches the facts in the system's memory with the contents in the rule base, i.e., the matching algorithm. It pushes the successfully matched rules into the next part.

[0099] (2) Agenda is a consistency coordinator, also known as an agenda. It is mainly used to manage the content pushed from the scheduler and to specify the execution order for these contents. It pushes the content to be executed into the next part.

[0100] (3) Execution Engine is also called the execution engine. It is responsible for executing the relevant content according to the rules pushed from the consistency coordinator, such as generating new facts or modifying certain facts.

[0101] The algorithm used in the design of the inference engine in the above system is the RETE algorithm. The original intention of this algorithm is to use the common parts of the domains between the rules to reduce the rule storage, and at the same time save the temporary results of the matching process to speed up the matching speed.

[0102] (1) Overview of the RETE algorithm (see Figure 4-5 (shown)

[0103] The RETE algorithm can be divided into two parts: rule network construction and runtime execution. The construction of the rule network is the process of building the reasoning network through the algorithm's reasoning process, and the runtime execution is the process of propagating the reasoning object in the network. The rule network is constructed based on the content of the existing rules; runtime execution refers to finding all matching objects and rules through network screening methods, which mainly includes two aspects: structural similarity and time redundancy.

[0104] Structural similarity means that all knowledge and all rules in the expert system's memory are represented in the same way, so the matching mode for all content is the same: Temporal redundancy means that changes in the fact base during the reasoning process have limited impact on the inferred content, that is, only the changed content needs to be redefined and matched.

[0105] (2) RETE Network

[0106] The RETE network is mainly divided into two parts: Alpha node network and Beta node network

[0107] 1) The left half is the Alpha node. Push the facts into the Root node, match the set that matches each pattern in the rule, and get the Alpha Memory. The Type node and the Select node are two nodes for filtering facts.

[0108] 2) The right half is the Beta node. The content in Alpha Memory is received through the Dummy node and pushed into the Join node. The function of the Join node is to save the content pushed through the BetaMemory node for integration and filtering and push the conclusion into the Terminal node. Among them, Terminal is the end node, indicating that a single matching rule matches all conditions.

[0109] (3) Application of RETE algorithm in this system

[0110] The facts in the fact base enter the Alpha network from the root node, enter the Alpha node after filtering, and then are pushed into the Beta node.

[0111] ALPHA Node:

[0112] C1: (Amplitude: control signal triangle wave, feedback signal triangle wave)

[0113] C2: (cycle: control signal triangular wave, feedback signal triangular wave)

[0114] C3: (Rise time: control signal triangle wave, feedback signal triangle wave) C4: (Fall time: control signal triangle wave, feedback signal triangle wave) Beta node

[0115] C1, C2: (control signal and feedback signal are both triangular waves)

[0116] C1, C2, C3, C4: (control signal and feedback signal are completely fitted) network graphs generated according to the above points;

[0117] Fact base module design:

[0118] The fact base is one of the important modules of the system. The knowledge representation form in the fact base will directly affect the system operation performance of the knowledge base. In the parameter reasoning system of the laser galvanometer scanning system, the knowledge is usually divided into three parts according to the nature of the knowledge: entity, relationship and rule. The entity refers to the definition of the amplitude, period, rise time and fall time of the triangular wave, the relationship refers to the relationship between the control signal triangular wave and the feedback signal triangular wave, and the rule refers to the error of the control signal triangular wave and the feedback signal triangular wave as a constant value.

[0119] Rule base module design:

[0120] Forward reasoning:

[0121] Forward reasoning is a fact-driven reasoning method. Forward reasoning starts from known facts and continuously generates new facts and adds them to the total database by continuously executing rules before reaching the target conclusion.

[0122] The general form of forward rule reasoning is "IF condition (X) THEN conclusion (Y)", where condition X can be one condition or multiple conditions. When the condition is not a single condition, it is generally connected using the same "AND" or "OR" tag. When the system schedules a rule, it generally determines whether "condition (X)" is completely true, that is, whether the condition can be matched by the current known facts. If so, action 2 in "conclusion (Y)" will be executed. If "condition (X)" is not completely true, the current reasoning level will be terminated and the next rule will be matched.

[0123] The general process of forward reasoning is as follows:

[0124] 1) Match the existing facts in the total database with the premise part in the rule base. 2) Obtain the action Y of the activated rule from the previous one and determine the order of precedence.

[0125] 3) Perform action Y and obtain new knowledge X', and add the new X' to the total database.

[0126] Reverse reasoning:

[0127] See also Figure 6As shown in the figure, reverse reasoning starts from the goal and uses the rules from the "conclusion". The reverse rule assumes that the "conclusion" exists, treats it as an existing fact, and searches the total database for the "condition" that can match this rule. Reverse reasoning is a goal-driven reasoning method.

[0128] The general form of the reverse reasoning rule is "Conclusion (Y) IF THEN Condition (X)", and the condition X can also be a negative number. From a formal point of view, the positions of X and Y in the forward rule reasoning are swapped. In the reasoning process, action Y is regarded as a known condition and matched with objective facts in the system. When the match is successful, action Z related to "Condition (X)" is called, and "Condition (X)" is input into the system as a fact. Therefore, this kind of rule reasoning method is generally called reverse rule reasoning or inverse rule reasoning.

[0129] Design of reverse reasoning tree:

[0130] For a single hypothesis "Assume", it consists of three parts: function part, variable part and conclusion part. The function part refers to the auxiliary function in a certain class of mathematical plane geometry relationship class, and the conclusion part is the target conclusion of this function. The variable part refers to the required input of the function, which is stored in a dictionary type.

[0131] In the process of converting from a sub-rule to a hypothesis, there is an attribute label "SUP" in the sub-rule. In the process of constructing the reasoning tree, when a sub-rule in a rule needs to be converted into a hypothesis, the "SUP" label in the sub-rule will be checked: if the value is TRUE, the sub-rule is normally converted into a hypothesis; if the value is FALSE, the sub-rule is skipped. In a single rule, there may be a sub-rule used to prevent a single rule from being repeatedly matched by several identical facts. This sub-rule used to prevent the matching system from repeatedly matching is not needed when performing reverse reasoning, that is, it does not need to be converted into a hypothesis. Because in the reverse reasoning engine, the controller will check the matching rules. When it is checked that the function content of the current sub-rule is already in the tree structure, and the variable name of the variable used by the function is the same as the variable name of the variable used by the node of the same function in the tree, the rule will be skipped to ensure that the same variable will not be repeatedly matched by the rule. Therefore, a "SUP" label is needed to indicate whether the current sub-rule needs to be used and converted in reverse reasoning.

[0132] For the construction of tree nodes, the main process is as follows:

[0133] (1) The controller finds the matching rule based on Assume. (2) It finds the PROCESS of the When part of the current matching rule, converts several sub-rules (p1, p2, ...) into Assume, and puts them into an Assume Collection. (3) The controller goes through each hypothesis in the Assume Collection one by one, and constructs each hypothesis into a tree node with only the upper half.

[0134] (4) Match the first node above, find a new matching rule from the rule base, and connect its "Then" part to the node below to form a complete node. The "when" part of the rule is connected to "Then below" (5) Execute the content in (4) for all remaining nodes in the current hypothesis set.

[0135] (6) Repeat step (1) for all the terminal “when” parts until the match is complete.

[0136] In the process of constructing the reverse reasoning tree, the rules in the rule base need to be matched multiple times. When using a single hypothesis to match the rules in the rule base, the content in the hypothesis needs to be converted. The construction of the rule is mainly divided into the When part and the Then part. In the PROCESS in the When part, the sub-rule will be constructed into a hypothesis. When using the hypothesis to match the rule, the Then part of the rule needs to be matched instead of the When part, so the matching method needs to be adjusted.

[0137] In the process of converting to hypothesis, this parameter will determine whether to keep this sub-rule. In the matching process, the main thing is to match the content of FUN in the hypothesis. In FUN, it is generally a function of judging the content. In this system, the control signal, feedback signal amplitude, cycle, rise time, and fall time difference functions are mainly compared. In the Then part of the rule, the content of the constant difference value is added to the execution function in the fact library.

[0138] Embodiment 2:

[0139] At the specific implementation level, based on Example 1, this example further specifically describes a control parameter reasoning method for a laser digital galvanometer scanning system in Example 1:

[0140] The operating parameters of the laser galvanometer operation task include: laser wavelength, spot diameter, focal depth, galvanometer angle range, scanning speed, galvanometer acceleration, motor voltage, motor current, and the output product image of the operation task;

[0141] Laser wavelength, spot diameter, and focal depth are collected based on spectrum analyzers, beam analyzers, and focusing detectors. The galvanometer angle range, scanning speed, and galvanometer acceleration are collected based on distance and speed sensors. The motor voltage and motor current are fed back in real time by the laser galvanometer control motor. The images of the running task output products are collected through industrial cameras.

[0142] When determining the validity of the laser galvanometer operation task, retrieve the laser galvanometer operation task operation parameters in the fact database, set the determination logic, and apply the determination logic in combination with the laser galvanometer operation task operation parameters to determine whether the laser galvanometer operation task is valid;

[0143] In the stage of retrieving the operating parameters of the laser galvanometer operation task, the target time domain is set for parameter retrieving, and the operating parameters of the laser galvanometer operation task of the corresponding time domain are retrieved from the fact database based on the target time domain for parameter retrieving. The retrieved operating parameters of the laser galvanometer operation task are applied to the judgment logic to judge whether the laser galvanometer operation task is valid;

[0144] The validity judgment logic of the laser galvanometer operation task is expressed as:

[0145]

[0146] Where: f is the validity judgment value of the laser galvanometer operation task; sim(p finish , p0) is the similarity between the output product image of the running task and the ideal product image in the ideal running parameters of the laser galvanometer running task; L max , L min are the maximum and minimum laser wavelengths in the laser galvanometer operation task parameters; L0 is the ideal laser wavelength; d max d min is the maximum and minimum spot diameter in the laser galvanometer operation task operating parameters; d0 is the ideal spot diameter; l max , l min is the maximum and minimum focal depth in the laser galvanometer operation task operating parameters; l0 is the ideal focal depth; A is the span of the actual galvanometer angle range in the laser galvanometer operation task operating parameters; A0 is the span of the ideal galvanometer angle range; n is the total amount of scanning speed acquisition results in the laser galvanometer operation task operating parameters; V i is the scanning speed corresponding to the i-th group of acquisition results; V0 is the ideal scanning speed; v max is the extreme value of the galvanometer acceleration; v0 is the ideal galvanometer acceleration; It is the maximum and minimum voltage-current ratio in the laser galvanometer operation task operating parameters; is the ideal voltage-to-current ratio; ω1, ω2, ω3, ω4, ω5, ω6, ω7 are weights;

[0147] in, Table pair The sum of weights ω1, ω2, ω3, ω4, ω5, ω6, and ω7 is 1. The value of each weight is customized by the user. After the validity judgment value f of the laser galvanometer operation task is obtained, the validity judgment threshold of the laser galvanometer operation task is further set, and the judgment value f is compared with the judgment threshold to determine whether the laser galvanometer operation task is valid;

[0148] The output product image of the running task and the ideal product image are both contour images of the product. During the acquisition phase of the output product image of the running task, the contour image of the output product image of the running task is synchronously output based on the Canny edge detection algorithm. The similarity sim(p finish ,p0) is obtained by the following formula:

[0149]

[0150] Where: Output the feature vector of the directional gradient histogram corresponding to the product image and the ideal product image for the running task; is the modulus of the eigenvector; K is the number of low-frequency Fourier coefficients retained; F(p finish ) j 、F(p0) j is the i-th element in the Fourier descriptor feature vector of the running task output product image and the ideal product image; γ is the correction factor;

[0151] Among them, the correction factor γ∈(0,1], and the more regular the image texture in the ideal product image is, the larger the value of the correction factor γ is, and vice versa, the smaller the value of the correction factor γ is;

[0152] When executing the judgment operation of the effectiveness of the laser galvanometer operation task, the output product image of the operation task applied by the logic formula and the ideal product image are both contour images, and the ideal operation parameters of the laser galvanometer operation task applied in the logic formula are all derived from the knowledge base.

[0153] In this embodiment, through the above-mentioned settings and the setting of logical formulas, further step execution logic support is provided for the execution of the method in Example 1, ensuring the stable execution of the method in Example 1.

[0154] In summary, during the execution of the method in the above embodiment, the laser galvanometer operating parameters are collected by the laser galvanometer operating parameter collection device, and then a fact base and a knowledge base are created to store the collected laser galvanometer operating parameters and the ideal operating parameters of the laser galvanometer. Further, based on the comprehensive analysis and evaluation of the stored parameters, the effectiveness of the laser galvanometer operating task is judged, and finally the judgment result is used as prior data to cooperate with the reasoning application stage of the method to realize the effectiveness judgment of the laser galvanometer operating task under incomplete operating parameter states.

[0155] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A control parameter inference method for a laser digital galvanometer scanning system, characterized in that: include: Preparation stage: Collect the operating parameters of the laser galvanometer operation task, create a fact library, store the collected operating parameters of the laser galvanometer operation task based on the fact library, determine the effectiveness of the laser galvanometer operation task according to the laser galvanometer operation task parameters stored in the fact library, and simultaneously input the results of the determination of the effectiveness of the laser galvanometer operation task into the fact library and bind them to the corresponding laser galvanometer operation task parameters for storage; Create a knowledge base, and the user end uploads the ideal operating parameters of the laser galvanometer operation task based on the knowledge base, and stores the ideal operating parameters of the laser galvanometer operation task in the knowledge base; Continuously collect the operating parameters of the laser galvanometer operation task and determine the effectiveness of the corresponding operation task; Reasoning application stage: Upload the operating parameters of the laser galvanometer under the running state, identify the integrity of the uploaded operating parameters of the laser galvanometer under the running state, and infer the validity of the current laser galvanometer operation task based on the uploaded operating parameters of the laser galvanometer under the running state.

2. The control parameter inference method of a laser digital galvanometer scanning system according to claim 1, characterized in that: The operating parameters of the laser galvanometer operation task include: laser wavelength, spot diameter, focal depth, galvanometer angle range, scanning speed, galvanometer acceleration, motor voltage, motor current, and the output product image of the operation task; The laser wavelength, spot diameter, and focal depth are collected based on a spectrum analyzer, a beam analyzer, and a focusing detector. The galvanometer angle range, scanning speed, and galvanometer acceleration are collected based on distance and speed sensors. The motor voltage and motor current are fed back in real time by the laser galvanometer control motor. The output product image of the running task is collected by an industrial camera. When determining the validity of the laser galvanometer operation task, the laser galvanometer operation task operating parameters are retrieved from the fact database, and the determination logic is set. The determination logic is applied in combination with the laser galvanometer operation task operating parameters to determine whether the laser galvanometer operation task is valid.

3. The control parameter reasoning method of a laser digital galvanometer scanning system according to claim 2 is characterized in that: In the stage of retrieving the operating parameters of the laser galvanometer operation task, the target time domain is set for parameter retrieving, and the operating parameters of the laser galvanometer operation task of the corresponding time domain are retrieved from the fact database based on the target time domain for parameter retrieving. The retrieved operating parameters of the laser galvanometer operation task are applied to the judgment logic to judge whether the laser galvanometer operation task is valid; The validity determination logic of the laser galvanometer operation task is expressed as: Where: f is the validity judgment value of the laser galvanometer operation task; sim(p finish , p0) is the similarity between the output product image of the running task and the ideal product image in the ideal running parameters of the laser galvanometer running task; L max , L min are the maximum and minimum laser wavelengths in the laser galvanometer operation task parameters; L0 is the ideal laser wavelength; d max ,d min is the maximum and minimum spot diameter in the laser galvanometer operation task operating parameters; d0 is the ideal spot diameter; l max , l min It is the maximum and minimum focal depth in the laser galvanometer operation task operating parameters; l0 is the ideal focal depth; A is the span of the actual galvanometer angle range in the laser galvanometer operation task operating parameters; A0 is the span of the ideal galvanometer angle range; n is the total amount of scanning speed acquisition results in the laser galvanometer operation task operation parameters; V i is the scanning speed corresponding to the i-th group of acquisition results; V0 is the ideal scanning speed; v max is the extreme value of the galvanometer acceleration; v0 is the ideal galvanometer acceleration; It is the maximum and minimum voltage-current ratio in the laser galvanometer operation task operating parameters; is the ideal voltage-to-current ratio; ω1, ω2, ω3, ω4, ω5, ω6, ω7 are weights; in, Table pair The sum of weights ω1, ω2, ω3, ω4, ω5, ω6, and ω7 is 1. The value of each weight is customized by the user. After the laser galvanometer operation task validity judgment value f is obtained, the laser galvanometer operation task validity judgment threshold is further set, and the judgment value f is compared with the judgment threshold to determine whether the laser galvanometer operation task is valid.

4. The control parameter reasoning method of a laser digital galvanometer scanning system according to claim 3 is characterized in that: The output product image of the running task and the ideal product image are both contour images of the product. In the phase of collecting the output product image of the running task, the contour image of the output product image of the running task is synchronously output based on the Canny edge detection algorithm. The similarity sim(p finish ,p0) is obtained by the following formula: Where: Y pfinish , Y p0 Output the feature vector of the directional gradient histogram corresponding to the product image and the ideal product image for the running task; ||Y pfinish ||、||Y p0 || is the modulus of the eigenvector; K is the number of low-frequency Fourier coefficients retained; F(p finish ) j 、F(p0) j is the i-th element in the Fourier descriptor feature vector of the running task output product image and the ideal product image; γ is the correction factor; Among them, the correction factor γ∈(0,1], and the more regular the image texture in the ideal product image is, the larger the value of the correction factor γ is, and vice versa, the smaller the value of the correction factor γ is.

5. The control parameter inference method of a laser digital galvanometer scanning system according to claim 3 or 4, characterized in that: When the validity of the laser galvanometer operation task is judged, the output product image of the operation task applied by the logic formula and the ideal product image are both contour images, and the ideal operation parameters of the laser galvanometer operation task applied in the logic formula are all derived from the knowledge base.

6. The control parameter reasoning method of a laser digital galvanometer scanning system according to claim 1, characterized in that: The regularity of the image texture in the ideal product image is evaluated by the following logic: The ideal product image is divided into several groups of sub-ideal product image blocks of equal size and specification, and the comprehensive similarity between each group of sub-ideal product image blocks is identified. The higher the similarity, the more regular the texture of the ideal product image is, and vice versa, the less regular the texture of the ideal product image is.

7. The control parameter reasoning method of a laser digital galvanometer scanning system according to claim 1, characterized in that: The operating parameters of the laser galvanometer operating state uploaded in the reasoning application stage are complete or incomplete operating parameters. When the operating parameters are identified as complete operating parameters, the validity of the current laser galvanometer operating task is determined based on the laser galvanometer operating task validity judgment logic in the preparation stage. When the operating parameters are identified as incomplete operating parameters, the reasoning and judgment of the validity of the laser galvanometer operating task are completed based on the laser galvanometer operating task validity reasoning logic.

8. The control parameter inference method of a laser digital galvanometer scanning system according to claim 7, characterized in that: The logic for identifying the integrity of the operating parameters in the laser galvanometer operating state is as follows: Traverse the operating parameters of the laser galvanometer in the running state uploaded in the inference application stage, and identify whether the operating parameters include laser wavelength, spot diameter, focal depth, galvanometer angle range, scanning speed, galvanometer acceleration, motor voltage, motor current, and running task output product image. The integrity of the operating parameters of the laser galvanometer in the running state is: Where: W is the percentage of the integrity of the operating parameters of the laser galvanometer under the operating state; f(·) is the judgment function; Among them, the weights ω1, ω2, ω3, ω4, ω5, ω6, and ω7 are consistent with the weight values ​​used when obtaining the laser galvanometer operation task validity judgment value f, and the judgment function f(·) takes a value of 1 or 0. When the parameter in the brackets of the judgment function f(·) is one of the operating parameters of the laser galvanometer under the operating state uploaded in the inference application stage, the judgment function f(·) = 1, otherwise, the judgment function f(·) = 0.

9. The control parameter inference method of a laser digital galvanometer scanning system according to claim 7, characterized in that: The reasoning logic of the effectiveness of the laser galvanometer operation task is: Identify the type of operating parameters when uploading the laser galvanometer operating state, and the corresponding application weight when calculating the effectiveness of the laser galvanometer operating task in the preparation phase: Where: f reason is the reasoning result of the effectiveness of the laser galvanometer operation task; u is the total amount of the uploaded operating parameters under the laser galvanometer operation state; α is the operating parameter α; ω α is the corresponding weight of the operating parameter α; θ α is a set of laser galvanometer operation task operation parameters from the laser galvanometer operation task operation parameters stored in the fact base in the preparation stage and the operation parameters corresponding to α; Among them, |α-θ α | min It represents the value of the minimum absolute value of a set of laser galvanometer operation task operation parameters among the laser galvanometer operation task operation parameters stored in the fact base and corresponding to α.

10. The control parameter reasoning method of a laser digital galvanometer scanning system according to claim 9, characterized in that: The reasoning results of the effectiveness of the laser galvanometer operation task f reason After the calculation, based on the effective judgment threshold of the running task set in the method preparation stage and f reason Compare and infer to determine the effectiveness of the current laser galvanometer operation task.

Images