Search

CN-121995800-A - Intelligent laser control and remote management method and system based on multi-mode driving

CN121995800ACN 121995800 ACN121995800 ACN 121995800ACN-121995800-A

Abstract

The invention discloses an intelligent control and remote management method and system for a laser based on multi-mode driving, which belong to the technical field of laser control and comprise the steps of collecting operation data of at least two different modes in the operation process of the laser and forming a process characteristic vector, wherein the operation data of the different modes at least comprise any two of laser output power data, retro-reflection data, spectrum data, temperature data, current or voltage ripple data. The multimode operation data in the laser operation process is constructed into a cross-modal physical consistency residual error and forms a process credibility evidence, meanwhile, remote proof or integrity check is carried out on a local control end to form a platform credibility evidence, and a network controllability index is calculated based on network delay and jitter to form a network credibility evidence, so that the three types of evidence can be fused into a credibility score, and the problem of unified constraint that the process state, the platform credibility and the network controllability are difficult to simultaneously quantify in a remote management scene is solved.

Inventors

  • ZHENG LONG
  • YAN ZHENJUN
  • Ou Xianjiao

Assignees

  • 深圳市龙创激光科技有限公司

Dates

Publication Date
20260508
Application Date
20260121

Claims (9)

  1. 1. The intelligent laser control and remote management method based on the multi-mode driving is characterized by comprising the following steps of: S1, acquiring operation data of at least two different modes in the operation process of a laser and forming a process feature vector, wherein the operation data of the different modes at least comprise any two types of laser output power data, retro-reflection data, spectrum data, temperature data, current or voltage ripple data, and synchronously acquiring network transmission state data, and the network transmission state data at least comprises a time delay parameter and a jitter parameter; S2, constructing a cross-modal physical consistency residual error based on the process feature vector, wherein the cross-modal physical consistency residual error is used for representing the deviation degree of different modal operation data under a preset physical constraint relation and outputting process credible evidence; s3, performing remote proof or integrity check on the local control end of the laser to output platform trusted evidence, wherein the platform trusted evidence is used for representing the degree that the software and hardware states of the local control end meet a preset trusted base line; S4, calculating a network controllability index based on the network transmission state data and outputting network credible evidence, wherein the network credible evidence is used for representing the influence degree of a remote control link on the stability of closed-loop control; S5, fusing the process credible evidence, the platform credible evidence and the network credible evidence to obtain a credibility score, and mapping the credibility score into a control authority set, wherein the control authority set at least comprises an allowed control mode set, an allowed instruction set, an allowed waveform set or waveform library subset and an energy input boundary or a power change rate boundary; S6, generating control waveform parameters and forming a laser driving control waveform under the constraint of the control authority set, and embedding a physical watermark sequence in the laser driving control waveform, wherein the physical watermark sequence is realized through time domain perturbation or frequency domain perturbation of a pulse sequence and meets the constraint of not exceeding a preset quality influence threshold.
  2. 2. The intelligent control and remote management method of a laser based on multi-mode driving according to claim 1, further comprising the steps of: s7, executing the laser driving control waveform by the local control end in a hard real-time mode to drive a laser to output, continuously monitoring the reliability score and the network controllability index, entering an autonomous degradation mode when the reliability score is detected to be lower than a threshold value or the network controllability index is detected to be degraded to a threshold value condition, limiting the control waveform to a safe waveform in the waveform library subset, and converging the energy input boundary or the power change rate boundary; And S8, receiving returned multi-mode operation data by a remote management end, detecting the physical watermark sequence based on the returned data to verify the execution consistency of the laser driving control waveform, simultaneously mapping the reliability score, the control authority set, and recording the autonomous degradation triggering event and the physical watermark sequence verification result to form a traceable remote management audit record.
  3. 3. The multi-mode drive based intelligent control and remote management method of lasers of claim 2, wherein said cross-modal physical consistency residual is obtained and used to form said process confidence evidence by: After time alignment and sampling window constraint processing are carried out on laser output power data and retro-reflective data in the process feature vector, a retro-reflective predicted value is generated according to a preset power retro-reflective coupling constraint relation, deviation between the retro-reflective data and the retro-reflective predicted value is used as an instantaneous residual, the instantaneous residual is further recursively accumulated to form a residual trend quantity, the residual trend quantity is mapped into the process credible evidence, so that the process credible evidence is monotonically reduced when the residual trend quantity is increased, and the process credible evidence is used as a leading quantity for convergence constraint on the control authority set in fusion calculation of the credibility score, so that entry conditions of the autonomous degradation mode are triggered in advance.
  4. 4. The intelligent control and remote management method of a laser based on multimode driving according to claim 2, wherein the power back-coupling function is an updatable coupling function, and updating of coupling parameters is only allowed to be triggered under the condition that platform trusted evidence meets a preset trusted base line and a remote management end group is characterized as executing consistency based on verification results of returned multimode operation data on a physical watermark sequence, spectral drift indexes are extracted from spectral data within a sampling window same as the residual trend after being allowed to be triggered, and the spectral drift indexes are written into an updating objective function of the coupling parameters to solve the updated coupling parameters, and the updated coupling parameters are utilized to correct back-reflection predicted values to obtain corrected residual trend amounts and update process trusted evidence accordingly, so that sensitivity of process trusted evidence decrease is improved when the spectral drift indexes are increased and the control authority set is further converged; And locking the coupling parameters and updating the process trusted evidence according to a monotonically decreasing rule when the platform trusted evidence does not meet a preset trusted baseline or the physical watermark sequence verification result is not characterized as executing consistency so as to forcedly converge the control authority set and trigger an autonomous degradation mode.
  5. 5. The intelligent control and remote management method based on multimode driving of laser according to claim 4, wherein the mapping from the credibility score to the control authority set is implemented by a stepwise converged segmentation rule, the segmentation rule sets a first credibility threshold and a second credibility threshold, the first credibility threshold is smaller than the second credibility threshold, and a authority status word is determined according to the comparison result of the credibility score with the first credibility threshold and the second credibility threshold, and the authority status word is used for uniquely determining an allowed control mode set and a control waveform available set corresponding to the allowed control mode set and synchronously determining an energy input boundary and a power change rate boundary; When the credibility score is not higher than the first credibility threshold, the authority status word is put into a strong convergence state so that the control waveform usable set converges to a safe waveform set, the energy input boundary and the power change rate boundary converge to preset minimum boundaries and an autonomous degradation mode is forcedly kept; When the confidence score is above the first confidence threshold and not above the second confidence threshold, the authority status word is put into a convergence state to allow only limited step-size updating of control waveform parameters and to keep the power change rate boundary from exceeding an upper limit corresponding to the network controllability index on the premise of autonomous degradation mode release; When the confidence score is higher than the second confidence threshold, the authority status word is put into a weak convergence state to unlock the call to the set of extension waveforms in the control waveform library and update the energy input boundary and the power change rate boundary according to a monotonically widening rule, so that the control authority set is unlocked step by step as the confidence score increases and is converged step by step as the confidence score decreases.
  6. 6. The multi-mode drive based laser intelligence control and remote management method of claim 5, wherein the updating of the rights state words follows a state transition rule with hysteresis constraints, the state transition rule comprising maintaining a state retention register amount corresponding to the rights state words at a local control end and setting a minimum number of retention windows for each state transition, wherein the transition of the rights state words from a weak to a medium convergence state and from a medium convergence state to a strong convergence state is configured as an instant effect transition that is not characterized as performing a consistency or the reliability score is not higher than the first reliability threshold as a trigger condition and locking the state retention register amount after triggering with the platform reliability proof not meeting a preset reliability baseline or the physical watermark sequence verification result; And the migration of the permission status word from the strong convergence status to the medium convergence status is allowed to trigger only when the platform trusted evidence continuously meets a preset trusted base line and the physical watermark sequence verification results are all characterized as executing consistency within a continuous first minimum holding window number, the migration of the permission status word from the medium convergence status to the weak convergence status is further allowed to trigger only when the network controllability index is not degraded within a continuous second minimum holding window number after the completion of the foregoing allowed triggering and the reliability score is continuously higher than the second reliability threshold, so that the permission unlocking must be subjected to a gradual release path from the medium convergence to the weak convergence and is constrained by the advanced gating of platform trusted and watermark consistency.
  7. 7. The intelligent control and remote management method of a laser based on multimode driving according to claim 6, wherein the physical watermark sequence is determined by a remote management terminal based on a proof random number generation seed bound with a platform credible evidence and synchronously forms a reference watermark sequence, the reference watermark sequence is mapped to a perturbation coding amount of the laser driving control waveform at a local control terminal to perform micro-modulation distributed by a window on a pulse width or a pulse repetition frequency of a pulse sequence, so that energy deviation introduced by the perturbation coding amount is always limited by the preset quality influence threshold; And the remote management end performs synchronous alignment and alignment on the returned retroreflective data under a time reference consistent with the sampling window after receiving the returned retroreflective data, extracts watermark features, obtains watermark confidence through correlation matching with the reference watermark sequence and takes the watermark confidence as quantitative representation of a physical watermark sequence verification result, wherein when the watermark confidence is lower than a preset threshold, the process trusted evidence is updated according to the monotonically decreasing rule, the authority status word is forcedly migrated to a strong convergence state, and the status is locked to keep the storage quantity, and when the watermark confidence continuously meets the preset threshold and meets the minimum keeping window number, the locking is allowed to be released and a step-by-step unlocking path is entered, so that the detectability of the physical watermark sequence becomes a necessary condition for controlling authority unlocking and forms a closed-loop constraint relation with the trusted score.
  8. 8. The intelligent control and remote management system of a laser based on multi-mode driving according to any one of claims 1-7, characterized by comprising a laser, a local control end, a remote management end, and a communication link for connecting the local control end and the remote management end; The local control end is connected with the laser and is configured to collect operation data of at least two types of different modes in the operation process of the laser to form a process feature vector, wherein the operation data of the different modes at least comprise any two types of laser output power data, retro-reflection data, spectrum data, temperature data, current or voltage ripple data, network transmission state data of the communication link are synchronously obtained, and the network transmission state data at least comprise delay parameters and jitter parameters; the local control end is further configured to construct a cross-modal physical consistency residual error based on the process feature vector to output process credible evidence, and perform remote attestation or integrity check on the local control end to output platform credible evidence, and calculate a network controllability index based on the network transmission state data to output network credible evidence; The local control end is further configured to perform fusion calculation on the process trusted evidence, the platform trusted evidence and the network trusted evidence to obtain a reliability score and generate a control authority set according to the reliability score, wherein the control authority set at least comprises an allowed control mode set, an allowed instruction set, an allowed waveform set or waveform library subset and an energy input boundary or a power change rate boundary; the local control end is further configured to generate a laser driving control waveform under the constraint of the control authority set, embed a physical watermark sequence in the laser driving control waveform, and execute the laser driving control waveform in a hard real-time manner to drive the laser to output, and enter an autonomous degradation mode when the reliability score is lower than a threshold value or the network controllability index is degraded to a threshold condition, limit the control waveform to a safe waveform in the waveform library subset, and converge the energy input boundary or the power change rate boundary.
  9. 9. The multi-mode drive-based laser intelligent control and remote management system of claim 8, wherein the local control terminal is configured to solidify the mapping result of the trust score into a rights state word and uniquely define the waveform library subset and the energy input boundary or the power change rate boundary with the rights state word, the rights state word is determined by a first trust threshold and a second trust threshold, the first trust threshold is smaller than the second trust threshold, and form a progressive unlocking path with hysteresis constraint by setting a minimum hold window number, so that the rights state word is immediately validated when migrating to a stronger convergence direction and only when platform trust evidence continuously meets a preset trust baseline and a physical watermark sequence verification result is continuously characterized as performing consistency within a continuous minimum hold window number and a network controllability index is allowed to trigger under the condition that the corresponding window is not degraded; The remote management end is further configured to determine a reference watermark sequence based on a proof random number generation seed bound with the platform trusted evidence and to determine the physical watermark sequence in accordance with the local control end, and obtain watermark confidence based on correlation matching of returned retro-reflective data and the reference watermark sequence and quantitatively characterize the physical watermark sequence verification result with the watermark confidence, so that when the watermark confidence is lower than a preset threshold, the local control end forcedly shifts a permission status word to a strong convergence state and locks the step-by-step unlocking path until the watermark confidence continuously meets the minimum retention window number.

Description

Intelligent laser control and remote management method and system based on multi-mode driving Technical Field The invention relates to the technical field of laser control, in particular to an intelligent laser control and remote management method and system based on multi-mode driving. Background The existing laser generally adopts local closed-loop control in the processes of welding, cutting, additive manufacturing and the like, and adjusts output power or pulse waveform by means of signals such as power, back reflection, spectrum, temperature, power supply ripple waves and the like so as to reduce penetration fluctuation, avoid key hole instability or inhibit heat accumulation; along with the increase of production line digitization and equipment centralized operation and maintenance requirements, more and more schemes introduce remote monitoring and remote parameter issuing, so that a remote terminal can update a waveform library, adjust the upper limit of power, change the change rate of power or issue a process formula, and carry out quality tracing and fault diagnosis through multi-mode data returned by a network; On the one hand, network delay and jitter can lead to closed loop feedback lag and even control command arrival sequence disorder, a remote end is difficult to judge whether a certain waveform is accurately executed in time, and mismatching of energy boundary setting or control failure of power change rate easily occurs, so that process risks such as overshoot, burn-through or expansion of a heat affected zone are caused; on the other hand, the remote control link introduces a problem of platform reliability and execution consistency, for example, a software version, a waveform library file, a parameter table or control logic of a local control end may deviate from a baseline due to abnormal upgrading, configuration drift or malicious tampering, the remote end may be replaced by different waveforms or replayed old instructions locally even if a correct strategy is issued, so that the remote end considers that the execution is not verifiable and inconsistent with the local actual output, the existing scheme is mostly remained on an alarm, log or simple authority control layer, a computable evidence closed loop is lacked to uniformly quantify a process multi-mode consistency, a platform reliability state and a network controllability into a censorable and reproducible control constraint variable, and further the allowable control mode, a callable waveform set and an energy input boundary and a power change rate boundary are subjected to hierarchical convergence and hysteresis release; Therefore, in the prior art, especially in a remote management scenario, the output risk caused by coupling of network uncertainty, control end credibility and waveform execution consistency can not be verified at the same time is considered, and a technical scheme capable of dynamically converging control authority based on multi-mode evidence and credible evidence and carrying out verifiable constraint on waveform execution consistency is lacking. Disclosure of Invention In order to achieve the above purpose, the present invention provides the following technical solutions: the intelligent laser control and remote management method based on the multi-mode driving comprises the following steps: S1, acquiring operation data of at least two different modes in the operation process of a laser and forming a process feature vector, wherein the operation data of the different modes at least comprise any two types of laser output power data, retro-reflection data, spectrum data, temperature data, current or voltage ripple data, and synchronously acquiring network transmission state data, and the network transmission state data at least comprises a time delay parameter and a jitter parameter; S2, constructing a cross-modal physical consistency residual error based on the process feature vector, wherein the cross-modal physical consistency residual error is used for representing the deviation degree of different modal operation data under a preset physical constraint relation and outputting process credible evidence; s3, performing remote proof or integrity check on the local control end of the laser to output platform trusted evidence, wherein the platform trusted evidence is used for representing the degree that the software and hardware states of the local control end meet a preset trusted base line; S4, calculating a network controllability index based on the network transmission state data and outputting network credible evidence, wherein the network credible evidence is used for representing the influence degree of a remote control link on the stability of closed-loop control; S5, fusing the process credible evidence, the platform credible evidence and the network credible evidence to obtain a credibility score, and mapping the credibility score into a control authority set, wherein the control authority set at l