A method and device for detecting the deflection of a galvanometer motor in a medical galvanometer scanning system

Abstract

The application relates to a detection method and equipment for deflection of a galvanometer motor in a medical galvanometer scanning system. The method part mainly comprises the following steps: establishing a dynamic threshold library, calling a dynamic threshold group suitable for a current scene, the dynamic threshold group comprising an angle error threshold, a vibration threshold and a driver temperature threshold; collecting an inner loop control signal in real time, the inner loop control signal comprising at least a motor angle error signal, a vibration signal and a driver temperature signal; quantitatively fusing the collected signals through a multi-signal weight fusion algorithm and performing preliminary judgment; comparing the signals subjected to the preliminary judgment with the dynamic threshold group, judging whether each signal meets the corresponding threshold condition, and if all meet the threshold condition, determining that the galvanometer motor is deflected to a position. Compared with the prior art, the application has significantly improved precision, avoids normal skin burn risk caused by position deviation, and meets the high-precision safety requirement of a medical scene.

Description

Technical Field

[0001] This application relates to the field of galvanometer scanning technology, and in particular to a method and device for detecting the deflection of the galvanometer motor in a medical galvanometer scanning system. Background Technology

[0002] Galvanometer scanning systems are an important component of medical laser equipment (such as carbon dioxide laser therapy devices). They adjust the laser projection angle by deflecting the galvanometer motor and driving the scanning mirror, enabling precise medical operations such as fractional laser therapy and cutting of minute lesions. The requirements for accuracy, safety, and adaptability of galvanometer scanning systems in medical settings are far higher than in industrial settings. It is essential to ensure that the laser projection position deviation is controlled within a safe range to avoid burning normal skin, while also adapting to the dynamic needs of different treatment modes and treatment sites.

[0003] Existing technologies include a basic solution that determines the deflection of the galvanometer motor by collecting the inner loop control signal of the motor driver. This solution is designed for general scenarios and solves some problems of traditional fixed delay solutions. However, it has significant drawbacks when applied to medical scenarios and cannot meet the specific needs of medical scenarios.

[0004] For example, when the existing basic solutions are applied to medical scenarios, the accuracy of judging the deflection of the galvanometer motor is insufficient, resulting in high medical safety risks: the existing solutions only use a single threshold or a simple combination of signals to judge the position, without considering the problem of laser energy superposition caused by the slight vibration of the galvanometer motor in medical scenarios, and without setting a specific high-precision judgment standard for medical scenarios. It is difficult to control the deviation of the laser projection position within the medical safety range, which can easily cause the risk of burns to normal skin.

[0005] Therefore, how to overcome the problem of insufficient accuracy in judging the deflection of the galvanometer motor in the existing technology and the high risk to medical safety is a problem to be solved in this technical field. Summary of the Invention

[0006] To address the aforementioned deficiencies or improvement needs of existing technologies, and to resolve the issues of insufficient accuracy in determining the deflection of the galvanometer motor and high medical safety risks associated with existing technologies, this application provides a method and device for detecting the deflection of the galvanometer motor in a medical galvanometer scanning system. This method collects multiple types of inner-loop control signals, analyzes them using a multi-signal weighted fusion algorithm, and sets specific judgment criteria for medical scenarios. This significantly improves accuracy compared to existing technologies, avoids the risk of burns to normal skin due to positional deviations, and meets the high-precision safety requirements of medical scenarios.

[0007] The embodiments of this application adopt the following technical solutions:

[0008] In a first aspect, this application provides a method for detecting the deflection of a galvanometer motor in a medical galvanometer scanning system, comprising:

[0009] Establish a dynamic threshold library and call the dynamic threshold group adapted to the current scene. The dynamic threshold group includes the angle error threshold, vibration threshold and driver temperature threshold.

[0010] Real-time acquisition of inner loop control signals, which include at least motor angle error signals, vibration signals, and driver temperature signals;

[0011] The collected signals are quantized and fused using a multi-signal weight fusion algorithm, and a preliminary judgment is made. In the weight allocation of the multi-signal weight fusion algorithm, the proportion of the motor angle error signal is greater than the sum of the proportions of the vibration signal and the driver temperature signal.

[0012] The signals initially judged are compared with the dynamic threshold group to determine whether each signal meets the corresponding threshold condition. If all conditions are met, the galvanometer motor is determined to be deflected into position.

[0013] By adopting the above technical solution, multi-dimensional and accurate determination of the deflection of the galvanometer motor is achieved: the motor angle error signal, which accounts for a higher proportion, is used as the core judgment basis, combined with vibration signal and driver temperature signal for auxiliary verification, avoiding misjudgment caused by a single signal judgment; at the same time, dynamic threshold groups are adapted to different medical scenarios to ensure that the judgment criteria for positioning are in line with clinical practice under different needs such as fine facial treatment and large-area trunk treatment, effectively improving the positional accuracy of medical operations and reducing safety risks such as normal skin burns.

[0014] In some implementations, the dynamic threshold library is equipped with a self-learning mechanism, which includes:

[0015] First run trigger: Automatically selects multiple sets of typical medical scenario instructions, sends them to the motor driver in sequence, synchronously collects the steady-state data of the motor under each set of instructions, and generates an initial threshold library after processing the data;

[0016] Triggered after a preset number of runs: The compensation value is determined by the deviation between the currently collected steady-state data and the historical threshold. The compensated threshold does not exceed the preset safety threshold limit.

[0017] By adopting the above technical solutions, the scenario-adaptive optimization of the dynamic threshold library is achieved: the initial threshold library is quickly built through self-learning on the first run, without the need for manual parameter configuration, reducing the complexity of medical operations; the dynamic compensation mechanism for each preset number of runs can adapt to performance fluctuations caused by motor aging and changes in treatment load in real time, and the constraint design of the upper limit of the safety threshold avoids excessive threshold deviation that may cause medical safety hazards, ensuring the stability of judgment accuracy during long-term use.

[0018] In some implementations, before each compensation is triggered, multiple sets of data are cross-validated to determine whether compensation should be performed.

[0019] If compensation is performed, the current threshold library is automatically backed up before each compensation. If, during treatment verification after compensation, a laser projection deviation is detected to be close to or exceed the safety limit, the system automatically rolls back to the threshold library before compensation.

[0020] By adopting the above technical solutions, the reliability and security of threshold compensation are further improved: multi-data cross-validation avoids miscompensation caused by single data anomalies, ensuring the rationality of the compensation logic; threshold backup and rollback mechanisms form a dual security guarantee, so that even if compensation deviation causes abnormal laser projection accuracy, it can quickly recover to the safe threshold state, avoiding the accumulation of deviations that affect the treatment effect, which meets the high safety redundancy requirements of medical equipment.

[0021] In some implementations, the initial threshold library is synchronized to the cloud platform, and the initial threshold library of the same model device is directly downloaded when a new device starts up, thus canceling the self-learning process triggered on the first run.

[0022] By adopting the above technical solutions, the deployment efficiency of multi-device batch application scenarios is optimized: new devices do not need to repeat the initial self-learning process and can directly reuse the mature threshold library of the same model of device, which greatly shortens the device startup preparation time; at the same time, the cloud synchronization mechanism facilitates the hospital to perform unified operation and maintenance management of multiple devices, reducing the cost of device calibration and maintenance.

[0023] In some implementations, auxiliary determination conditions are also included when determining whether the galvanometer motor has deflected to the correct position:

[0024] Whether the angle error signal remains within the angle error threshold range for a time greater than or equal to a preset time;

[0025] Is the drive current signal within the steady-state current range?

[0026] If any one of the conditions is not met, the galvanometer motor is determined not to have deflected into position.

[0027] By adopting the above technical solutions, the rigor of the positioning judgment is enhanced: the constraint condition of the duration of the angle error can filter out the instantaneous fluctuations during the motor deflection process, such as the overshoot jitter during the step command response, to ensure that the motor is in a stable state when the positioning is determined; the steady-state verification of the drive current further verifies the rationality of the motor's operating load, avoiding the triggering of laser start-stop when the motor is not fully stable or the load is abnormal, and further improving the safety of medical operations and the consistency of treatment effects.

[0028] In some implementations, after determining that the galvanometer motor has deflected into position, the laser's operating parameters are dynamically adjusted based on the real-time vibration signal amplitude. If the vibration signal amplitude is not greater than half of the vibration threshold, the laser outputs power according to the set value. If the vibration signal amplitude is greater than half of the vibration threshold but less than the vibration threshold, the laser automatically reduces the preset power and shortens the preset pulse width.

[0029] By adopting the above technical solution, dynamic collaborative optimization of laser output and motor operation status is achieved: the laser parameters are adjusted according to the amplitude of the vibration signal, and the laser output is carried out at the set power when the motor is running stably to ensure treatment efficiency; when there is slight vibration in the motor, the laser energy is avoided by reducing the power and shortening the pulse width to ensure uniform dot matrix treatment and neat cutting edges; at the same time, it adapts to the laser energy requirements of different medical scenarios, taking into account both treatment effect and safety redundancy.

[0030] In some implementations, it is determined whether the next instruction is a small-angle jump instruction; if it is, the sampling frequency is increased.

[0031] By adopting the above technical solutions, the response efficiency of small-angle jump scenarios can be improved: For small-angle scenarios such as adjacent point jumps in facial fractional treatment, increasing the sampling frequency can increase the signal acquisition density, capture the motor's positioning status faster, and shorten the positioning judgment delay; it can achieve rapid synchronization between laser start / stop and motor positioning, improve the continuity and efficiency of fractional treatment, and reduce the patient's treatment waiting time.

[0032] In a second aspect, this application provides a detection device for the deflection of the galvanometer motor in a medical galvanometer scanning system, which applies the detection method for the deflection of the galvanometer motor in a medical galvanometer scanning system as described in the first aspect, including an intelligent positioning detection device, an application controller, a motor driver, a galvanometer motor, and a laser.

[0033] The application controller establishes control or communication connections with the motor driver, laser, and intelligent positioning detection device, respectively, to generate motor deflection commands according to scenario requirements and send them to the motor driver, while synchronizing current scenario information with the intelligent positioning detection device.

[0034] The motor driver establishes a drive connection with the galvanometer motor to drive the galvanometer motor to perform deflection action and outputs an inner loop control signal reflecting the motor's operating status to the intelligent positioning detection device.

[0035] The intelligent positioning detection device is configured to: establish a dynamic threshold library adapted to different scenarios; call the dynamic threshold group corresponding to the current scenario based on the scenario information synchronized by the application controller, wherein the dynamic threshold group includes an angle error threshold, a vibration threshold, and a driver temperature threshold; collect the inner loop control signal output by the motor driver in real time, wherein the inner loop control signal includes at least a motor angle error signal, a vibration signal, and a driver temperature signal; quantize and fuse the collected inner loop control signal through a multi-signal weight fusion algorithm and complete a preliminary judgment, wherein in the weight allocation of the multi-signal weight fusion algorithm, the proportion of the motor angle error signal is greater than the sum of the proportions of the vibration signal and the driver temperature signal; compare the signal that has passed the preliminary judgment with the called dynamic threshold group item by item to determine whether each signal meets the corresponding threshold condition; if all conditions are met, it is determined that the galvanometer motor has deflected to the correct position, and a positioning feedback signal is sent to the application controller;

[0036] After receiving the bit feedback signal, the application controller sends a switch control signal to the laser to synchronize the laser start / stop with the motor's position status.

[0037] By adopting the above technical solutions, a complete closed-loop hardware architecture of command issuance, signal acquisition, position determination, and laser control was constructed: each component has a clear division of labor and close cooperation. The application controller realizes the overall scheduling of scene commands, the intelligent position detection device completes the core position determination logic, and finally realizes the precise synchronization of laser start / stop and motor position status, providing hardware support for high precision and high safety of medical operations.

[0038] In some embodiments, the intelligent positioning detection device is further configured to preprocess the acquired inner loop control signal using electromagnetic shielding and a Kalman filtering algorithm.

[0039] By adopting the above technical solutions, the acquisition quality of the inner loop control signal is effectively improved: electromagnetic shielding can filter electromagnetic interference generated by equipment such as MRI and ECG in the medical environment, while the Kalman filter algorithm removes signal noise, increasing the signal-to-noise ratio of the filtered signal and avoiding positional distortion caused by interference signals; ensuring that the acquired angle error, vibration and other signals remain intact and accurate in complex medical electromagnetic environments, providing a reliable data foundation for accurate judgment.

[0040] In some implementations, the intelligent positioning detection device has a built-in fault self-testing module. The fault self-testing module is used to monitor the integrity of the inner loop control signal in real time. If a signal disconnection, excessive noise, or driver temperature exceeding a preset value is detected, it immediately sends dual safety signals. One of them is a fault alarm signal, which is transmitted to the application controller. The other is an emergency stop signal, which is transmitted directly to the laser to control the laser to shut down in an emergency.

[0041] By adopting the above technical solutions, a medical-grade safety protection system has been constructed: the fault self-diagnosis module monitors the core signals and equipment status in real time, enabling early detection and handling of abnormal situations; the dual-channel safety signal design forms redundancy protection, so even if the application controller fails, the emergency stop signal can directly trigger the laser to shut down, reducing response delay and minimizing the risk of medical accidents in abnormal scenarios such as signal disconnection and temperature exceeding the standard.

[0042] In summary, this application includes at least the following beneficial technical effects:

[0043] 1. Achieved multi-dimensional and accurate determination of the deflection of the galvanometer motor: The motor angle error signal, which accounts for a higher proportion, is used as the core judgment basis, combined with vibration signal and driver temperature signal for auxiliary verification, avoiding misjudgment caused by a single signal judgment; At the same time, the dynamic threshold group is adapted to different medical scenarios to ensure that the judgment criteria for positioning are in line with clinical practice under different needs such as fine facial treatment and large-area trunk treatment, effectively improving the positional accuracy of medical operations and reducing safety risks such as normal skin burns;

[0044] 2. Achieve scenario-adaptive optimization of dynamic threshold library: The initial threshold library is quickly built through self-learning on the first run, without the need for manual parameter configuration, reducing the complexity of medical operations; the dynamic compensation mechanism after each preset number of runs can adapt to performance fluctuations caused by motor aging and changes in treatment load in real time, and the constraint design of the upper limit of the safety threshold avoids excessive threshold deviation that may cause medical safety hazards, ensuring the stability of judgment accuracy during long-term use.

[0045] 3. Achieve dynamic collaborative optimization between laser output and motor operation status: Adjust laser parameters according to vibration signal amplitude, output at the set power when the motor is running stably to ensure treatment efficiency; when there is slight vibration in the motor, reduce power and shorten pulse width to avoid laser energy superposition, ensuring uniform dot matrix treatment and neat cutting edges; at the same time, adapt to the laser energy requirements of different medical scenarios, taking into account both treatment effect and safety redundancy. Attached Figure Description

[0046] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly described below. Obviously, the drawings described below are merely some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0047] Figure 1 This application provides an architectural diagram of a detection device for the deflection of a galvanometer motor in a medical galvanometer scanning system.

[0048] Figure 2This is a block diagram of the intelligent positioning detection device provided in the embodiments of this application;

[0049] Figure 3 This is a flowchart illustrating a method for detecting the deflection of a galvanometer motor in a medical galvanometer scanning system, as provided in an embodiment of this application. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. Furthermore, the technical features involved in the various embodiments described below can be combined with each other as long as they do not conflict with each other.

[0051] It should be noted that, when the existing basic solutions are applied to medical scenarios, in addition to the insufficient accuracy in judging the deflection of the galvanometer motor and the high risk to medical safety, the following defects also exist:

[0052] Poor scene adaptability and inability to meet dynamic medical needs: The basic solution of the existing technology lacks a scene-based parameter adaptation mechanism. It cannot automatically adjust the judgment criteria according to different treatment areas such as the face (high precision requirements) and the torso (large area treatment requirements). It also cannot adapt to the differentiated needs of different medical operations such as laser fractional treatment (multiple fractional shapes) and laser cutting (continuous / intermittent modes). Medical staff need to manually adjust the parameters, which increases the complexity of operation and the risk of error.

[0053] Lack of medical-specific safety and anti-interference design: The basic solution of the existing technology does not have a dedicated anti-interference mechanism for medical devices, making it difficult to resist electromagnetic interference generated by equipment such as MRI and ECG in the medical environment, which may lead to signal distortion and affect the accuracy of the diagnosis; at the same time, there is no medical-grade fault self-check and safety redundancy design, which cannot meet the electromagnetic compatibility and safe operation requirements of medical devices, and poses risks to medical operation.

[0054] Insufficient synergy in laser control leads to poor treatment results: The existing basic solution only achieves a simple linkage of "motor in position → laser start and stop", without dynamically adjusting laser parameters according to the motor's operating status. When there is slight vibration in the motor, problems such as laser energy superposition or rough cutting edges may occur, affecting the medical treatment effect.

[0055] In addition to addressing the issues of insufficient accuracy in judging the deflection of the galvanometer motor and high medical safety risks, this application also proposes corresponding solutions to the aforementioned problems.

[0056] The present application will now be described in detail with reference to the accompanying drawings and embodiments. Example 1

[0057] like Figure 1 As shown, Embodiment 1 of this application provides a detection device for the deflection of the galvanometer motor in a medical galvanometer scanning system, which mainly includes an intelligent positioning detection device, an application controller, a medical-grade motor driver, a medical high-precision galvanometer motor, a medical laser, and a medical equipment cloud platform.

[0058] The application controller has a bidirectional connection to a medical-grade motor driver, used to send motor deflection commands (including dot matrix / cutting mode and treatment site adaptation parameters) to the driver and to receive status feedback from it. It also has a unidirectional connection to a medical laser, outputting laser switching signals and dynamic parameters (power, pulse width) to the laser. Furthermore, it has a bidirectional connection to an intelligent positioning detection device, sending pre-judgment information and receiving positioning feedback and fault alarm signals. Finally, it has a unidirectional connection to a medical device cloud platform, used to upload device operating data and download optimized parameters for the same model of equipment.

[0059] The medical-grade motor driver is bidirectionally connected to a high-precision medical galvanometer motor, used to output drive signals to control motor deflection and to receive angle and speed feedback signals from the motor. The medical-grade motor driver is unidirectionally connected to an intelligent positioning detection device, used to output inner-loop control signals (angle error, drive current, vibration amplitude, driver temperature, etc.) to the intelligent positioning detection device in real time; a closed-loop structure for position, current, and vibration suppression is adopted to ensure the real-time performance and stability of signal acquisition.

[0060] The intelligent positioning detection device connects unidirectionally to the medical laser to send a safety shutdown signal in abnormal scenarios (signal disconnection, excessive noise) to urgently shut down the laser. It also connects bidirectionally to the medical equipment cloud platform to upload dynamic threshold libraries, fault data, and download batch application optimization parameters.

[0061] The high-precision medical galvanometer motor has no direct electrical connection with the medical laser. It achieves precise synchronization between mechanical deflection and laser projection through a logical linkage of motor deflection into position → feedback from the detection device → controller command → laser start and stop.

[0062] The core working principle of the detection equipment for galvanometer motor deflection in a medical galvanometer scanning system is explained below:

[0063] Command generation and issuance stage: The application controller generates appropriate motor deflection commands based on medical operation requirements (such as facial spiral dot matrix treatment, cutting of minor skin lesions) (dot matrix mode corresponds to step signal sequence, cutting mode corresponds to continuous / intermittent signal), and simultaneously sends the deflection direction and angle increment of the next command to the intelligent positioning detection device to complete the prediction preparation.

[0064] Signal acquisition and real-time analysis stage: The medical-grade motor driver drives the galvanometer motor to deflect, and simultaneously outputs inner-loop control signals (angle error, drive current, vibration amplitude, driver temperature) to the intelligent positioning detection device in real time; the detection device filters out electromagnetic interference from the medical environment (such as MRI and ECG interference) through a medical-grade isolation sampling circuit and Kalman filter algorithm to ensure signal integrity; a multi-signal weighted fusion algorithm is activated (e.g., 60% weight for angle error, 20% weight for vibration signal, and 20% weight for temperature signal), combined with a dynamic threshold library (e.g., when adapting to a high-precision medical galvanometer motor with an accuracy level ≥ ±0.002°, the facial treatment angle error is ±0.003°, and the trunk treatment angle error is ±0.008°) to analyze the signal.

[0065] Position determination and feedback phase: When the three conditions are met—angle error ≤ corresponding dynamic threshold and lasting for 30μs, vibration amplitude ≤ 0.01g, and temperature ≤ 60℃—the detection device determines that the motor deflection is in place and generates a medical-grade isolated feedback signal (5V amplitude, 30μs pulse width), which is sent to the application controller through the shielded SPI interface. If an abnormal signal is detected (disconnection, excessive noise), a safety shutdown signal is immediately sent to the application controller and the medical laser, triggering the laser emergency shutdown and alarm.

[0066] Laser precision control stage: After receiving the position feedback signal, the application controller delays for 30-50μs to match the response time of the medical laser, and dynamically adjusts the laser parameters according to the motor vibration state (for example, for trunk treatment, if the vibration is ≤0.005g, the set power is output; if the vibration exceeds the standard, the power is reduced by 10% and the pulse width is shortened by 30%), and then a switch control signal is sent. In laser fractional mode, after a single fractional treatment (20μs) is completed, the application controller automatically sends the next step command and repeats the above process. When switching the cutting mode trajectory, the detection device determines the motor jump state (current >90mA, speed >0.8° / ms), keeps the laser off, and turns it on again after the motor jumps to the correct position.

[0067] Dynamic optimization and data synchronization phase: Every 100 complete treatment commands (including single continuous operations of dot matrix / cutting) are executed by the device, the intelligent positioning detection device automatically starts the self-learning module to compensate for motor aging errors (e.g., threshold fine-tuning ±0.002°) and update the dynamic threshold library; the dynamic threshold library and fault data are uploaded to the medical device cloud platform on a regular basis. When a new device is started, the optimized parameters of the same model can be downloaded directly without repeated self-learning.

[0068] Furthermore, regarding the working principle of the intelligent positioning detection device, it is the core sensing and decision-making unit of the medical galvanometer scanning system. Its core logic is self-adaptation to medical scenarios + multi-dimensional signal fusion + end-to-end safety redundancy. Through a closed-loop process of signal acquisition - preprocessing - intelligent judgment - feedback control - dynamic optimization, it accurately determines the positioning status of the galvanometer motor deflection, while simultaneously meeting the high precision, high safety, and strong anti-interference requirements of medical scenarios. Its detailed working principle is as follows:

[0069] I. Core Hardware Components: Reference Figure 2 As shown, the intelligent positioning detection device integrates a medical-grade core processing unit, a multi-channel signal acquisition unit, a medical-grade anti-interference unit, a medical-grade communication unit, and a safety unit. Specifically, the core processing unit includes a medical-grade FPGA (e.g., model XC7K325T), responsible for signal analysis, algorithm calculation, and instruction generation, with a sampling frequency supporting dynamic switching from 20kHz to 50kHz; the signal acquisition unit includes a multi-channel isolated sampling circuit (medical-grade isolation voltage ≥2.5kV), adapting to the acquisition of four types of signals: angle error, drive current, vibration, and temperature; the anti-interference unit includes an electromagnetic shielding module (compliant with medical EMC standard EN 60601-1-2) and a Kalman filter algorithm hardware acceleration module; the communication unit includes a medical-grade shielded SPI interface (transmission rate 10Mbps) and a cloud communication module (supporting medical device data security transmission protocols); the safety unit includes a fault self-test circuit and an emergency stop signal generation module.

[0070] II. Phased Working Principle:

[0071] 1. Initialization and self-learning phase (laying the foundation for adaptation to medical scenarios):

[0072] After the intelligent positioning detection device is activated, it first executes a "medical scenario self-learning process" to establish an adaptation benchmark for subsequent positioning judgments: It receives medical scenario parameters from the application controller (e.g., treatment mode: fractional / cutting; treatment area: face / torso; device model: SG-10M galvanometer motor, laser parameters); automatically traverses three sets of typical motor deflection commands in the current scenario (e.g., small-angle step command for facial fractional cutting, medium-angle continuous command for torso continuous cutting, large-angle jump command for treatment trajectory switching); and collects steady-state motor operation data (e.g., angle error fluctuation range, steady-state value of drive current, vibration amplitude, driver temperature); based on the collected data... The system generates a "dynamic threshold library" which is stored hierarchically according to medical scenarios. For example, for facial fractional radiography: dynamic threshold for angle error ±0.003°, steady-state current 30-40mA, vibration threshold ≤0.005g; for trunk fractional radiography: dynamic threshold for angle error ±0.008°, steady-state current 35-50mA, vibration threshold ≤0.01g; for continuous cutting mode: steady-state current 30-50mA, speed threshold ≤0.05° / ms, temperature threshold ≤60℃. After every 100 runs, the system automatically repeats the above self-learning process to compensate for errors caused by motor aging and load changes (e.g., threshold fine-tuning ±0.002°) and updates the dynamic threshold library.

[0073] 2. Signal Acquisition and Preprocessing Stage (Ensuring Signal Reliability in the Medical Environment):

[0074] During motor deflection, the intelligent positioning detection device collects and purifies signals in real time to avoid interference from the medical environment affecting the accuracy of judgment: Through a multi-channel isolated sampling circuit, it synchronously collects four types of inner-loop control signals output by the medical-grade motor driver. Among them, the angle error signal is extracted from the motor driver's position loop, reflecting the deviation between the current angle of the motor and the preset angle; the drive current signal is extracted from the current loop, reflecting the motor's operating load status; the vibration signal is extracted from the motor driver's existing vibration suppression system (or collected through the built-in acceleration sensor), reflecting the motor's operating stability; and the temperature signal is collected from the driver's temperature sensor, reflecting the long-term operating safety of the equipment. The collected raw signals are subjected to "medical-grade preprocessing": electromagnetic shielding filters electromagnetic interference generated by medical equipment such as MRI and ECG machines; Kalman filtering algorithm removes signal noise (the signal-to-noise ratio after filtering is ≥60dB); and the isolated sampling circuit avoids high-voltage crosstalk, ensuring that the signal meets medical safety standards.

[0075] 3. Multi-dimensional intelligent judgment stage (core decision-making process, logic specific to medical scenarios):

[0076] The intelligent positioning detection device determines whether the motor is in position based on pre-processed signals using a weighted fusion algorithm and multiple condition constraints. Specifically, it calls a dynamic threshold library to obtain the judgment threshold corresponding to the current scene; it then activates a multi-signal weighted fusion algorithm to quantify and analyze four types of signals. For the angle error signal (weight 60%), it checks whether its absolute value is less than or equal to the dynamic threshold and its duration is greater than or equal to 30μs (to avoid misjudgment due to minor motor vibrations); for the vibration signal (weight 20%), it checks whether its amplitude is less than or equal to the vibration threshold corresponding to the scene (e.g., ≤0.005g for facial treatment); for the temperature signal (weight 20%), it checks whether the driver temperature is less than or equal to 60℃ (to avoid motor performance drift due to high temperatures); and for the drive current signal (auxiliary judgment), it checks whether it is within the steady-state current range as a supplementary verification condition. The judgment logic is as follows: "Deflection in position" is determined only when all three conditions are met simultaneously (angle error meets the condition, vibration amplitude meets the condition, and temperature meets the condition); if any condition is not met, it is determined as "not in position," and signal acquisition continues for repeated judgment.

[0077] 4. Feedback signal generation and transmission stage (medical-grade secure communication):

[0078] The intelligent positioning detection device generates corresponding feedback signals based on the positioning judgment result, ensuring safe coordination with the application controller and laser. Specifically, if "positioned" is determined: a medical-grade isolated positioning feedback signal (amplitude 5V, pulse width 30μs, rise time ≤1μs) is generated and transmitted to the application controller via a medical shielded SPI interface, notifying it to control the laser's start and stop; if "not positioned" is determined: no positioning feedback signal is sent, and signal acquisition is continuously repeated for judgment until the positioning condition is met; if an "abnormal signal" is detected (such as signal disconnection, excessive noise, temperature exceeding 65℃): dual safety signals are immediately generated, one sending a fault alarm signal to the application controller via the SPI interface, and the other directly sending an emergency stop signal to the medical laser, controlling the laser to shut down urgently and triggering the equipment alarm (compliant with medical safety standard EN 60601-1).

[0079] 5. Path prediction and dynamic optimization stage (improving the synchronization of medical operations):

[0080] To address the needs of "continuous cutting trajectory switching" and "rapid dot matrix jumps" in medical scenarios, a predictive optimization logic has been added: The intelligent positioning detection device receives "next instruction prediction information" (e.g., deflection direction, angle increment, instruction type) from the application controller; based on the prediction information, the working parameters are dynamically adjusted. If the prediction is a "small angle jump instruction" (e.g., jump between adjacent points in the dot matrix): the sampling frequency is increased from 20kHz to 50kHz to increase signal acquisition density and shorten judgment delay; if the prediction is a "medium angle continuous instruction" (e.g., cutting a long trajectory): the sampling frequency is maintained at 20kHz, and the filtering algorithm parameters are optimized to balance response speed and signal stability; when switching cutting trajectories, the current and speed signals collected are used to determine whether the motor is in a "jump state" (current > 90mA, speed > 0.8° / ms). At this time, the positioning judgment is paused, the laser is kept off, and the judgment process is resumed only after the motor jumps to the new trajectory and meets the positioning conditions.

[0081] 6. Data synchronization and batch adaptation stage (supporting multi-device medical applications):

[0082] The intelligent positioning detection device regularly uploads key data to the medical equipment cloud platform, supporting batch application across multiple devices. Specifically, the uploaded data includes: dynamic threshold library, fault records (signal anomaly time and type), and self-learning logs. When a new device is started, an optimized dynamic threshold library for the same model can be downloaded from the cloud, eliminating the need to re-execute the complete self-learning process (only verification command calibration is required), thus adapting to the unified operation and maintenance needs of multiple devices in a hospital.

[0083] The intelligent positioning detection device, based on the above solution, offers at least the following advantages:

[0084] Multi-signal fusion judgment: Unlike the single-signal judgment of existing technologies, it avoids misjudgments caused by minor motor vibrations and temperature drift in medical scenarios through multi-signal weight fusion; Dynamic threshold self-adaptation: Based on self-learning in medical scenarios, rather than fixed thresholds, it adapts to differences in different treatment sites and equipment states; Medical-grade anti-interference: The dual design of hardware shielding and algorithm filtering meets the electromagnetic compatibility requirements of the medical environment and avoids laser projection deviation caused by interference; Full-process safety redundancy: Fault self-checking and emergency shutdown dual protection meet medical equipment safety standards and reduce medical risks. Finally, the intelligent positioning detection device provided in this application can achieve an angle error recognition accuracy of ≤±0.002°, a signal sampling delay of ≤30μs, an electromagnetic interference suppression ratio of ≥60dB, and a laser emergency shutdown delay of ≤30μs in case of failure. It is also compatible with circular / square / spiral dot arrays and continuous / intermittent cutting modes without the need for manual parameter adjustment. Example 2

[0085] Based on the detection device for galvanometer motor deflection in the medical galvanometer scanning system provided in Embodiment 1, this Embodiment 2 provides a method for detecting galvanometer motor deflection in the medical galvanometer scanning system.

[0086] refer to Figure 3 As shown, the method includes the following steps:

[0087] Step 101: Establish a dynamic threshold library and call the dynamic threshold group adapted to the current scene. The dynamic threshold group includes the angle error threshold, vibration threshold and driver temperature threshold.

[0088] Step 102: Real-time acquisition of inner loop control signals, which include at least motor angle error signals, vibration signals, and driver temperature signals.

[0089] Step 103: Quantize and fuse the collected signals using a multi-signal weight fusion algorithm and make a preliminary judgment; wherein, in the weight allocation of the multi-signal weight fusion algorithm, the proportion of the motor angle error signal is greater than the sum of the proportions of the vibration signal and the driver temperature signal.

[0090] Step 104: Compare the signals that have been preliminarily judged with the dynamic threshold group to determine whether each signal meets the corresponding threshold condition. If all of them meet the condition, the galvanometer motor is determined to be deflected into position.

[0091] The main purpose of steps 101-104 above is to address the technical problem that existing technologies lack sufficient accuracy in determining the laser projection position. These technologies rely on a single threshold or simple signal combinations, failing to consider minor motor vibrations and the resulting potential for excessive laser projection position deviation, which could lead to burns to normal skin. Steps 101-104 improve the laser projection / cutting position deviation to ≤0.01mm, significantly improving accuracy compared to existing technologies. This avoids the risk of burns to normal skin due to positional deviations and meets the high-precision safety requirements of medical settings.

[0092] Specifically, the core of the multi-signal weighted fusion algorithm is to prioritize and quantify multiple core signals, ultimately outputting a comprehensive fit index, rather than judging whether a single signal meets the standard. Among these, the motor angle error signal (weight 60%) is used to analyze whether the positional accuracy meets the standard. This signal directly reflects the deviation between the current deflection angle of the motor and the preset treatment angle (e.g., ±0.003° for facial dot matrix and ±0.008° for the torso), and is the core basis for judging "whether it is in place," with the highest weight. The vibration signal (weight 20%) is used to analyze whether the operational stability meets the standard. This signal captures the minute vibrations after motor deflection (e.g., the high-frequency vibration of a lightweight medical motor), avoiding laser energy superposition due to "instantaneous angle meeting the standard but vibration not disappearing," and is a key supplement to judging "whether it is stably in place." The driver temperature signal (weight 20%) is used to analyze whether the equipment status is safe. This signal monitors the temperature of the driver after long-term operation, avoiding motor performance drift caused by high temperatures (e.g., angle error caused by changes in coil resistance), and is an important condition for judging "whether it is safely in place."

[0093] Analysis logic of multi-signal weighted fusion algorithm (quantization fusion process):

[0094] Step 1: Signal Standardization Processing. The three types of raw signals are converted into standardized scores of 0-100 (the higher the score, the closer to the "stable and safe" state). Specifically: Angle Error Signal: 100 points if error = 0; 0 points if error reaches twice the dynamic threshold upper limit; intermediate values ​​are scored linearly (e.g., a facial scene error of 0.003° at the dynamic threshold upper limit earns 50 points). Vibration Signal: 100 points if vibration amplitude = 0; 0 points if it reaches twice the scene adaptation threshold upper limit; intermediate values ​​are scored linearly (e.g., a facial scene vibration of 0.005g at the dynamic threshold upper limit earns 50 points). Temperature Signal: 100 points if temperature is less than or equal to 30℃ (ideal steady state); 0 points if temperature = 60℃ (threshold upper limit); intermediate values ​​are scored linearly (e.g., 50 points for a temperature of 45℃).

[0095] Step 2: Calculate the overall fit using a weighted summation. The total score is calculated based on the weighted distribution: Overall Fit = Angular Error Standardized Score × 60% + Vibration Standardized Score × 20% + Temperature Standardized Score × 20%.

[0096] Step 3: Output the analysis results. Only when the overall fit score is ≥60 points (which can be adjusted according to the accuracy requirements of the medical scenario) and the "angle error continuously meets the standard for ≥30μs" condition is the subsequent linkage judgment with the dynamic threshold library proceeded; if the total score is <60 points, it is directly judged as "not in place" without further judgment.

[0097] The multi-signal weighted fusion algorithm acts as a "pre-screening device," while the dynamic threshold library specific to medical scenarios serves as the "final verification standard." The two are linked to form a dual-judgment process of "quantitative fusion followed by precise verification," with the specific steps as follows:

[0098] 1. First Step: Dynamic Threshold Library Pre-Adapted to Scenarios. The intelligent positioning detection device has generated a dynamic threshold library adapted to different treatment needs through a self-learning process for medical scenarios (e.g., facial dot matrix: angle error threshold ±0.003°, vibration threshold ≤0.005g; torso cutting: angle error threshold ±0.008°, vibration threshold ≤0.01g). When the application controller sends a motor deflection command, it simultaneously informs the intelligent positioning detection device of the current scenario (e.g., "spiral dot matrix + facial treatment"). The detection device automatically calls the corresponding threshold group for the scenario, preparing for subsequent verification.

[0099] 2. Second step: Pre-screening using a multi-signal weighted fusion algorithm. The intelligent positioning detection device collects three types of signals in real time and calculates the comprehensive fit using the algorithm described above. If the comprehensive fit score is <60 points: it is directly judged as "not in position" and the signal is re-collected (e.g., when the motor has just started, the total score is <30 points, so it is judged as not in position). If the comprehensive fit score is ≥60 points: it means that the motor is close to the "stable and safe in position" state, triggering the next threshold verification step (e.g., in a facial scene, the standardized scores for angle error, vibration, and temperature are 80 points, 60 points, and 83.3 points respectively, so the comprehensive fit score = 80×60% + 60×20% + 83.3×20% ≈ 76.7 points, which meets the pre-screening conditions). In addition, when performing pre-screening, the multi-signal weighted fusion algorithm can also introduce parameters such as steady-state current and rotation speed threshold for auxiliary screening based on different scenarios. For example, in the facial fractional treatment scenario, a steady-state current of 30-40mA is introduced. If the steady-state current condition is not met, the screening will not pass. For example, in the continuous cutting mode, a steady-state current of 30-50mA and a rotation speed threshold of ≤0.05° / ms are introduced. If the steady-state current and rotation speed threshold conditions are not met, the screening will not pass.

[0100] 3. Third Step: Final Precision Verification of the Dynamic Threshold Library. Signals that passed the initial screening are verified item by item according to the dynamic threshold group for the current scenario. Verification 1: Whether the motor angle error is ≤ the corresponding dynamic threshold (e.g., ≤ ±0.003° for facial scenarios), and the duration is ≥ 30μs (avoiding instantaneous compliance). Verification 2: Whether the vibration amplitude is ≤ the scene-appropriate vibration threshold (e.g., ≤ 0.005g for facial scenarios). Verification 3: Whether the driver temperature is ≤ 60℃ (a universal safety threshold, suitable for all medical scenarios). Only when all three verifications pass is the final determination that "the galvanometer motor deflection is in place"; if any verification fails (e.g., the angle error meets the standard but the vibration exceeds the standard), the process returns to the initial screening stage, and the overall compatibility is recalculated.

[0101] 4. Fourth step: Output positioning feedback signal. After determining that the laser is in position, the intelligent positioning detection device generates a medical-grade isolated positioning feedback signal and sends it to the application controller to trigger the laser to start and stop precisely (e.g., in a facial fractional laser scenario, the laser outputs 15W power and continues for 20μs to complete a single fractional laser treatment).

[0102] The following example (facial spiral fractional treatment) will further illustrate the above solution. Scene Adaptation: The intelligent positioning detection device uses the "facial fractional" dynamic threshold group (angle error ±0.003°, vibration ≤0.005g). Signal Acquisition: During motor deflection, the intelligent positioning detection device acquires the following signals: angle error 0.003°, vibration 0.006g, temperature 32℃. Algorithm Analysis: Standardized scores are 50 (angle error), 40 (vibration), and 93.3 (temperature). The overall adaptation score = 50×60% + 40×20% + 93.3×20% ≈ 56.7 points < 60 points, indicating inadequate positioning; acquisition continues. Signal Update: As the motor gradually stabilizes, the following signals are acquired: angle error 0.003°, vibration 0.004g, temperature 32℃. Algorithm Analysis: Standardized scores were 50, 60, and 93.3 respectively. The overall fit was calculated as 50×60% + 60×20% + 93.3×20% ≈ 60.7 points, which is greater than or equal to 60 points. Furthermore, the angle error remained at this level for 12μs, triggering threshold verification. Threshold Verification: Angle error 0.003° ≤ ±0.003° (lasting 12μs), vibration 0.004g ≤ 0.005g, and temperature 32℃ ≤ 60℃. All three criteria were met, indicating successful placement. Feedback Control: The intelligent placement detection device sent a placement feedback signal. The application controller delayed for 15μs to activate the laser, completing the single-point fractional treatment.

[0103] In the above process, the role of the multi-signal weighted fusion algorithm is to filter out signals that are "close to being in place" through "quantitative weighting," avoiding medical risks caused by misjudgment due to a single signal and improving the reliability of the judgment. The combination logic of the multi-signal weighted fusion algorithm and the dynamic threshold library is as follows: the algorithm performs "coarse screening" (excluding obviously incomplete states), and the threshold library performs "fine calibration" (accurate judgment according to the scenario). The combination of the two ensures the high accuracy requirements of medical scenarios while taking into account the efficiency and safety of the judgment, completely solving the defects of single threshold judgment and lack of scenario adaptation in existing technologies. Example 3

[0104] Based on the solutions in Embodiments 1 and 2, Embodiment 3 further establishes a self-learning mechanism for the dynamic threshold library. This self-learning mechanism includes: upon initial operation, automatically selecting multiple sets of typical medical scenario instructions and sequentially issuing them to the motor driver; synchronously collecting steady-state motor data under each set of instructions; and generating an initial threshold library after data processing. Upon triggering a preset number of runs, the compensation value is determined by the deviation between the currently collected steady-state data and historical thresholds, ensuring that the compensated threshold does not exceed a preset safety threshold upper limit. Furthermore, before each compensation trigger, multiple sets of data are cross-validated to determine whether compensation should be performed. If compensation is performed, the current threshold library is automatically backed up before each compensation. If, during treatment verification after compensation, a laser projection deviation is detected to be close to or exceed the safety upper limit, the system automatically rolls back to the threshold library before compensation. Furthermore, the initial threshold library is synchronized to the cloud platform. When a new device starts, the initial threshold library of the same model is directly downloaded, canceling the self-learning process triggered upon initial operation.

[0105] The main purpose of the above solution is to address the technical problem of poor adaptability of existing technologies to different scenarios, which fails to meet dynamic medical needs. Through this solution, medical staff no longer need to manually adjust parameters; the system automatically adapts to different treatment sites, modes, and equipment statuses, significantly reducing operational complexity, minimizing the risk of human error, and improving the usability of medical equipment.

[0106] Specifically, the self-learning process for medical scenarios represents an improvement from "passively receiving parameters" to "actively adapting to the scenario." The dynamic threshold library serves as a "precise adaptation dictionary" for medical scenarios, and is the core output of the self-learning process. Essentially, it's a set of "approval criteria" categorized and stored according to medical scenarios. For example, for facial dot matrix scenarios: angle error threshold ±0.003°, vibration amplitude threshold ≤0.005g, steady-state current threshold 30-40mA, and settling time 30μs. For continuous torso cutting scenarios: angle error threshold ±0.008°, rotation speed threshold ≤0.05° / ms, steady-state current threshold 35-50mA, and temperature threshold ≤60℃. For spiral dot matrix scenarios: based on the facial / torso thresholds, the duration threshold of the angle error is further optimized (15μs) to adapt to the continuous deflection characteristics of the spiral trajectory. Update Mechanism: The self-learning process, which automatically compensates for dynamic changes, performs an incremental update to the threshold library every 100 runs. If motor aging causes an increase in angle error of 0.002°, the corresponding angle error threshold is automatically fine-tuned by +0.002° (e.g., facial dot matrix from ±0.003° to ±0.005°), ensuring the judgment standard always adapts to the current device state. If, during multiple treatments, the vibration amplitude is found to be generally low in a certain scenario (e.g., pediatric facial treatment), the threshold library automatically optimizes the vibration threshold for that scenario (e.g., lowering it from ≤0.005g to ≤0.004g), improving accuracy.

[0107] The core of the initial self-learning run is "automatic scene adaptation calibration when the device is used preclinically for the first time after leaving the factory," without manual intervention from medical staff. The specific steps are as follows:

[0108] 1. Triggering Conditions (Automatic triggering, no manual operation required). After unpacking and installing the equipment, connect the power and start the medical laser therapy instrument (such as a carbon dioxide laser therapy instrument); after the system completes the power-on self-test (checking whether the core components such as the motor, laser, and intelligent positioning detection device are normal), if it detects that "the dynamic threshold library is empty" (first run indicator), the self-learning process will be automatically triggered, and the operation interface will pop up a "Self-learning in progress, please wait" prompt (this does not affect other functions, it only runs in the background).

[0109] 2. Self-learning execution process (runs automatically in the background, approximately 30 seconds in total). The system automatically selects 3 sets of typical medical scenario instructions (no doctor setting required) and sends them to the motor driver in sequence: First set, small angle step instruction (angle increment 0.05°, corresponding to the smallest deflection unit of facial fractional treatment); Second set: medium angle continuous instruction (angle increment 0.5°, continuous output for 5 seconds, corresponding to continuous torso cutting); Third set: large angle jump instruction (angle increment 2°, corresponding to the maximum deflection amplitude of treatment trajectory switching). The intelligent positioning detection device synchronously collects steady-state data of the motor under each set of commands: Under small-angle step commands, it collects the angle error fluctuation range (e.g., 0.001°-0.003°), vibration amplitude (e.g., 0.002g-0.004g), and steady-state current (e.g., 32mA-38mA) after the motor reaches steady state; Under medium-angle continuous commands, it collects the speed stability of the motor during continuous operation (e.g., 0.03° / ms-0.04° / ms), driver temperature change (e.g., 30℃-35℃), and current fluctuation (e.g., 35mA-45mA); Under large-angle jump commands, it collects the transition time from start-up to steady state of the motor (e.g., 150μs), angle error after jump (e.g., 0.002°-0.005°), and vibration decay time (e.g., 20μs). Data Processing and Threshold Library Generation: The FPGA module of the intelligent positioning detection device filters and removes outliers from the collected raw data (e.g., removing instantaneous current peaks), and generates a dynamic threshold library categorized by medical scenario: For facial dot matrix scenarios, the angle error threshold is the maximum value collected + 0.002° (e.g., 0.003° + 0.002° = ±0.005°, reserving safety redundancy), and the vibration threshold is the maximum value (e.g., 0.004g); For torso cutting scenarios, the rotation speed threshold is the maximum value collected (e.g., 0.04° / ms), and the steady-state current threshold is the upper and lower limits of the collected range (e.g., 35mA-45mA); For trajectory switching scenarios, the angle error threshold is ±0.005°, and the stabilization time is 30μs (to avoid minor jitter after switching). Self-learning Completion: The operation interface displays "Self-learning successful," and the dynamic threshold library is stored locally on the device. Subsequent treatments can directly call upon this library without repeated execution.

[0110] Special case handling: If a component abnormality is detected during the self-learning process (such as excessive motor angle error or signal acquisition failure), the system will automatically retry twice; if it still fails, a "Self-learning failed, please contact maintenance" prompt will pop up, and the factory-preset general medical threshold library will be enabled (to ensure that basic treatment can be carried out normally, and the threshold library will be updated after maintenance personnel come to calibrate it on-site).

[0111] Furthermore, the self-learning compensation after every 100 runs is not a fixed compensation, but a dynamic adaptive compensation based on the actual collected data. The core principle of compensation is small-range fine-tuning + upper limit constraint to avoid excessive threshold deviation leading to safety hazards. Example 1: If slight aging of the motor causes the angle error to increase by 0.0008° compared to the historical threshold, then compensate by +0.0008°; Example 2: If the recent treatment load is relatively light (e.g., mostly facial treatments for children with low skin hardness), and the angle error decreases by 0.0005° compared to the historical threshold, then compensate by -0.0005°; Example 3: If there is no significant change in the data (deviation <0.0003°), then no compensation is performed, and the original threshold remains unchanged. Compensation upper limit constraint: To avoid excessive threshold deviation, a "medical safety threshold upper limit" is set, and the compensated threshold for all scenarios must not exceed this upper limit. For example, in facial dot matrix scenarios: the upper limit of the angle error threshold is ±0.008° (even with multiple compensations, this value will not be exceeded, ensuring that the laser projection deviation is always ≤0.01mm, with no risk of burns); in torso cutting scenarios: the upper limit of the steady-state current threshold is 60mA (the upper limit of the safe current for medical galvanometer motors, to avoid overload); if the threshold approaches the upper limit after a certain compensation (e.g., the facial dot matrix reaches ±0.0075°), then the next self-learning will only allow compensation of ≤0.0005°, until it does not exceed the upper limit.

[0112] Furthermore, before each compensation trigger, three sets of data are cross-validated to avoid miscompensation caused by a single data anomaly. For example, cross-validation 1: compare the deviation trends of the three types of data collected this time—angle error, current, and vibration—to see if they are consistent (e.g., if the angle error increases while the current also rises slightly, it indicates that the motor is indeed aging and can be compensated; if only a single data is abnormal, it is judged as an occasional fluctuation and no compensation is given); cross-validation 2: compare the average steady-state data of the last 10 treatments. If the deviation of the current data from the average is <5%, it is judged as normal fluctuation and no compensation is given; cross-validation 3: check whether the driver temperature is within the normal range (≤60℃). If the temperature exceeds the limit and causes data anomalies, a cooling prompt is triggered first, and no compensation is given.

[0113] Furthermore, a "threshold rollback mechanism" is set up. Before each compensation, the current threshold library is automatically backed up; if, during treatment verification after compensation, the laser projection deviation is detected to be close to the safety limit (such as a facial dot matrix deviation of 0.009mm), or if medical staff trigger "accuracy abnormality" feedback, the system will automatically roll back to the threshold before compensation and pop up a "threshold abnormality, calibration recommended" prompt.

[0114] With the above solution, the first self-learning run is "automatically completing scenario adaptation after the device leaves the factory", which is executed in the background without manual operation. The generated threshold library directly serves subsequent treatment. The compensation every 100 runs is a dual design of "dynamic adaptation + safety constraints". It not only solves the accuracy drift caused by equipment aging and load changes, but also avoids excessive threshold deviation through upper limit constraints, cross-validation and rollback mechanisms. There are no safety hazards and it is fully adapted to the high safety requirements of medical scenarios. Example 4

[0115] Based on the aforementioned solution, in order to further address the lack of medical-specific safety and anti-interference design in existing technologies, this application also adopts a medical-grade anti-interference design for the intelligent positioning detection device. This includes electromagnetic shielding and Kalman filtering algorithms that comply with the medical EMC standard EN60601-1-2 to preprocess the acquired inner-loop control signals and remove electromagnetic interference generated by equipment such as MRI and ECG in the medical environment. It also includes a built-in fault self-test module to monitor the integrity of the inner-loop control signals in real time. If a signal disconnection, excessive noise, or driver temperature exceeding 65°C is detected, dual safety signals are immediately sent: one is a fault alarm signal transmitted to the application controller, and the other is an emergency stop signal directly transmitted to the medical laser to control the laser to shut down in an emergency, with a laser emergency shutdown delay of ≤30μs.

[0116] Existing technologies lack dedicated anti-interference mechanisms, and complex electromagnetic interference in medical environments can easily lead to signal distortion. The above-mentioned solution in this application provides dual protection through "hardware shielding + algorithm filtering": the electromagnetic shielding uses medical-grade shielding materials, which can suppress more than 80% of external electromagnetic radiation; the Kalman filter algorithm is optimized for the signal noise characteristics of medical scenarios, which can effectively filter high-frequency pulse interference (such as the pulse signal generated when the ECG machine is working), and the signal-to-noise ratio of the pre-processed signal is ≥60dB, ensuring the accuracy of the judgment.

[0117] The self-diagnostic module's monitoring scope covers three core risk points: signal transmission, signal quality, and equipment temperature. These include signal disconnection (such as the connection between the motor driver and the detection device coming loose), excessive noise (such as electromagnetic interference causing signal amplitude fluctuations to exceed 30% of the normal range), and temperature exceeding 65℃ (overheating caused by the driver working continuously for a long time). It comprehensively covers common fault scenarios of medical equipment.

[0118] Dual protection of safety signals: The emergency stop signal adopts a dual-path design of "direct transmission + controller linkage". Even if the application controller fails, the intelligent positioning detection device can directly send a shutdown signal to the laser to ensure that the laser can be shut down within 30μs in case of failure, avoiding medical accidents caused by the expansion of the fault; the fault alarm signal simultaneously triggers the equipment's audible and visual alarm to remind medical staff to deal with it in time, which meets the safety redundancy requirements of medical equipment.

[0119] Based on the aforementioned scheme, in order to further address the problem of insufficient synergy in existing laser control and poor treatment effects, after determining that the galvanometer motor has deflected into position, the laser's operating parameters are dynamically adjusted according to the real-time vibration signal amplitude. If the vibration signal amplitude is not greater than half of the vibration threshold, the laser outputs power according to the set value. If the vibration signal amplitude is greater than half of the vibration threshold but less than the vibration threshold, the laser automatically reduces the preset power and shortens the preset pulse width. Specifically, after the application controller receives the positioning feedback signal from the intelligent positioning detection device, it dynamically adjusts the operating parameters of the medical laser based on the vibration signal amplitude synchronously transmitted by the intelligent positioning detection device. Taking the treatment of the torso as an example, if the vibration amplitude is ≤0.005g, the medical laser outputs power according to the set power, of which the fractional treatment power is 10-20W and the cutting power is 25-40W; if the vibration amplitude is 0.005g<A≤0.01g, the medical laser automatically reduces the power by 10% and shortens the pulse width by 30%. When the laser cutting mode switches tracks, the intelligent positioning detection device judges the motor jump status through the inner loop control signal. When the drive current rises to above 90mA and the speed rises to above 0.8° / ms, the medical laser is kept off until the motor jumps to the new track and meets the positioning judgment conditions, and then a positioning feedback signal is sent to control the laser to turn on.

[0120] The core purpose of dynamic laser parameter adjustment is to address the shortcomings of existing technologies that only fix the laser switch and do not consider the impact of motor vibration on treatment effectiveness. The above solution uses the linkage between "vibration state and laser parameters" to avoid laser energy superposition or irregular cutting. When there is slight vibration in the motor (vibration amplitude 0.005g < A ≤ 0.01g), reducing the power by 10% can reduce the laser energy output per unit time, and shortening the pulse width by 30% can avoid overlapping laser projection positions caused by vibration. The combination of these two measures ensures uniform energy in the treatment area and avoids problems such as local burns or uneven depth of the laser beam.

[0121] Power and pulse width parameter matching: the fractional treatment power is set to 10-20W (which meets the safe power range for fractional skin treatment), and the cutting power is 25-40W (which meets the energy requirements for the excision of small lesions); the default pulse width is 20μs, which is shortened to 14μs when vibration exceeds the limit, so as to ensure the treatment effect and avoid safety risks.

[0122] Coordinated control of cutting trajectory switching: Existing technologies do not consider the cutting deviation caused by motor instability during trajectory switching. The above solution determines the switching state through current and speed signals (drive current > 90mA, speed > 0.8° / ms corresponds to the motor acceleration switching stage). At this time, the laser is kept off to avoid missed or multiple cuts caused by laser projection position drift during the switching process. When the motor switches to the new trajectory, the current drops to 30-50mA, the speed is ≤ 0.05° / ms, and the angle error meets the standard, the laser is then turned on to ensure that the smoothness deviation of the cutting edge is ≤ 0.01mm, which is 80% better than the existing technology.

[0123] Based on the aforementioned solution, to further address the shortcomings of insufficient response coordination and long synchronization delay in existing technologies, the intelligent positioning detection device can also incorporate a path prediction feedforward module. When the application controller sends the current motor deflection command, it simultaneously transmits the deflection direction and angle increment of the next command to the intelligent positioning detection device. The intelligent positioning detection device determines the command type based on the angle increment as prediction information. The intelligent positioning detection device dynamically adjusts the sampling frequency based on the prediction information: if the prediction is a small angle jump command (angle increment ≤ 0.1°), the sampling frequency is increased from 20kHz to 50kHz. After receiving the position feedback signal, the application controller sends a switch control signal to the medical laser after a 30μs delay, with the delay time precisely matched to the response time of the medical laser (≤ 30μs).

[0124] The purpose of the path prediction feedforward design is as follows: Existing technologies lack a prediction mechanism and have a fixed sampling frequency. When making small-angle jumps, insufficient sampling density leads to judgment delays. The above scheme dynamically adjusts the sampling frequency through prediction information. When making small-angle jumps (such as the jump between adjacent points in fractional therapy, where the angle increment is ≤0.1°), the sampling frequency is increased to 50kHz, and the sampling interval is shortened from 50μs to 20μs, which can capture the motor's position status more quickly and shorten the response delay.

[0125] Sampling frequency adjustment is suitable for the following scenarios: Small-angle jump scenarios (such as fractional therapy) have high requirements for response speed, and increasing the sampling frequency can ensure the real-time judgment of the position; Large-angle continuous command scenarios (such as long trajectory cutting) have lower requirements for sampling density, and maintaining a sampling frequency of 20kHz can balance response speed and signal processing efficiency, avoiding resource waste caused by oversampling.

[0126] Laser response time matching design: The typical response time of medical lasers (high-precision medical scenarios) is ≤30μs. Existing technologies do not take this delay into account. Directly sending a switch signal will cause the laser to start and stop or advance. This solution sets a 30μs delay to ensure that the laser starts immediately after the motor is in position and is precisely synchronized with the motor stop state when the laser is turned off. The synchronization response delay is ≤30μs, which is significantly shorter than existing technologies and further improves the accuracy of treatment effects.

[0127] The data in each embodiment of this application are examples under specific scenarios. The actual data settings should be based on clinical trials. In addition, the specific data required for different medical scenarios are different, and the required accuracy is also different. They will not be listed one by one here.

[0128] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for detecting the deflection of a galvanometer motor in a medical galvanometer scanning system, characterized in that, include: Establish a dynamic threshold library and call the dynamic threshold group adapted to the current scene. The dynamic threshold group includes the angle error threshold, vibration threshold and driver temperature threshold. Real-time acquisition of inner loop control signals, which include at least motor angle error signals, vibration signals, and driver temperature signals; The collected signals are quantized and fused using a multi-signal weight fusion algorithm, and a preliminary judgment is made. In the weight allocation of the multi-signal weight fusion algorithm, the proportion of the motor angle error signal is greater than the sum of the proportions of the vibration signal and the driver temperature signal. The signals initially judged are compared with the dynamic threshold group to determine whether each signal meets the corresponding threshold condition. If all conditions are met, the galvanometer motor is determined to be deflected into position.

2. The method of claim 1, wherein the step of detecting the deflection of the galvanometer motor in the medical galvanometer scanning system is characterized by, The dynamic threshold library is equipped with a self-learning mechanism, which includes: First run trigger: Automatically selects multiple sets of typical medical scenario instructions, sends them to the motor driver in sequence, synchronously collects the steady-state data of the motor under each set of instructions, and generates an initial threshold library after processing the data; Triggered after a preset number of runs: The compensation value is determined by the deviation between the currently collected steady-state data and the historical threshold. The compensated threshold does not exceed the preset safety threshold limit.

3. The method of claim 2, wherein the step of detecting the deflection of the galvanometer motor in the medical galvanometer scanning system is characterized by, Before each compensation is triggered, multiple sets of data are used to cross-validate whether compensation should be performed. If compensation is performed, the current threshold library is automatically backed up before each compensation. If, during treatment verification after compensation, a laser projection deviation exceeding the safety limit is detected, the system automatically rolls back to the threshold library before compensation.

4. The method of claim 2, wherein the step of detecting the deflection of the galvanometer motor in the medical galvanometer scanning system is characterized by, The initial threshold library is synchronized to the cloud platform. When a new device starts up, the initial threshold library of the same model device is downloaded directly, and the self-learning process triggered on the first run is canceled.

5. The method of claim 1, wherein the step of detecting the deflection of the galvanometer motor in the medical galvanometer scanning system is characterized by, When determining whether the galvanometer motor has deflected to the correct position, auxiliary determination conditions are also included: Whether the angle error signal remains within the angle error threshold range for a time greater than or equal to a preset time; Is the drive current signal within the steady-state current range? If any one of the conditions is not met, the galvanometer motor is determined not to have deflected into position.

6. The method of claim 1, wherein the step of detecting the deflection of the galvanometer motor in the medical galvanometer scanning system further comprises the step of: After determining that the galvanometer motor has deflected into position, the laser's operating parameters are dynamically adjusted based on the real-time vibration signal amplitude. If the vibration signal amplitude is not greater than half of the vibration threshold, the laser outputs power according to the set value. If the vibration signal amplitude is greater than half of the vibration threshold but less than the vibration threshold, the laser automatically reduces the preset power and shortens the preset pulse width. ​ 7. A device for detecting deflection of a galvanometer motor in a medical galvanometer scanning system, which applies the method for detecting deflection of a galvanometer motor in a medical galvanometer scanning system according to any one of claims 1 to 6, characterized in that, It includes an intelligent positioning detection device, an application controller, a motor driver, a galvanometer motor, and a laser; The application controller establishes communication connections with the motor driver, laser, and intelligent positioning detection device, respectively, to generate motor deflection commands according to scenario requirements and send them to the motor driver, while simultaneously synchronizing current scenario information with the intelligent positioning detection device; The motor driver establishes a drive connection with the galvanometer motor to drive the galvanometer motor to perform deflection action and outputs an inner loop control signal reflecting the motor's operating status to the intelligent positioning detection device. The intelligent positioning detection device is configured to: establish a dynamic threshold library adapted to different scenarios; call the dynamic threshold group corresponding to the current scenario based on the scenario information synchronized by the application controller, wherein the dynamic threshold group includes an angle error threshold, a vibration threshold, and a driver temperature threshold; collect the inner loop control signal output by the motor driver in real time, wherein the inner loop control signal includes at least a motor angle error signal, a vibration signal, and a driver temperature signal; and perform quantization and fusion of the collected inner loop control signal and complete a preliminary judgment through a multi-signal weight fusion algorithm, wherein in the weight allocation of the multi-signal weight fusion algorithm, the proportion of the motor angle error signal is greater than the sum of the proportions of the vibration signal and the driver temperature signal. The signals initially judged will be compared with the dynamic threshold group called one by one to determine whether each signal meets the corresponding threshold condition. If all of them meet the condition, the galvanometer motor will be determined to be in position and a position feedback signal will be sent to the application controller. After receiving the bit feedback signal, the application controller sends a switch control signal to the laser to synchronize the laser start / stop with the motor's position status.

8. The apparatus for detecting the deflection of a galvanometer motor in a medical galvanometer scanning system according to claim 7, wherein The intelligent positioning detection device is also configured to preprocess the acquired inner loop control signal using electromagnetic shielding and Kalman filtering algorithms.

9. The apparatus for detecting the deflection of a galvanometer motor in a medical galvanometer scanning system according to claim 7, wherein The intelligent positioning detection device has a built-in fault self-testing module, which is used to monitor the integrity of the inner loop control signal in real time. If a signal disconnection, excessive noise, or driver temperature exceeds a preset value is detected, it immediately sends dual safety signals: one is a fault alarm signal, which is transmitted to the application controller; the other is an emergency stop signal, which is transmitted directly to the laser to control the laser to shut down in an emergency.

Images