A parameter optimization method of shuluo instrument based on subjective and objective data fusion

Abstract

The application provides a kind of based on subjective and objective data fusion's shuoluo instrument parameter optimization method, it is related to shuoluo instrument technical field, including the following steps: according to the initial electrophysiological signal of shuoluo instrument to target acupoint detection, generate initial electrical stimulation parameters and start output;In the initial stage of starting output, receive the subjective trigger signal of user feedback, which characterizes tolerable critical point, and record the stimulation intensity value corresponding to subjective trigger signal as the current personal comfort threshold of this user;With current personal comfort threshold as the basis intensity, carry out continuous stimulation, and at the same time, the following parallel steps are executed: real-time acquisition of real-time subjective comfort feedback signal input by user through interactive interface, continuous or discrete;Real-time monitoring of the dynamic electrophysiological signal of target acupoint, and analyzing its change trend to generate objective adjustment suggestion signal;Fusion analysis is carried out on real-time subjective comfort feedback signal and objective adjustment suggestion signal, and current electrical stimulation parameters are dynamically adjusted in real time accordingly.

Description

Technical Field

[0001] This application relates to the field of Shuluo instrument technology, specifically to a method for optimizing Shuluo instrument parameters based on the fusion of subjective and objective data. Background Technology

[0002] In existing technologies, meridian therapy devices utilizing electrical stimulation primarily rely on two modes for setting output parameters. One mode automatically generates a fixed set of stimulation parameters based on the instrument's detection of initial electrical parameters (such as resistance and impedance) at the user's acupoints, combined with a preset general or graded model. This method ignores the dynamic changes in subjective feelings and tolerance among individual users during real-time therapy, often resulting in excessive stimulation in pursuit of effectiveness, causing significant discomfort such as burning pain and muscle tension, leading to poor user compliance. The other mode completely empowers the user to adjust the parameters manually to a comfortable intensity based on their own feelings. While this method avoids strong discomfort, the lack of guidance on objective physiological response trends often results in stimulation levels remaining low, failing to reach the effective therapy threshold and leading to poor results.

[0003] Therefore, how to make the output parameters of the Shuluo device respond simultaneously to the user's real-time subjective feelings and objective physiological signals, and dynamically approach the effective intensity while ensuring comfort, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0004] In view of the above-mentioned defects or deficiencies in the prior art, this application aims to provide a method for optimizing the parameters of a shuluo instrument based on the fusion of subjective and objective data, including the following steps:

[0005] Based on the initial electrophysiological signals detected by the Shuluo instrument at the target acupoint, initial electrical stimulation parameters are generated, and the output is started based on the initial electrical stimulation parameters;

[0006] In the initial stage of starting the output, the system receives a subjective trigger signal from the user that represents the tolerable threshold, and records the stimulus intensity value corresponding to the subjective trigger signal as the user's current personal comfort threshold.

[0007] Continuous stimulation is applied based on the current personal comfort threshold, while simultaneously performing the following parallel steps:

[0008] Real-time subjective comfort feedback signals, whether continuous or discrete, are acquired from user input via the interactive interface.

[0009] Real-time monitoring of dynamic electrophysiological signals at target acupoints and analysis of their changing trends to generate objective regulatory suggestion signals;

[0010] The real-time subjective comfort feedback signal and the objective adjustment suggestion signal are fused and analyzed, and the current electrical stimulation parameters are dynamically adjusted in real time accordingly.

[0011] According to the technical solution provided in this application, the step of fusing and analyzing the real-time subjective comfort feedback signal and the objective adjustment suggestion signal, and dynamically adjusting the current electrical stimulation parameters in real time accordingly, includes the following steps:

[0012] When the trend of the dynamic electrophysiological signal meets the first preset condition in terms of duration and amplitude, the trigger validity threshold is determined and an objective effective signal is generated.

[0013] When the amplitude and speed of the real-time subjective comfort feedback signal exceeding the preset comfort range meet the second preset condition, it is determined that the discomfort trigger threshold is triggered and an discomfort activation signal is generated.

[0014] Based on the objective effectiveness signal or the inappropriate effectiveness signal, the current electrical stimulation parameters are dynamically adjusted in real time.

[0015] According to the technical solution provided in this application, the real-time dynamic adjustment of the current electrical stimulation parameters based on the objective effectiveness signal or the discomfort effectiveness signal includes the following steps:

[0016] If the discomfort activation signal is triggered, a protective attenuation command is immediately generated to reduce the current stimulation intensity to no higher than the current personal comfort threshold.

[0017] If the objective activation signal is triggered and the real-time subjective comfort feedback signal stabilizes within the preset comfort range, a tentative enhancement instruction is generated; the magnitude of the stimulus intensity increment controlled by the tentative enhancement instruction is dynamically calculated according to preset rules.

[0018] The preset rule satisfies the following: the magnitude of the stimulus intensity increment is positively correlated with the rate of change of the dynamic electrophysiological signal, and negatively correlated with the difference between the current stimulus intensity value and the stimulus intensity value before the most recent reduction in intensity due to triggering the discomfort effect signal in the historical record.

[0019] According to the technical solution provided in this application, the method further includes the following steps:

[0020] If the following conditions are met simultaneously during a complete treatment cycle, the current personal comfort threshold will be updated at the end of the current treatment cycle:

[0021] No new adverse effect signal was triggered.

[0022] By performing adjustments based on the objective effectiveness signal, the stimulation intensity is successfully increased and stabilized at a level higher than the current personal comfort threshold recorded at the beginning of this conditioning cycle, and this stable state continues for a preset duration.

[0023] According to the technical solution provided in this application, updating the current personal comfort threshold at the end of the current conditioning cycle includes the following steps:

[0024] From the time-series data of stimulation intensity in this conditioning cycle, extract the continuous data segment from the moment when the stimulation intensity first reaches and remains above the current personal comfort threshold recorded in the initial record until the end of conditioning or the end of the stable state, as the effective reinforcement data segment;

[0025] Based on the real-time subjective comfort feedback signal and dynamic electrophysiological signal corresponding to the effective enhanced data segment, the dynamic credibility weight of each sampling point in the effective enhanced data segment is calculated.

[0026] The stimulus intensity values ​​of all sampling points within the effective enhanced data segment are weighted and averaged according to their corresponding dynamic confidence weights, and the calculation result is used as the updated current personal comfort threshold.

[0027] According to the technical solution provided in this application, the calculation of the dynamic confidence weight of each sampling point within the effective enhanced data segment includes the following steps:

[0028] For any sampling point within the effective enhanced data segment, its dynamic credibility weight is calculated based on the subjective and objective data at that point, where:

[0029] The absolute value of the deviation of the real-time subjective comfort feedback signal value at the sampling point from the center value of the preset comfort range is obtained, and the absolute value is input into the first mapping function to obtain the subjective factor value. The first mapping function is configured such that the output value decreases monotonically as the input value increases.

[0030] The positive change amplitude of the dynamic electrophysiological signal value at the sampling point compared with the resting baseline value before conditioning is obtained, and the change amplitude is input into the second mapping function to obtain the objective factor value. The second mapping function is configured such that the output value increases monotonically with the increase of the input value.

[0031] The subjective factor value is multiplied by the objective factor value and then normalized to obtain the dynamic credible weight.

[0032] According to the technical solution provided in this application, the calculation of the dynamic confidence weight of each sampling point within the effective enhanced data segment includes the following steps:

[0033] For each sampling point, the instantaneous subjective confidence of the corresponding real-time subjective comfort feedback signal and the instantaneous objective confidence of the dynamic electrophysiological signal are calculated respectively.

[0034] Based on the numerical relationship between the instantaneous subjective confidence and the instantaneous objective confidence, the dynamic confidence weight of each sampling point is determined;

[0035] The instantaneous subjective confidence is positively correlated with the stability of the user's operation of the interactive interface and / or the smoothness of the real-time subjective comfort feedback signal over time; the instantaneous objective confidence is positively correlated with the signal-to-noise ratio of the dynamic electrophysiological signal and / or the continuity of its changing trend within a short time window.

[0036] According to the technical solution provided in this application, determining the dynamic confidence weight of each sampling point based on the numerical relationship between the instantaneous subjective confidence and the instantaneous objective confidence includes the following steps:

[0037] If both the instantaneous subjective confidence and the instantaneous objective confidence are higher than their respective preset high confidence thresholds, then the sampling point is assigned the highest basic weight value.

[0038] If one of them is lower than its preset low confidence threshold, the basic weight is determined mainly based on the high confidence signal, and a compensatory attenuation is applied to it based on the historical confidence pattern.

[0039] If both the instantaneous subjective confidence and the instantaneous objective confidence are within the medium confidence range, the basic weight is determined based on the weighted sum of their confidence levels, and the gain is adjusted in conjunction with the consistency of their trends. Gain is obtained when the trends are consistent, and attenuation is achieved when the trends diverge.

[0040] According to the technical solution provided in this application, the method further includes constructing and updating a long-term personalized user model, comprising the following steps:

[0041] Record key data for each complete treatment cycle to form a personal historical dataset; the key data includes: the current personal comfort threshold at the beginning of the current treatment cycle, whether the threshold is updated at the end of the cycle and the updated current personal comfort threshold, the total number of discomfort signals triggered during the cycle, and the average dynamic confidence weight of the effective reinforcement data segment.

[0042] Based on the personal historical dataset, a predictive relationship between the current personal comfort threshold and the user's basic physiological information and historical conditioning effects is established using machine learning models or statistical regression models.

[0043] The long-term personalized model is used to optimize at least one of the following:

[0044] Predict and recommend initial electrical stimulation parameters before the start of subsequent treatment cycles;

[0045] The determination threshold in the first preset condition and / or the second preset condition is dynamically adjusted.

[0046] According to the technical solution provided in this application, the method further includes the following steps:

[0047] Based on the aforementioned long-term personalized model, a phased treatment path is planned when a user starts a new treatment cycle. The phased treatment path includes at least a first phase and a second phase executed sequentially.

[0048] In the first stage, a conservative enhancement strategy is adopted to safely confirm the current personal comfort threshold, wherein the magnitude of the stimulus intensity increment controlled by the tentative enhancement instruction is limited to a first upper limit value.

[0049] In the second stage, an active enhancement strategy is adopted to guide the stimulus intensity toward the recommended intensity range predicted by the model, wherein the incremental amplitude of the stimulus intensity is set to be no less than a second lower limit value, and the second lower limit value is higher than the first upper limit value.

[0050] The phased treatment path is dynamically adjusted according to the frequency of the discomfort effect signals triggered in real time; when the frequency of the discomfort effect signals triggered in the second phase exceeds the pre-calculated safety threshold, the second phase is interrupted and the process returns to the first phase.

[0051] Compared with existing technologies, the beneficial effects of this application are: achieving a dynamic balance between effectiveness and comfort, and improving user experience and treatment compliance. This method introduces the current personal comfort threshold as a benchmark for personalized regulation. This threshold is directly derived from the user's own tolerance threshold, ensuring from the outset that the initial intensity of the stimulus is within a safe and comfortable range. The scheme acquires real-time subjective comfort feedback signals and objective dynamic electrophysiological signals in parallel, and performs fusion analysis on both, thereby driving real-time dynamic adjustment of parameters. This makes the treatment process no longer a unidirectional, fixed output, but a closed-loop system of perception-feedback-regulation: when the user feels comfortable, the system can cautiously enhance the stimulus to improve the effect based on the effectiveness trend of objective physiological signals; simultaneously, once it receives subjective feedback of discomfort, it can immediately make protective adjustments. Thus, in a single treatment session, it can dynamically find optimized parameters that approximate the user's personal effective-comfort boundary, fundamentally solving the contradiction of the difficulty in balancing these two aspects in traditional methods. Attached Figure Description

[0052] Figure 1 The flowchart of the steps of the Shuluo instrument parameter optimization method based on subjective and objective data fusion provided in this application. Detailed Implementation

[0053] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0054] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0055] Example 1

[0056] As mentioned in the background section, to address the problems in the existing technology, this application proposes a method for optimizing the parameters of a shuluo instrument based on the fusion of subjective and objective data, such as... Figure 1 As shown, it includes the following steps:

[0057] S1. Based on the initial electrophysiological signal detected by the Shuluo instrument on the target acupoint, generate initial electrical stimulation parameters, and start the output based on the initial electrical stimulation parameters;

[0058] S2. In the initial stage of starting the output, receive the subjective trigger signal from the user that represents the tolerable threshold, and record the stimulus intensity value corresponding to the subjective trigger signal as the user's current personal comfort threshold.

[0059] S3. Provide continuous stimulation based on the current personal comfort threshold, while simultaneously performing the following parallel steps:

[0060] S31. Real-time subjective comfort feedback signals, whether continuous or discrete, input by the user through the interactive interface.

[0061] S32. Monitor the dynamic electrophysiological signals of the target acupoints in real time and analyze their changing trends to generate objective regulatory suggestion signals;

[0062] S4. The real-time subjective comfort feedback signal and the objective adjustment suggestion signal are fused and analyzed, and the current electrical stimulation parameters are dynamically adjusted in real time accordingly.

[0063] Specifically, this method begins with the detection of initial electrophysiological signals at the target acupoint. Here, the target acupoint is pre-selected based on Traditional Chinese Medicine (TCM) diagnostic criteria or treatment goals, such as common acupoints like Zusanli (ST36) and Hegu (LI4). The initial electrophysiological signal mainly refers to the electrical characteristic parameters of the skin at the acupoint, such as impedance and conductivity, measured by the Shuluo device's electrodes before stimulation is applied. Generating initial electrical stimulation parameters means that the system's built-in algorithm model (such as a lookup table based on common population data or a simple linear regression model) maps the initial stimulation intensity, frequency, and waveform parameters based on the detected initial electrophysiological signal values. For example, an acupoint with higher resistance might correspond to a medium-intensity initial stimulation parameter. Subsequently, the Shuluo device initiates electrical pulse output based on this set of initial parameters.

[0064] In the initial phase after output is initiated (e.g., the first 30 seconds to 1 minute), the system prompts the user for feedback through an interactive interface (such as a button on the device or a virtual slider on the accompanying mobile app). As the stimulation intensity gradually increases from zero, a subjective trigger signal is activated when the user first experiences a clear, tolerable, mild tingling or throbbing sensation (rather than sharp pain). This signal can be a single click or long press by the user. The system captures the moment this action occurs and immediately records the output stimulation intensity value at that moment, defining it as the user's current personal comfort threshold for the current therapy session. This threshold serves as the absolute safety baseline for personalized therapy.

[0065] Subsequently, the system provides continuous stimulation based on this individual comfort threshold. Simultaneously, the system executes two key data acquisition processes in parallel. First, it continuously acquires real-time subjective comfort feedback signals input by the user through an interactive interface. This interface can be a continuous slider, which the user can slide left or right at any time to indicate the intensity of the sensation, from too weak to too strong; or it can be a series of discrete buttons, such as weaken, suitable, and strengthen. Second, it monitors the dynamic electrophysiological signals of the target acupoints in real time. Under continuous stimulation, parameters such as conductivity of the acupoints change due to variations in local blood circulation and nerve excitability. The system obtains a curve showing the change of this signal over time through continuous sampling.

[0066] The next step is the fusion analysis. The system does not process these two signals independently, but rather performs a correlation analysis. For example, a simple fusion logic might be: when the objective dynamic electrophysiological signal shows a stable upward trend (suggesting that the stimulus may be producing a physiological effect), and the user's subjective comfort feedback remains stable within the comfort range, the system determines that there is room for enhancement. Conversely, if the user's subjective feedback rapidly becomes too strong, the system determines that it needs to be weakened regardless of the objective signal. Based on the results of this fusion analysis, the system generates control commands to dynamically adjust the currently outputting electrical stimulation parameters (mainly intensity) in real time, with small amplitude adjustments, such as fine-tuning once per second or every few seconds.

[0067] This implementation method breaks through the limitations of traditional nerve-relieving devices that either require fixed parameters or are entirely manually adjustable. For the first time, it establishes a dynamic closed-loop adjustment system in a single treatment session, with the user's real-time comfort as a safety constraint and changes in objective physiological signals as a therapeutic guide. Its technical principle lies in transforming the user from a passive recipient to an active feedback participant, digitizing this subjective feeling, and synchronizing and correlating it with the objective electrophysiological data collected by the instrument for decision-making. This allows the output intensity to fluctuate in an evidence-based and responsive manner above a safe, personalized baseline. This achieves personalized treatment from a one-size-fits-all approach on a macro level to a tailored approach for each individual, and on a micro level, it achieves a dynamic balance and synergistic optimization of comfort and effectiveness, significantly improving user experience and treatment adherence.

[0068] In a preferred embodiment, the step of fusing and analyzing the real-time subjective comfort feedback signal and the objective adjustment suggestion signal, and dynamically adjusting the current electrical stimulation parameters in real time accordingly, includes the following steps:

[0069] When the trend of the dynamic electrophysiological signal meets the first preset condition in terms of duration and amplitude, the trigger validity threshold is determined and an objective effective signal is generated.

[0070] When the amplitude and speed of the real-time subjective comfort feedback signal exceeding the preset comfort range meet the second preset condition, it is determined that the discomfort trigger threshold is triggered and an discomfort activation signal is generated.

[0071] Based on the objective effectiveness signal or the inappropriate effectiveness signal, the current electrical stimulation parameters are dynamically adjusted in real time.

[0072] Specifically, firstly, the system quantifies the changing trends of the monitored dynamic electrophysiological signals. This trend can be obtained by calculating the first derivative (rate of change) or mean change of the signal within a sliding time window (e.g., the past 10 seconds). The first preset condition is a comprehensive requirement for both duration and amplitude; for example, it can be set to a positive rate of change for 30 consecutive seconds, with the cumulative amplitude exceeding 5% of the initial value. When this condition is met, the system determines that the current stimulus may be producing an effective physiological response, thereby triggering an internal effectiveness threshold and generating an objective effectiveness signal. This signal is a logical flag indicating the existence of objective evidence of enhancing stimulation.

[0073] Secondly, the system analyzes real-time subjective comfort feedback signals. The preset comfort range is a pre-defined range on the interactive interface; for example, in a range of 0-100, 40-60 is considered the comfort range. The system not only monitors whether the feedback signal exceeds this range but also monitors the magnitude and speed of its exceedance. A second preset condition is a strict determination of this exceedance behavior. For example, it can be set that the feedback signal spikes from 55 to 80 within 3 seconds (a large and rapid exceedance), rather than a slow movement from 58 to 65 within 10 seconds. This is to distinguish between accidental user errors or subtle sensory fluctuations and genuine discomfort. When this second preset condition is met, the system determines that the user is experiencing clear discomfort, triggers the discomfort trigger threshold, and generates a discomfort activation signal.

[0074] Finally, the system drives parameter adjustments based on these two generated activation signals. This simplifies the decision-making logic: the adjustment is triggered directly by these two explicit event signals, rather than by complex, continuous calculations of the original data. For example, when an objective activation signal is set, an enhancement process can be triggered; when an inappropriate activation signal is set, a protective attenuation process is immediately triggered.

[0075] The technical principle of this implementation lies in transforming continuous subjective and objective data streams into discrete, meaningful events through preset conditional judgments based on duration and amplitude. This brings multiple benefits: First, it improves system stability and avoids frequent erroneous adjustments caused by minor signal fluctuations; second, it enhances interpretability, providing clear signal basis for every important decision made by the system; and third, it provides clear triggering conditions for subsequent, more complex adjustment strategies. It transforms both subjective feelings (discomfort) and objective bodily reactions (effectiveness) into standardized events that machines can recognize and respond to, representing a crucial step in building intelligent and automated adjustment systems.

[0076] In a preferred embodiment, the real-time dynamic adjustment of the current electrical stimulation parameters based on the objective effectiveness signal or the discomfort effectiveness signal includes the following steps:

[0077] If the discomfort activation signal is triggered, a protective attenuation command is immediately generated to reduce the current stimulation intensity to no higher than the current personal comfort threshold.

[0078] If the objective activation signal is triggered and the real-time subjective comfort feedback signal stabilizes within the preset comfort range, a tentative enhancement instruction is generated; the magnitude of the stimulus intensity increment controlled by the tentative enhancement instruction is dynamically calculated according to preset rules.

[0079] The preset rule satisfies the following: the magnitude of the stimulus intensity increment is positively correlated with the rate of change of the dynamic electrophysiological signal, and negatively correlated with the difference between the current stimulus intensity value and the stimulus intensity value before the most recent reduction in intensity due to triggering the discomfort effect signal in the historical record.

[0080] Specifically, when the system detects that a discomfort signal has been triggered, its response is rapid and rigid. The system immediately generates a protective attenuation command, the core action of which is to quickly reduce the current stimulus intensity (which may already be higher than the individual's comfort threshold), with the goal of not exceeding the previously recorded current personal comfort threshold. This is a safety mechanism to ensure that whenever the user clearly indicates discomfort, the intensity can quickly drop back to a known safe baseline.

[0081] When the system detects that an objective activation signal has been triggered, and simultaneously the real-time subjective comfort feedback signal remains stable within a preset comfort range (e.g., no discomfort signal is triggered for 10 consecutive seconds, and the feedback value is in the middle of the comfort range), the system generates a tentative enhancement instruction. The core of this instruction is to calculate and apply an increment in stimulus intensity. The magnitude of this increment is not a fixed value, but is determined according to a dynamically calculated preset rule. This rule contains two mutually constraining factors:

[0082] First, the magnitude of the increase is positively correlated with the rate of change of dynamic electrophysiological signals. This means that the more positive and stronger the objective physiological response (e.g., the faster the conductivity increases), the more confident the system is, and the greater the allowable increase can be.

[0083] Secondly, the increment magnitude is negatively correlated with the difference between the current stimulus intensity value and the stimulus intensity value before the most recent reduction in intensity due to triggering an discomfort signal in the historical record. Here, the difference refers to how far the current intensity is from the dangerous intensity that last triggered discomfort. The closer the distance (smaller the difference), the closer it is to the safety boundary, and the enhancement must be extremely cautious, with the increment magnitude being very small or even zero; the farther the distance (larger the difference), the larger the safety buffer space, and the relatively larger the increment magnitude can be.

[0084] For example, the system can set a basic increment unit ΔI_base. The final increment ΔI = ΔI_base × (current rate of change / reference rate of change) × (safety margin / maximum historical margin). Among them, the rate of change factor promotes enhancement, while the safety margin factor restricts enhancement.

[0085] This implementation achieves tentative enhancement through intelligent, gradual exploration within safe boundaries. Its technical principle involves introducing the key concept of a safe distance based on historical experience and coupling it with the current objective response intensity. This simulates the human learning process: the system not only reacts to current positive signals (objective effectiveness) but also remembers the intensity levels of past negative outcomes (discomfort), using this as a basis to weigh the aggressiveness of current actions. This makes the system's enhancement behavior no longer blind or linear, but possesses the ability to remember and judge. While encouraging effective reinforcement, it incorporates a braking mechanism that gradually strengthens as it approaches historical danger points, thereby greatly improving the safety, stability, and intelligence of exploring the upper limit of effectiveness within the comfort zone. This is the core algorithmic guarantee for achieving dynamic equilibrium.

[0086] In a preferred embodiment, the method further includes the following steps:

[0087] If the following conditions are met simultaneously during a complete treatment cycle, the current personal comfort threshold will be updated at the end of the current treatment cycle:

[0088] No new adverse effect signal was triggered.

[0089] By performing adjustments based on the objective effectiveness signal, the stimulation intensity is successfully increased and stabilized at a level higher than the current personal comfort threshold recorded at the beginning of this conditioning cycle, and this stable state continues for a preset duration.

[0090] Specifically, this is implemented at the end of a complete treatment cycle (e.g., a 20-minute treatment). Updating the current personal comfort threshold does not occur every time, but requires two strict prerequisites to be met simultaneously:

[0091] Condition 1: No new discomfort signals were triggered during the entire treatment period. This means that the entire treatment was completed within the user's subjective comfort and safety range, and no discomfort events requiring emergency intervention occurred. This is the updated safety prerequisite.

[0092] Condition 2: By implementing adjustments based on objective effectiveness signals (i.e., tentative enhancements), the stimulation intensity was successfully increased and stabilized at a level higher than the current personal comfort threshold recorded at the beginning of this treatment cycle, and this stable state was maintained for a preset duration. For example, the initial comfort threshold was intensity 5. During the treatment, through several enhancements, the intensity was increased to 7, and in the last 8 minutes, the intensity remained stable around 7 without dropping. The preset duration here can be set to 5 minutes. This demonstrates that the user not only tolerated the higher intensity but that their body may have adapted, which is a prerequisite for the effectiveness of the update.

[0093] The system will initiate an update procedure at the end of the current treatment cycle only if both of the above conditions are met simultaneously. The update operation involves replacing the previously stored current personal comfort threshold with the higher intensity value that was successfully reached and maintained stably (or the average intensity value during that stable phase). For example, the threshold will be updated from 5 to 7.

[0094] This implementation achieves dynamic evolution of individual comfort thresholds, extending the personalized adaptability of the device from a single session to continuous optimization across sessions. Its technical principle lies in solidifying a successful, safe intensity increase experience as a new, higher tolerance baseline for the user. It defines a clear learning moment: only when the user stably accepts higher intensity stimulation without discomfort does the system recognize that the user's tolerance level may have improved and update their personal model accordingly. This breaks the traditional model of unchanging device parameters and differs from instantaneous adjustments relying solely on single feedback. It allows the device to increase its intensity as the user's physical condition improves or their adaptability increases, guiding a gradual and safe increase in therapeutic dosage. In the long term, this effectively prevents users from remaining in an inefficient comfort zone for extended periods, instead encouraging them to gradually increase the stimulation dosage under painless conditions, thus enabling the attainment of better and deeper therapeutic effects.

[0095] In a preferred embodiment, updating the current personal comfort threshold at the end of the current treatment cycle includes the following steps:

[0096] From the time-series data of stimulation intensity in this conditioning cycle, extract the continuous data segment from the moment when the stimulation intensity first reaches and remains above the current personal comfort threshold recorded in the initial record until the end of conditioning or the end of the stable state, as the effective reinforcement data segment;

[0097] Based on the real-time subjective comfort feedback signal and dynamic electrophysiological signal corresponding to the effective enhanced data segment, the dynamic credibility weight of each sampling point in the effective enhanced data segment is calculated.

[0098] The stimulus intensity values ​​of all sampling points within the effective enhanced data segment are weighted and averaged according to their corresponding dynamic confidence weights, and the calculation result is used as the updated current personal comfort threshold.

[0099] Specifically, the implementation begins after the conditioning cycle ends and the update conditions are met. First, the system extracts a specific effective reinforcement data segment from all the time-series data of stimulus intensity recorded in this conditioning cycle. The starting point of this data segment is defined as the moment when the stimulus intensity first reaches and remains above the initially recorded current personal comfort threshold. For example, if the initial threshold is intensity 5, and the intensity first starts at 5.1 in the 3rd minute after the conditioning begins and remains above 5.1 for at least the next minute, then this 3rd minute is the starting point of the data segment. Its ending point is the moment the conditioning ends (e.g., at the 20-minute mark), or the moment the stable state before the conditioning ends (e.g., if the intensity drops back and remains below the initial threshold of 5 for some reason at the 15th minute). This continuous data segment is identified by the system as the core phase of effective reinforcement in this conditioning period, and is direct evidence that the user has successfully tolerated and may benefit from higher intensities.

[0100] Secondly, based on this, the system effectively strengthens the real-time subjective comfort feedback signal and dynamic electrophysiological signal corresponding to each sampling point (e.g., one data point per second) within the data segment, and calculates a dynamic reliability weight for each point. The core idea of ​​this weight is that the intensity value at every moment within the data segment is not equally reliable or worthy of being used as a reference for a new threshold. A more reliable moment should be the moment when the user feels more comfortable subjectively and the objective physiological response is more positive. For example, the system can design an algorithm that gives higher weights to points where the comfort feedback is closer to the center value of a preset interval (e.g., 50 points) and the electrophysiological signal shows a greater increase compared to the baseline. Conversely, points with high intensity but where the user feedback is close to the comfort zone boundary (e.g., 59 points) or where the electrophysiological response is flat have lower weights.

[0101] Finally, the system performs a weighted average calculation. It effectively amplifies the stimulus intensity value recorded at each sampling point within the data segment, multiplies it by its corresponding dynamic confidence weight, sums all the products, and then divides by the sum of all weights. The result is not a simple arithmetic mean, but a confidence-weighted average. This final calculated value is then set as the updated current personal comfort threshold.

[0102] In a preferred embodiment, calculating the dynamic confidence weight of each sampling point within the effective enhanced data segment includes the following steps:

[0103] For any sampling point within the effective enhanced data segment, its dynamic credibility weight is calculated based on the subjective and objective data at that point, where:

[0104] The absolute value of the deviation of the real-time subjective comfort feedback signal value at the sampling point from the center value of the preset comfort range is obtained, and the absolute value is input into the first mapping function to obtain the subjective factor value. The first mapping function is configured such that the output value decreases monotonically as the input value increases.

[0105] The positive change amplitude of the dynamic electrophysiological signal value at the sampling point compared with the resting baseline value before conditioning is obtained, and the change amplitude is input into the second mapping function to obtain the objective factor value. The second mapping function is configured such that the output value increases monotonically with the increase of the input value.

[0106] The subjective factor value is multiplied by the objective factor value and then normalized to obtain the dynamic credible weight.

[0107] Specifically, the implementation method performs the following calculations for each sampling point (e.g., time point t) within the effective enhanced data segment: First, subjective data is processed. The system reads the value of the real-time subjective comfort feedback signal corresponding to that point, for example, a scalar S_t between 0 (very weak) and 100 (very strong). The preset comfort range may be 40 to 60. The system calculates the absolute value of the difference between this value S_t and the center value of the comfort range (e.g., 50), i.e., |S_t - 50|, denoted as ΔS_t. This ΔS_t represents the degree to which the subjective feeling deviates from the most comfortable state. Then, ΔS_t is input into a first mapping function f_S. This function is configured such that the output value monotonically decreases as the input value ΔS_t increases. This means that the more the subjective feeling deviates from the center (whether weak or strong), the smaller the output subjective factor value. The specific form of this function can be linear, such as f_S(ΔS_t) = 1 - k1 ×ΔS_t (k1 is a positive constant to ensure the result is non-negative), or it can be non-linear, such as in the form of exponential decay.

[0108] The second step is to process the objective data. The system reads the value O_t of the dynamic electrophysiological signal corresponding to this point and calculates the positive change magnitude of this signal compared to the resting baseline value O_base measured before conditioning, i.e., max(0, O_t - O_base), denoted as ΔO_t. This ΔO_t represents the degree of positive change in the objective physiological response relative to the resting state. ΔO_t is input into a second mapping function f_O, which is configured such that the output value monotonically increases with the input value ΔO_t. This means that the more positive the objective physiological response, the larger the output objective factor value. This function can also be linear, such as f_O(ΔO_t) = k2 × ΔO_t, or a nonlinear form with saturated growth.

[0109] The third step is fusion calculation. The subjective factor value f_S(ΔS_t) obtained from the sampling point is directly multiplied by the objective factor value f_O(ΔO_t) to obtain an initial weight value W_t' = f_S(ΔS_t) × f_O(ΔO_t). This multiplicative relationship means that a low value of any factor will significantly reduce the final weight, reflecting the logic that both subjective and objective conditions must be satisfied simultaneously.

[0110] The fourth step is normalization. Since the range of W_t' values ​​may differ between different sampling points, the initial weights of all sampling points within the entire effective augmented data segment need to be normalized for a fair weighted average. A common approach is to divide all W_t' values ​​by the maximum value or by the sum of all W_t' values, ensuring that the final dynamic confidence weight W_t for each sampling point falls within the range of 0 to 1 and is relatively comparable.

[0111] In a preferred embodiment, calculating the dynamic confidence weight of each sampling point within the effective enhanced data segment includes the following steps:

[0112] For each sampling point, the instantaneous subjective confidence of the corresponding real-time subjective comfort feedback signal and the instantaneous objective confidence of the dynamic electrophysiological signal are calculated respectively.

[0113] Based on the numerical relationship between the instantaneous subjective confidence and the instantaneous objective confidence, the dynamic confidence weight of each sampling point is determined;

[0114] The instantaneous subjective confidence is positively correlated with the stability of the user's operation of the interactive interface and / or the smoothness of the real-time subjective comfort feedback signal over time; the instantaneous objective confidence is positively correlated with the signal-to-noise ratio of the dynamic electrophysiological signal and / or the continuity of its changing trend within a short time window.

[0115] Specifically, the implementation also processes each sampling point in steps: First, the instantaneous subjective confidence level is calculated. This confidence level does not directly originate from the specific numerical value of the user feedback (e.g., 55 points), but rather from the quality of the process that generates this feedback. It is mainly evaluated through two dimensions: First, the stability of the user's interactive interface. For example, if the user provides feedback through a continuous slider, the system can analyze the standard deviation or coefficient of variation of the slider position in a short period of time (e.g., before and after 5 seconds) near the sampling point. The more stable the slider and the less jitter, the more cautious and certain the user feedback is, and the higher the instantaneous subjective confidence level. Conversely, if the slider fluctuates violently, it may indicate that the user is uncertain or has touched it unconsciously, and the confidence level will decrease. Second, the smoothness of the real-time subjective comfort feedback signal over time. The system can evaluate whether its change is gradual by calculating the first or second derivative of the signal within a short time window. A smoothly changing signal is usually more likely to reflect the real change in feeling than a signal with a sharp jump, so the higher the smoothness, the higher the confidence level. The evaluation results of these two dimensions can be weighted and summed to form the instantaneous subjective confidence score CS_t of the sampling point.

[0116] The second step is to calculate the instantaneous objective confidence score. This confidence score assesses the quality of the acquired dynamic electrophysiological signal itself. It mainly considers two dimensions: First, the signal-to-noise ratio (SNR). The system can estimate the SNR through frequency domain analysis (e.g., calculating the ratio of energy in a specific frequency band to noise energy across the entire frequency band) or time domain analysis (e.g., calculating the ratio of signal amplitude to background fluctuation amplitude within a sliding window). A high SNR means clear physiological signals, less susceptibility to environmental interference, and high confidence. Second, the continuity of the signal change trend within a short time window. For example, the system can examine the signal values ​​of several past sampling points (e.g., 5 points) to see if their direction of change is consistent and their amplitude is continuous. A continuous, stable upward or downward trend is more likely to reflect the true physiological process than a chaotic, frequently reversing trend; therefore, the higher the continuity, the higher the confidence score CO_t.

[0117] The third step is to determine the final weights. After obtaining the instantaneous subjective confidence level CS_t and instantaneous objective confidence level CO_t for each sampling point, the system determines the dynamic confidence weight for that point based on the numerical relationship between the two through a pre-defined decision logic. This decision logic is more complex and intelligent than simple multiplication, and is designed to handle complex situations with different combinations of the two confidence levels.

[0118] Furthermore, determining the dynamic confidence weight of each sampling point based on the numerical relationship between the instantaneous subjective confidence and the instantaneous objective confidence includes the following steps:

[0119] If both the instantaneous subjective confidence and the instantaneous objective confidence are higher than their respective preset high confidence thresholds, then the sampling point is assigned the highest basic weight value.

[0120] If one of them is lower than its preset low confidence threshold, the basic weight is determined mainly based on the high confidence signal, and a compensatory attenuation is applied to it based on the historical confidence pattern.

[0121] If both the instantaneous subjective confidence and the instantaneous objective confidence are within the medium confidence range, the basic weight is determined based on the weighted sum of their confidence levels, and the gain is adjusted in conjunction with the consistency of their trends. Gain is obtained when the trends are consistent, and attenuation is achieved when the trends diverge.

[0122] Specifically, the implementation involves three scenario processing scenarios: The first scenario is high-confidence consensus. The system presets two high-confidence thresholds, one for subjective confidence (e.g., C_S_high) and one for objective confidence (e.g., C_O_high). When both the instantaneous subjective confidence (CS_t) and instantaneous objective confidence (CO_t) of a sampling point are higher than their respective high-confidence thresholds, the system determines that the subjective and objective data at this moment are extremely reliable. In this case, the system directly assigns the highest base weight value to the sampling point, for example, W_base = 1.0. This indicates that the data quality at this point is optimal and should have the greatest say in the calculation of the new threshold.

[0123] The second scenario involves a single low-confidence signal. The system also presets two low-confidence thresholds (C_S_low, C_O_low). When either C_t or C_O_t falls below its low-confidence threshold, while the other is above or equal to it, the system determines that one signal is unreliable. In this case, the decision logic does not simply average or discard the signal entirely, but primarily determines the base weight based on the high-confidence signal. For example, if C_t is very low and C_t is very high, the base weight is determined proportionally based on the value of C_t. After determining the base weight, a compensatory attenuation based on historical confidence patterns is applied. For example, the system records the user's historical performance in similar situations (subjective low confidence). If the final confirmation of validity was also low in historical cases, the attenuation is increased. This simulates skepticism towards unreliable sources, and this skepticism is quantified and calibrated based on historical experience.

[0124] The third scenario is moderately fuzzy consensus. When neither CS_t nor CO_t reaches the high confidence threshold, nor falls below the low confidence threshold (i.e., both are within the moderate confidence interval), the system first determines a base weight based on the weighted sum of their confidence levels (e.g., aCS_t + bCO_t). Then, a key adjustment factor is introduced: the degree of consistency in their trends. The system calculates the direction of change of CS_t and CO_t within a short time window (e.g., rising, falling, or stable). If their trends are consistent (e.g., both are rising gently), a gain is given to the base weight, because the consistency between the signals strengthens the possibility that they jointly reflect the true state. If the trends diverge (e.g., one rises and the other falls), the base weight is attenuated, because inconsistency increases uncertainty. The magnitude of this gain or attenuation can be correlated with the strength of the trend or the angle of divergence.

[0125] In a preferred embodiment, the method further includes building and updating a long-term user personalization model, comprising the following steps:

[0126] Record key data for each complete treatment cycle to form a personal historical dataset; the key data includes: the current personal comfort threshold at the beginning of the current treatment cycle, whether the threshold is updated at the end of the cycle and the updated current personal comfort threshold, the total number of discomfort signals triggered during the cycle, and the average dynamic confidence weight of the effective reinforcement data segment.

[0127] Based on the personal historical dataset, a predictive relationship between the current personal comfort threshold and the user's basic physiological information and historical conditioning effects is established using machine learning models or statistical regression models.

[0128] The long-term personalized model is used to optimize at least one of the following:

[0129] Predict and recommend initial electrical stimulation parameters before the start of subsequent treatment cycles;

[0130] The determination threshold in the first preset condition and / or the second preset condition is dynamically adjusted.

[0131] Specifically, its implementation begins with the systematic accumulation of data. At the end of each complete treatment cycle, regardless of whether a threshold update is triggered, the system extracts and stores a set of key data, constituting the user's personal historical dataset. This key data is carefully selected to characterize the features and results of that treatment, and specifically includes:

[0132] The current personal comfort threshold at the beginning of this treatment cycle: represents the initial safety baseline for this treatment.

[0133] Whether to update the threshold at the end of the period and the updated threshold: Record whether the user's tolerance level has improved during this period and the extent of the improvement.

[0134] The total number of discomfort signals triggered within the cycle: This quantifies the frequency of subjective discomfort experienced by the user during the treatment process and serves as a reverse indicator of safety.

[0135] Effectively enhance the average dynamic credibility weight of the data segment: characterize the overall level of subjective and objective data collaboration quality during the intensity enhancement phase, reflecting the quality of this enhancement.

[0136] In addition, the dataset can also be linked to basic physiological information actively input by the user or acquired by the device, such as age, gender, and the location or degree of discomfort reported by the user before the treatment.

[0137] Based on a continuously accumulated personal historical dataset, the system periodically (e.g., after every 10 data points) calls a machine learning model or statistical regression model for training. For example, linear regression, support vector regression, or a lightweight neural network can be used. The model aims to establish a predictive relationship between the current personal comfort threshold and the user's basic physiological information and historical treatment effects. Here, historical treatment effects can be characterized by the aforementioned key data (such as the previous threshold, number of discomfort episodes, and average confidence weight) or their derived features (such as threshold growth trends and changes in discomfort frequency). Through training, the model learns personalized patterns, such as for user A, when reporting mild knee discomfort, their initial comfort threshold is usually about 10% lower than the average, or when user B has zero discomfort episodes in three consecutive treatments and a high average confidence weight, their next treatment threshold has an 80% probability of being raised.

[0138] The trained long-term personalized model is then put into application to optimize subsequent conditioning, mainly in at least one of the following aspects:

[0139] First, before the start of a subsequent treatment cycle, the system predicts and suggests initial electrical stimulation parameters. When the user is ready to begin a new treatment, the system inputs the user's current basic physiological information (such as the selected treatment area) and the latest historical effect characteristics into the model. The model outputs a predicted value for the current personal comfort threshold. The system no longer relies entirely on initial general electrophysiological measurements, but uses this predicted value as an important reference, combining it with real-time detection for fine-tuning, generating personalized initial parameters that are closer to the user's current state, achieving a warm start and shortening the adaptation period.

[0140] Secondly, the judgment thresholds in the first and / or second preset conditions are dynamically adjusted. For example, the model can adjust the duration or magnitude of change required to trigger an objective effect signal (first preset condition) based on the user's long-term performance. For users with good tolerance and positive responses, the conditions can be slightly relaxed, allowing the system to more sensitively capture the effect trend; for sensitive or slow-reacting users, the conditions may be tightened to avoid false triggers leading to overly frequent adjustments. Similarly, the amplitude and speed thresholds for triggering discomfort effect signals (second preset condition) can also be fine-tuned based on the user's historical discomfort patterns to better suit the user's feedback habits.

[0141] In a preferred embodiment, the method further includes the following steps:

[0142] Based on the aforementioned long-term personalized model, a phased treatment path is planned when a user starts a new treatment cycle. The phased treatment path includes at least a first phase and a second phase executed sequentially.

[0143] In the first stage, a conservative enhancement strategy is adopted to safely confirm the current personal comfort threshold, wherein the magnitude of the stimulus intensity increment controlled by the tentative enhancement instruction is limited to a first upper limit value.

[0144] In the second stage, an active enhancement strategy is adopted to guide the stimulus intensity toward the recommended intensity range predicted by the model, wherein the incremental amplitude of the stimulus intensity is set to be no less than a second lower limit value, and the second lower limit value is higher than the first upper limit value.

[0145] The phased treatment path is dynamically adjusted according to the frequency of the discomfort effect signals triggered in real time; when the frequency of the discomfort effect signals triggered in the second phase exceeds the pre-calculated safety threshold, the second phase is interrupted and the process returns to the first phase.

[0146] Specifically, its implementation begins in the planning phase of a new treatment cycle. When a user initiates a new treatment, the system first invokes the aforementioned long-term personalized model. The model not only predicts the initial parameters for this treatment but also, based on the user's historical response patterns, risk records, and current status, outputs two key planning criteria: first, the recommended intensity range for this treatment (a target interval); and second, the risk intensity boundaries that require attention. Based on this, the system dynamically generates a phased treatment path, which is clearly divided into at least two phases that must be executed sequentially, with different strategy objectives and operational rules set for each phase.

[0147] The goal of the first phase is to safely confirm the current personal comfort threshold. In this phase, the system employs a conservative enhancement strategy. Specifically, any tentative enhancement commands that may be generated during this phase have their intensity increment strictly limited, confined to a first upper limit. This first upper limit is a small constant (e.g., each enhancement does not exceed 1% of the base intensity), its purpose being to ensure that the intensity increase in this phase is extremely slow and cautious. The core task is to smoothly and reliably find and confirm the user's current real-time comfort threshold, avoiding triggering discomfort directly due to overly aggressive initial enhancements.

[0148] The second phase aims to guide the stimulus intensity towards the recommended range predicted by the model, after safely confirming the individual's comfort threshold. At this point, the system switches to an aggressive enhancement strategy. Accordingly, the incremental stimulus intensity in this phase is set to be no less than a second lower limit, which is set higher than a first upper limit. This means that in the second phase, the system is allowed to enhance the stimulus intensity in larger, more aggressive steps, aiming to more efficiently increase the stimulus intensity to a recommended range that the model deems potentially more effective for the user.

[0149] Crucially, this pre-defined path is not static. During the execution and monitoring of the path, the system continuously tracks the frequency of real-time triggered discomfort signals. This frequency data is compared in real-time with a safety threshold pre-calculated by the model based on historical user data. This safety threshold may be the maximum number of discomfort signals allowed per minute. The core logic of dynamic path adjustment is as follows: when the system detects that the frequency of triggered discomfort signals exceeds this pre-calculated safety threshold in the second phase, the system determines that the currently adopted aggressive enhancement strategy is too risky for the user's current state. In response, the system immediately interrupts the execution of the second phase and reverts the control flow to the first phase, re-activating the conservative enhancement strategy until the user's state stabilizes. This achieves a flexible control mechanism of advancing two steps and retreating one, or advancing one step and probing one.

[0150] This implementation method represents a leap from parameter-level optimization to strategy-level optimization, enhancing the safety, effectiveness, and user experience of the treatment process in complex individual cases. Its technical principle lies in its adoption of the phased treatment concept in clinical rehabilitation. By setting stages with different risk tolerance levels (first confirming safety, then approaching efficiency), the system proactively manages risk exposure during the treatment process. Using long-term models to calculate risk intensity boundaries and frequency safety thresholds makes this path planning highly personalized, rather than a fixed template. Its dynamic fallback mechanism establishes a robust real-time safety net, ensuring that even during the active enhancement phase, if a risk signal exceeding historical experience appears, the system can immediately brake and return to a safe mode. This creates a treatment process that is both proactive (actively approaching the recommended range) and resilient (retreating upon encountering risks). It effectively resolves the contradiction between plateau breakthroughs and managing state fluctuations that may occur in long-term treatment, enabling the device to dynamically adjust the treatment pace and intensity based on real-time user feedback, much like an experienced therapist—a high-level form of personalized, adaptive health intervention.

[0151] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. The above descriptions are only preferred embodiments of this application. It should be noted that due to the limitations of written expression, while there are objectively infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this invention, and can also combine the above technical features in an appropriate manner. These improvements, modifications, changes, or combinations, or the direct application of the inventive concept and technical solution to other situations without modification, should all be considered within the scope of protection of this application.

Claims

1. A system for implementing a parameter optimization method for a shuluo instrument based on the fusion of subjective and objective data, characterized in that, Includes the following steps: Based on the initial electrophysiological signals detected by the Shuluo instrument at the target acupoint, initial electrical stimulation parameters are generated, and the output is started based on the initial electrical stimulation parameters; In the initial stage of starting the output, the system receives a subjective trigger signal from the user that represents the tolerable threshold, and records the stimulus intensity value corresponding to the subjective trigger signal as the user's current personal comfort threshold. Continuous stimulation is applied based on the current personal comfort threshold, while simultaneously performing the following parallel steps: Real-time subjective comfort feedback signals, whether continuous or discrete, are acquired from user input via the interactive interface. Real-time monitoring of dynamic electrophysiological signals at target acupoints and analysis of their changing trends to generate objective regulatory suggestion signals; The real-time subjective comfort feedback signal and the objective adjustment suggestion signal are fused and analyzed, and the current electrical stimulation parameters are dynamically adjusted in real time accordingly. as well as, The process of fusing and analyzing the real-time subjective comfort feedback signal and the objective adjustment suggestion signal, and then dynamically adjusting the current electrical stimulation parameters in real time, includes the following steps: When the trend of the dynamic electrophysiological signal meets the first preset condition in terms of duration and amplitude, the trigger validity threshold is determined and an objective effective signal is generated. When the amplitude and speed of the real-time subjective comfort feedback signal exceeding the preset comfort range meet the second preset condition, it is determined that the discomfort trigger threshold is triggered and an discomfort activation signal is generated. Based on the objective effectiveness signal or the discomfort effectiveness signal, the current electrical stimulation parameters are dynamically adjusted in real time. The method further includes the following steps: If the following conditions are met simultaneously during a complete treatment cycle, the current personal comfort threshold will be updated at the end of the current treatment cycle: No new adverse effect signal was triggered. By performing adjustments based on the objective effectiveness signal, the stimulation intensity is successfully increased and stabilized at a level higher than the current personal comfort threshold recorded at the beginning of this conditioning cycle, and this stable state continues for a preset duration. The step of updating the current personal comfort threshold at the end of the current treatment cycle includes the following steps: From the time-series data of stimulation intensity in this conditioning cycle, extract the continuous data segment from the moment when the stimulation intensity first reaches and remains above the current personal comfort threshold recorded in the initial record until the end of conditioning or the end of the stable state, as the effective reinforcement data segment; Based on the real-time subjective comfort feedback signal and dynamic electrophysiological signal corresponding to the effective enhanced data segment, the dynamic credibility weight of each sampling point in the effective enhanced data segment is calculated. The stimulus intensity values ​​of all sampling points within the effective enhanced data segment are weighted and averaged according to their corresponding dynamic confidence weights, and the calculation result is used as the updated current personal comfort threshold.

2. The system according to claim 1, characterized in that: The real-time dynamic adjustment of the current electrical stimulation parameters based on the objective effectiveness signal or the discomfort effectiveness signal includes the following steps: If the discomfort activation signal is triggered, a protective attenuation command is immediately generated to reduce the current stimulation intensity to no higher than the current personal comfort threshold. If the objective activation signal is triggered and the real-time subjective comfort feedback signal stabilizes within the preset comfort range, a tentative enhancement instruction is generated; the magnitude of the stimulus intensity increment controlled by the tentative enhancement instruction is dynamically calculated according to preset rules. The preset rule satisfies the following: the magnitude of the stimulus intensity increment is positively correlated with the rate of change of the dynamic electrophysiological signal, and negatively correlated with the difference between the current stimulus intensity value and the stimulus intensity value before the most recent reduction in intensity due to triggering the discomfort effect signal in the historical record.

3. The system according to claim 1, characterized in that: The calculation of the dynamic confidence weight of each sampling point within the effective enhanced data segment includes the following steps: For any sampling point within the effective enhanced data segment, its dynamic credibility weight is calculated based on the subjective and objective data at that point, where: The absolute value of the deviation of the real-time subjective comfort feedback signal value at the sampling point from the center value of the preset comfort range is obtained, and the absolute value is input into the first mapping function to obtain the subjective factor value. The first mapping function is configured such that the output value decreases monotonically as the input value increases. The positive change amplitude of the dynamic electrophysiological signal value at the sampling point compared with the resting baseline value before conditioning is obtained, and the change amplitude is input into the second mapping function to obtain the objective factor value. The second mapping function is configured such that the output value increases monotonically with the increase of the input value. The subjective factor value is multiplied by the objective factor value and then normalized to obtain the dynamic credible weight.

4. The system according to claim 1, characterized in that: The calculation of the dynamic confidence weight of each sampling point within the effective enhanced data segment includes the following steps: For each sampling point, the instantaneous subjective confidence of the corresponding real-time subjective comfort feedback signal and the instantaneous objective confidence of the dynamic electrophysiological signal are calculated respectively. Based on the numerical relationship between the instantaneous subjective confidence and the instantaneous objective confidence, the dynamic confidence weight of each sampling point is determined; The instantaneous subjective confidence is positively correlated with the stability of the user's operation of the interactive interface and / or the smoothness of the real-time subjective comfort feedback signal over time; the instantaneous objective confidence is positively correlated with the signal-to-noise ratio of the dynamic electrophysiological signal and / or the continuity of its changing trend within a short time window.

5. The system according to claim 4, characterized in that: The step of determining the dynamic confidence weight of each sampling point based on the numerical relationship between the instantaneous subjective confidence and the instantaneous objective confidence includes the following steps: If both the instantaneous subjective confidence and the instantaneous objective confidence are higher than their respective preset high confidence thresholds, then the sampling point is assigned the highest basic weight value. If one of them is lower than its preset low confidence threshold, the basic weight is determined mainly based on the high confidence signal, and a compensatory attenuation is applied to it based on the historical confidence pattern. If both the instantaneous subjective confidence and the instantaneous objective confidence are within the medium confidence range, the basic weight is determined based on the weighted sum of their confidence levels, and the gain is adjusted in conjunction with the consistency of their trends. Gain is obtained when the trends are consistent, and attenuation is achieved when the trends diverge.

6. The system according to claim 2, characterized in that: The method also includes building and updating a long-term user personalization model, including the following steps: Record key data for each complete treatment cycle to form a personal historical dataset; the key data includes: the current personal comfort threshold at the beginning of the current treatment cycle, whether the threshold is updated at the end of the cycle and the updated current personal comfort threshold, the total number of discomfort signals triggered during the cycle, and the average dynamic confidence weight of the effective reinforcement data segment. Based on the personal historical dataset, a predictive relationship between the current personal comfort threshold and the user's basic physiological information and historical conditioning effects is established using machine learning models or statistical regression models. The long-term personalized model is used to optimize at least one of the following: Predict and recommend initial electrical stimulation parameters before the start of subsequent treatment cycles; The determination threshold in the first preset condition and / or the second preset condition is dynamically adjusted.

7. The system according to claim 6, characterized in that: The method also includes the following steps: Based on the aforementioned long-term personalized model, a phased treatment path is planned when a user starts a new treatment cycle. The phased treatment path includes at least a first phase and a second phase executed sequentially. In the first stage, a conservative enhancement strategy is adopted to safely confirm the current personal comfort threshold, wherein the magnitude of the stimulus intensity increment controlled by the tentative enhancement instruction is limited to a first upper limit value. In the second stage, an active enhancement strategy is adopted to guide the stimulus intensity toward the recommended intensity range predicted by the model, wherein the incremental amplitude of the stimulus intensity is set to be no less than a second lower limit value, and the second lower limit value is higher than the first upper limit value. The phased treatment path is dynamically adjusted according to the frequency of the discomfort effect signals triggered in real time; when the frequency of the discomfort effect signals triggered in the second phase exceeds the pre-calculated safety threshold, the second phase is interrupted and the process returns to the first phase.

Images