Intelligent insole and control method

Abstract

The invention relates to the technical field of intelligent control, and discloses an intelligent insole and a control method. The method comprises the steps that temperature data, humidity data and contact resistance data of a half sole area and an arch area in the intelligent insole are collected, and a foot environment parameter matrix is formed; performing micropore activity calculation and distribution analysis on the foot environment parameter matrix to obtain current distribution output parameters; calculating a humidity gradient ratio of a half sole area to an arch area according to the current distribution output parameters, and implementing a coordinated direct current control strategy to obtain a real-time current control parameter matrix; and inputting the real-time current control parameter matrix into an adaptive fuzzy control model for gradient descent optimization processing to obtain an optimal current intensity parameter and an optimal pulse frequency parameter. The foot environment is regulated and controlled according to needs, and meanwhile the stability of the humidity of the contact surface is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent control, and in particular, to an intelligent insole and a control method therefor. Background Art

[0002] Traditional insoles only provide passive physical support functions and are difficult to meet the needs of modern people for active regulation of the foot environment. Currently, the intelligent insole products on the market mainly focus on data collection and simple analysis functions, lacking the ability to sense and dynamically intervene in the foot microenvironment in real time. Especially in high-temperature and high-humidity environments, the problem of abnormal contact surface humidity caused by excessive secretion of liquid by micropores has not been effectively solved. Existing technologies usually adopt single-parameter monitoring and fixed-power output strategies, unable to perform differential control according to the environmental characteristics of different regions, with low energy utilization efficiency and limited control accuracy.

[0003] Although current-intervention type intelligent insoles have emerged, there are technical bottlenecks such as uneven current distribution, excessive power consumption, and short usage time. Especially in micro-current regulation, existing solutions generally adopt fixed-parameter control, lacking an adaptive optimization mechanism, resulting in difficulty in precise intervention according to the differences in micropore activity and contact resistance in different regions of the foot in practical applications. In addition, the current control system of traditional insoles lacks the ability to sense changes in walking states and cannot make real-time adjustments according to environmental changes during movement, causing energy waste and affecting the user experience. Summary of the Invention

[0004] The present invention provides an intelligent insole and a control method therefor, which realizes on-demand regulation of the foot environment and at the same time ensures the stability of the contact surface humidity.

[0005] In a first aspect, the present invention provides a control method for an intelligent insole, and the control method for the intelligent insole includes:

[0006] Collect temperature data, humidity data, and contact resistance data in the forefoot area and the arch area of the intelligent insole to form a foot environment parameter matrix;

[0007] Perform micropore activity calculation and distribution analysis on the foot environment parameter matrix to obtain current distribution output parameters;

[0008] Calculate the humidity gradient ratio between the forefoot area and the arch area according to the current distribution output parameters and implement a coordinated direct current control strategy to obtain a real-time current control parameter matrix;

[0009] Input the real-time current control parameter matrix into an adaptive fuzzy control model for gradient descent optimization processing to obtain optimal current intensity parameters and optimal pulse frequency parameters.

[0010] In a second aspect, the present invention provides an intelligent insole, which comprises:

[0011] a collection module, configured to collect temperature data, humidity data, and contact resistance data in the forefoot area and the arch area of the intelligent insole, and form a foot environment parameter matrix;

[0012] a distribution analysis module, configured to perform micropore activity calculation and distribution analysis on the foot environment parameter matrix to obtain current distribution output parameters;

[0013] a coordination module, configured to calculate the humidity gradient ratio between the forefoot area and the arch area according to the current distribution output parameters and implement a coordinated direct current control strategy to obtain a real-time current control parameter matrix;

[0014] an optimization module, configured to input the real-time current control parameter matrix into an adaptive fuzzy control model for gradient descent optimization processing to obtain an optimal current intensity parameter and an optimal pulse frequency parameter.

[0015] In the technical solution provided by the present invention, a modular multi-electrode array structure layout is adopted, 5 humidity sensors and 3 temperature sensors are set in the forefoot area, 3 humidity sensors and 2 temperature sensors are set in the arch area, and the contact resistance is measured by the four-point probe method, so as to realize high-density and multi-dimensional monitoring of foot environmental parameters, and greatly improve the spatial resolution and parameter acquisition accuracy of environmental feature identification. Based on micropore density analysis and activity distribution vector diagram technology, the present invention can accurately identify the micropore activity difference between the forefoot area and the arch area, calculate the current distribution parameters in a targeted manner, realize precise regulation of different areas, and effectively solve the problem of uneven current distribution in traditional technologies. By constructing a five-layer fuzzy reasoning network structure and a fuzzy rule base of 25 IF-THEN rules, the present invention realizes adaptive optimization of current intensity and pulse frequency. The system can automatically adjust the control parameters according to environmental changes, which greatly improves the accuracy and intelligence level of environmental regulation. The present invention designs a linear extended state observer and an anti-interference control law, which can effectively suppress random interference during walking, ensure control stability in a dynamic environment, and significantly improve the adaptability of the system in actual application scenarios. By adopting an intermittent pulse power limiting model with an initial working cycle ratio of 30 seconds working / 120 seconds pause, combined with an adaptive adjustment mechanism for the remaining battery capacity, the present invention achieves the long-term working ability of the smart insole, extends the single-charge usage time compared to traditional technologies, and improves the practicality of the device. By replacing the traditional square wave with a trapezoidal wave with a rise time and a fall time of 50ms, the present invention reduces the impact of current pulses on the electrodes and contact surfaces, improves the comfort and safety of current regulation, and at the same time reduces the aging rate of electrode materials, extending the service life of the equipment. By real-time acquisition and analysis of the change rate of contact surface humidity, the electrode array is intelligently triggered to power on when the regional humidity exceeds the threshold, thereby achieving on-demand regulation of the foot environment, avoiding unnecessary energy consumption, and ensuring the stability of the contact surface humidity. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0017] Figure 1 FIG. 1 is a schematic diagram of an embodiment of a control method of a smart insole in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] An embodiment of the present invention provides an intelligent insole and a control method. Terms such as "first", "second", "third", "fourth", etc. (if any) in the specification, claims and the above-mentioned drawings of the present invention are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the term "comprising" or "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or are inherent to these processes, methods, products or devices.

[0019] For ease of understanding, the specific process of the embodiment of the present invention will be described below. Please refer to Figure 1 An embodiment of the control method of the intelligent insole in the embodiment of the present invention includes:

[0020] Step S101, collect temperature data, humidity data and contact resistance data in the forefoot area and the arch area of the intelligent insole to form a foot environment parameter matrix;

[0021] It can be understood that the execution subject of the present invention can be an intelligent insole, or a terminal or a server, and specific details are not limited here. The embodiment of the present invention will be described by taking the server as the execution subject as an example.

[0022] Specifically, E humidity sensors and F temperature sensors are set in the forefoot area of the smart insole, and at the same time, M humidity sensors and N temperature sensors are configured in the arch area, so as to construct a temperature and humidity sensor array covering the main sweat gland distribution areas of the forefoot and arch. The sensor array uses capacitive humidity sensors and NTC thermistor temperature sensors as core components, and has high sensitivity and fast response characteristics. The entire sensor array is driven by a low-power microcontroller to continuously work at a preset sampling frequency, and the temperature and humidity change data of the foot contact surface are obtained in real time. The humidity sampling frequency is set to 1 Hz, and the temperature sampling frequency is set to 0.5 Hz. After converting the analog signal output by the sensor into a digital signal, it is filtered by a digital low-pass filter to filter out high-frequency interference and sensor noise, and a two-dimensional humidity data matrix and a temperature data matrix are obtained respectively. The rows and columns of the matrix correspond to different sampling times and sensor position indexes respectively. On the premise of forming a stable electrical contact between the foot skin and the electrodes, the four-point probe method is used to perform contact resistance measurement operations on the forefoot area and the arch area respectively. Among them, a constant current source is used to excite the two outer electrodes and measure the voltage drop between the two inner electrodes, so as to eliminate the influence of the contact electrode resistance on the measurement result, and the skin resistance value is calculated through Ohm's law to construct a contact resistance data matrix. Each element in this matrix represents the skin resistance value at a specific position at a certain moment, reflecting the local characteristics of the change in sweat conductivity and skin conductivity. In order to unify the data expression ranges of various physical quantities, enhance the convergence performance of subsequent algorithms, and facilitate matrix fusion processing, standardization processing is performed on the humidity data matrix, temperature data matrix, and contact resistance data matrix respectively. The original values in each matrix are linearly mapped and converted according to a preset reference interval, so that all values are normalized to the [0,1] interval or other unified dimension systems. The specific method adopts the Z-score standardization or Min-Max normalization strategy to ensure consistency and comparability in the dimensions of time, position, and physical units. Finally, the humidity data matrix, temperature data matrix, and contact resistance data matrix are three-dimensionally fused and spliced to construct a foot environmental parameter matrix.

[0023] Step S102: Calculate the micropore activity and distribution analysis of the foot environmental parameter matrix to obtain the current distribution output parameter;

[0024] Specifically, based on the contact resistance data and the foot contact surface temperature data contained in the foot environmental parameter matrix, a micro-pore activity index model reflecting the local sweat gland function state is constructed. This model jointly evaluates the resistance value and temperature value of each sensor monitoring point to judge the conductivity and metabolic heat state of the local skin area, thereby indirectly inferring the activity level of the sweat glands in this area, that is, the strength of the micro-pore opening and closing activity. In this process, an increase in temperature and a decrease in resistance are regarded as signs of enhanced sweat gland activity, so the system assigns a higher activity score to such a state. Based on this series of evaluation results, the micro-pore activity indexes of the forefoot area and the arch area are integrated by region, and considering the differences in their anatomical structures, especially the anatomical feature that the sweat gland density in the forefoot area is much higher than that in the arch area, different weights are imposed on the activity index, so that the influence between regions has a distinction in the overall map. The micro-pore activity distribution vector map generated after weighted integration describes the spatial distribution structure of the micro-sweat gland activity at different positions on the intelligent insole. According to this activity distribution map and the real-time contact resistance data, the initial current output parameters for the forefoot side and the arch side are respectively derived. Since the current is related to the skin conduction ability, the smaller the contact resistance, the more sweat and better conductivity in this area, so the current output intensity of such areas is appropriately increased to improve the local penetration effect of iontophoresis. These preliminary current parameters form the initial current framework for the forefoot and arch areas, and on this basis, further coordination processing at the power level is implemented. To ensure the battery usage efficiency and treatment balance, the initial current parameters are substituted into the dynamic power distribution model to establish a power balance relationship between the forefoot side and the arch side, ensuring that the total electrical input of the two regions is adaptively adjusted within the established power consumption limit range. At the same time, the change trend of the foot contact surface humidity on the time axis is extracted from the environmental parameter matrix to judge the dynamic process of sweat secretion. When it is detected that the humidity rises rapidly, it indicates that the sweat gland activity has entered the peak period. At this time, the proportionality coefficient involved in the power distribution relationship will be adjusted to quickly respond and inhibit excessive sweating. After obtaining the new current distribution coefficient and combining it with the previously generated micro-pore activity distribution vector map, the current output value of each electrode unit on the intelligent insole is calculated one by one, and the current is distributed according to the local activity level, conductivity state, and regional priority level, so that the current distribution is more in line with the physiological reality of the sole. Finally, a set of current distribution output parameters is formed as the basic instruction for the system to implement precise iontophoresis stimulation subsequently.

[0025] In this embodiment, a structural region division is performed on the micro-pore activity distribution vector diagram. Based on the physical distribution pattern of the electrodes on the smart insole, all electrode units are divided into two independent region sets: the forefoot control region and the arch control region, and it is ensured that the electrodes in each region correspond to specific activity sub-vectors and contact resistance subsets. Based on the electrode unit region division results, a region matching calculation is performed on the contact resistance data. Taking each electrode region as a boundary, the contact resistance values are aggregated and integrated to obtain the average contact resistance value of the forefoot region and the average contact resistance value of the arch region respectively. According to the average contact resistance value of the forefoot region and the average contact resistance value of the arch region, an appropriate current control model is established for the forefoot region and the arch region respectively. This model takes the regional average impedance as one of the input variables, and defines the current model calculation path through the voltage-current response relationship obtained by simulation or experimental calibration, so as to quickly obtain the electrical response behavior of the macroscopic region without directly solving the complex distributed conduction system. To improve the local adaptability of the current output, the micro-pore activity distribution vector diagram is called again and used as an adjustment factor to dynamically control the voltage parameters applied to the forefoot side and the arch side, that is, a slightly higher driving voltage is applied in the region with a higher micro-pore activity index to enhance the response ability of the current density in this region, and vice versa, to ensure that the current distribution of the entire platform can not only reflect the differential stimulation requirements but also control the overall power consumption level. After the voltage parameters are adjusted, the voltage values of the forefoot side and the arch side are respectively substituted into the current models established for them, and combined with the regional average contact resistance value for calculation to obtain the initial current values of these two regions. These current values represent the initial stimulation intensity settings of the system for the forefoot and the arch at this moment. To refine this initial current to the electrode unit level, a distribution calculation based on spatial distribution is performed on the above results. Considering the activity scores corresponding to the electrode positions and the resistance differences, a corresponding proportion of the current output value is allocated to each electrode unit, so that it not only meets the quota requirements of the total regional current but also reflects the differential expression of local stimulation requirements. The initial current parameters of the forefoot side and the arch side are obtained.

[0026] Step S103: Calculate the humidity gradient ratio of the forefoot region and the arch region according to the current distribution output parameters and implement a coordinated direct current control strategy to obtain a real-time current control parameter matrix;

[0027] Specifically, continuously obtain the real-time contact surface humidity data of the forefoot and arch regions from the sensor array, and calculate the humidity change rate of each region at fixed time intervals, denoted as the second humidity change rate. This change rate reflects the dynamic trend of local sweat gland secretion in the time dimension. Calculate the ratio of the humidity change rate of the forefoot region to the humidity change rate of the arch region to obtain the humidity gradient ratio between the two. This ratio is used to characterize the relative difference in sweating activity of different regions of the current sole. If the humidity change in the forefoot region is more intense, it indicates that the activity of the sweat glands in this region is increasing, and the current output intensity should be preferentially enhanced; conversely, if the arch region shows a greater humidity change, it needs to be strengthened. Based on the above humidity gradient ratio, generate a set of regional coordination control coefficients. This control coefficient reflects the priority adjustment direction of the current intensity of each region in the current state, and accordingly dynamically corrects the established reference current value in the current distribution output parameters, so that the originally statically set current output can be adaptively adjusted under real-time conditions, forming a control strategy that better fits the physiological fluctuations of the foot. Through the adjusted current value, generate the real-time current intensity data for the forefoot side and the arch side respectively. To make the output current maintain a smooth transition with the change of the environment and avoid discomfort caused by sudden changes, perform differential adjustment on the above real-time current value, that is, by measuring the change amplitude and direction of the current value between adjacent time points, and combining the speed level of the second humidity change rate, set appropriate current change rates for the forefoot side and the arch side respectively to obtain dynamic current adjustment parameters, which determine the increment or decrement amplitude of the current that can be adjusted per unit time, making the control process more flexible and safe. On this basis, introduce the real-time micropore activity index and the contact surface temperature data as correction factors to perform balance and temperature compensation processing on the current distribution result. Among them, the current balance strategy is used to balance the uneven current output caused by the activity difference of multiple electrode units in the same region, ensuring uniform distribution of stimulation; while the temperature compensation mechanism is used to prevent skin discomfort or burns caused by local overheating. When the real-time temperature exceeds the set threshold, the system automatically reduces the current intensity to ensure the safety of the user. Integrate various adjustment parameters, activity indicators, and temperature states to generate a real-time current control parameter matrix.

[0028] Step S104: Input the real-time current control parameter matrix into the adaptive fuzzy control model for gradient descent optimization processing to obtain the optimal current intensity parameter and the optimal pulse frequency parameter.

[0029] Specifically, the real-time current control parameter matrix is used as the basic input of the current plantar electrostimulation regulation state and introduced into the built-in adaptive fuzzy control model for processing. This model is constructed using a five-layer fuzzy inference network structure, and its overall architecture consists of an input layer, three hidden layers, and an output layer connected in series in sequence. It also integrates multiple groups of fuzzy rules, activation functions, and learning control units internally to achieve the non-linear mapping relationship between current parameters and external physiological states. The inputs of this model include the current intensity and distribution of each electrode unit in the real-time current control matrix, as well as auxiliary information such as the humidity change trend, skin resistance level, and current battery state. After weighted conversion by the input layer, all input variables flow into the hidden layer structure respectively for distributed fuzzy mapping and feature compression. In the forward propagation stage, through set activation functions, such as ReLU, Tanh, and Sigmoid, the signal calculation is completed layer by layer, gradually forming a non-linear combination expression of the hidden state, and initially predicting the current intensity and pulse frequency. Since this model needs to make an optimal response to the dynamic environment, its internal network parameters are continuously updated through continuous optimization means. For this purpose, the root mean square propagation algorithm is introduced as the core gradient descent optimization strategy. By calculating the mean square error difference between the model output and the expected control target, a loss function is generated, and this loss function is used to gradually iteratively update the weight matrix and bias term in the model. This algorithm has the characteristic of an adaptive learning rate, which can effectively suppress gradient oscillation and improve convergence stability. At the same time, the mean square value of historical gradients is retained in each update, so that the parameter update direction has a stronger overall trend judgment ability. As the multi-round optimization iteration progresses, the system feeds the new network weights back to the forward inference path after each round of parameter update, and corrects the error by inverting the current intensity and pulse frequency of the model output to continuously approach the global optimal solution. At the same time, to prevent the problem of parameter divergence or over-adjustment in the training process of the model, an adaptive step size control mechanism is implemented. When the model error convergence speed slows down after several consecutive iterations, the system automatically reduces the learning rate step size to ensure the model stability. And a range limit processing mechanism is introduced to set upper and lower threshold values for the output current intensity and pulse frequency to ensure that they operate within the physically implementable range and avoid the risk of skin irritation caused by excessive current or too dense pulse frequency. Based on the updated network parameter values obtained throughout the above process, the optimal current intensity parameter and optimal pulse frequency parameter for the current treatment cycle are output through the adaptive fuzzy control model.

[0030] In this embodiment, according to the optimal current intensity parameter, combined with the contact resistance data corresponding to each electrode unit, an intermittent pulse power limit model is constructed. This model realizes the effective control of the average power consumption by limiting the ratio of the effective working duration to the rest time of the electrode unit within a unit time. In the model initialization stage, a basic working cycle structure is set, such as working for 30 seconds and pausing for 120 seconds. Based on this basic cycle parameter, combined with the optimal current intensity, the actual power output values of each electrode in the current cycle are calculated. These power values are used to evaluate the overall power consumption and construct a battery load prediction model. To enable the current output mode to adapt to the dynamic posture changes of the user during walking, gait feature data of the user, including step frequency, acceleration change rate, and exercise intensity index, are collected at a high frequency through an acceleration sensor, and combined with preset thresholds to determine whether the current state belongs to one of the four different levels: stationary, mild activity, moderate activity, or high activity. According to the recognition result, the actual power values of each electrode unit are dynamically adjusted, and the original fixed working cycle ratio is replaced with dynamic working cycle control parameters. For example, during moderate activity, the working duration is shortened and the rest time is extended, thereby effectively reducing the physiological interference and discomfort during the stimulation process and improving the wearing experience. On this basis, a battery power awareness mechanism is introduced, that is, the remaining capacity level of the battery is detected in real time, and the generated dynamic working cycle control parameters are adaptively adjusted according to the current battery power state. When it is detected that the battery power drops below a certain threshold, the pulse output rhythm is compressed by shortening the working duration, reducing the frequency, or extending the rest period, etc., to obtain a more energy-saving target working cycle control value. At the same time, to cope with external disturbances caused by mechanical perturbations, posture changes, or pressure distribution transfer of the feet during walking, a linear extended state observer is designed to estimate the manifestation form of the disturbance signal in the current output path in real time. This observer can dynamically extract the influence degree of the disturbance source on the current system in the form of state variables, and fuse this disturbance estimation value with the optimal current intensity parameter to construct a control law with anti-disturbance ability. When implementing the control input adjustment, this control law can actively offset the current output fluctuation caused by walking activities, thereby ensuring that the system still maintains stable output under high disturbance conditions. Based on the current control input value with anti-disturbance ability, pulse waveform optimization operation is performed on the optimal pulse frequency parameter. To reduce the skin tingling sensation and improve user comfort, a trapezoidal waveform is used instead of the traditional square wave structure. By setting appropriate rise and fall edge times, the current stimulation is made smoother and the edges are softer, and this waveform parameter and the adjusted pulse frequency are combined to form an electrical stimulation timing structure. The optimized current pulse waveform parameters and the target working cycle control value are fused, and combined with the local activity level, physiological resistance state, and distribution position of each electrode unit, corresponding current intensity, pulse frequency, working cycle, and waveform parameters are assigned to it to generate an electrode-level target current control sequence.

[0031] In an embodiment of the present invention, a modular multi-electrode array structure layout is adopted, 5 humidity sensors and 3 temperature sensors are set in the forefoot area, 3 humidity sensors and 2 temperature sensors are set in the arch area, and the contact resistance is measured by the four-point probe method, so as to realize high-density and multi-dimensional monitoring of foot environmental parameters, and greatly improve the spatial resolution and parameter acquisition accuracy of environmental feature identification. Based on micropore density analysis and activity distribution vector diagram technology, the present invention can accurately identify the micropore activity difference between the forefoot area and the arch area, calculate the current distribution parameters in a targeted manner, realize accurate regulation of different areas, and effectively solve the problem of uneven current distribution in traditional technologies. By constructing a five-layer fuzzy reasoning network structure and a fuzzy rule base of 25 IF-THEN rules, the present invention realizes adaptive optimization of current intensity and pulse frequency, and the system can automatically adjust the control parameters according to environmental changes, which greatly improves the accuracy and intelligence level of environmental regulation. The present invention designs a linear expansion state observer and an anti-interference control law, which can effectively suppress random interference during walking, ensure control stability in a dynamic environment, and significantly improve the adaptability of the system in actual application scenarios. By adopting an intermittent pulse power limiting model with an initial working cycle ratio of 30 seconds working / 120 seconds pause, combined with an adaptive adjustment mechanism for the remaining battery capacity, the present invention achieves the long-term working ability of the smart insole, extends the single-charge usage time compared to traditional technologies, and improves the practicality of the device. By replacing the traditional square wave with a trapezoidal wave with a rise time and a fall time of 50ms, the present invention reduces the impact of current pulses on the electrodes and contact surfaces, improves the comfort and safety of current regulation, and at the same time reduces the aging rate of electrode materials, extending the service life of the equipment. By real-time acquisition and analysis of the change rate of contact surface humidity, the electrode array is intelligently triggered to power on when the regional humidity exceeds the threshold, thereby achieving on-demand regulation of the foot environment, avoiding unnecessary energy consumption, and ensuring the stability of the contact surface humidity.

[0032] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0033] E humidity sensors and F temperature sensors are set in the forefoot area of ​​the smart insole, and M humidity sensors and N temperature sensors are set in the arch area to form a sensor array;

[0034] The humidity data of the foot contact surface is collected by the humidity sensor in the sensor array, and the humidity data of the foot contact surface is processed by digital low-pass filtering to obtain a humidity data matrix;

[0035] The temperature data of the foot contact surface are collected by the temperature sensors in the sensor array, and the temperature data of the foot contact surface are processed by digital low-pass filtering to obtain a temperature data matrix;

[0036] The four-point probe method is used to measure the contact resistance data in the forefoot area and the arch area, obtaining a contact resistance data matrix, and the humidity data matrix, the temperature data matrix, and the contact resistance data matrix are standardized to obtain a foot environmental parameter matrix.

[0037] Specifically, according to the sweat gland distribution, bioactivity differences, and physiological pressure characteristics in different regions of the sole, the forefoot and the arch are configured independently as key monitoring regions. Set E humidity sensors and F temperature sensors in the forefoot region, while set M humidity sensors and N temperature sensors in the arch region to form a spatially covering distributed sensor array. During the sensor installation process, it is necessary to ensure that each device has good flexible adhesion to the sole skin. At the same time, to avoid local structural bulges or pressure loss, a polymer flexible film encapsulation material is used to embed the sensors integrally under the insole surface, which not only ensures the measurement accuracy but also does not affect the wearing comfort. All types of sensors are connected to the microcontroller unit through a unified multi-channel data acquisition bus. Among them, the humidity sensor uses a high-sensitivity capacitive sensing structure, which can quickly respond to the increase in local environmental humidity caused by sweat secretion on the sole; the temperature sensor uses an NTC thermistor, which has the ability to continuously track the subtle changes in skin temperature. The acquisition module preliminarily quantifies all sensor signals relying on a 12-bit analog-to-digital converter and transmits the converted digital signals to the processing unit uniformly. To ensure the stability of subsequent analysis and filter out high-frequency noise caused by foot movement, momentary poor contact, or environmental jitter, the system performs digital low-pass filtering on the original signals of humidity and temperature after receiving them. This filter is implemented using a first-order recursive algorithm, and the cut-off frequency is set to 5 Hz, so that the low-frequency components in the signal are retained, while the high-frequency fluctuations are effectively smoothed. The filtered humidity data and temperature data are respectively processed into two-dimensional matrices according to the sensor number and position index to form a humidity data matrix and a temperature data matrix. Each row represents the values measured by each sensor at a certain moment, and each column represents the humidity or temperature evolution sequence of a specific position. At the same time, to obtain conductive data that can better reflect the relationship between skin conductance state and sweating behavior, contact resistance measurements are respectively performed on the forefoot and arch regions by the four-point probe method. This method applies a weak constant current through two outer electrodes and measures the voltage drop value between two inner electrodes at the same time to exclude the interference of electrode contact resistance on the overall impedance measurement, so as to more accurately reflect the true conduction ability of the skin tissue to the current. During the measurement, the system automatically controls each electrode to work in turn and realizes concurrent acquisition through a multi-channel switch to obtain the contact resistance matrices of the forefoot and arch regions. Considering that the three different physical quantities have different units, dimensions, and magnitude characteristics, direct calculation will cause the model to be biased towards a single index. Therefore, the humidity data matrix, temperature data matrix, and contact resistance data matrix are respectively standardized. During the standardization process, the Min-Max normalization method is used to map all values to the [0,1] interval. At the same time, for parameters with physiological limits such as temperature and resistance, upper and lower limit clipping processing is performed to prevent interference from extreme values on the stability of subsequent calculations. The three standardized data matrices are structurally fused according to the time, position, and index dimensions to form a three-dimensional structure of the foot environmental parameter matrix.

[0038] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0039] Calculate the micropore activity index based on the contact resistance data and the foot contact surface temperature data in the foot environment parameter matrix to obtain the micropore activity indexes of the forefoot area and the arch area;

[0040] Conduct regional weight analysis on the micropore activity index according to the structural characteristics of the micropore density in the forefoot area and the arch area to obtain a micropore activity distribution vector diagram;

[0041] Calculate the initial current parameters on the forefoot side and the arch side based on the micropore activity distribution vector diagram and the contact resistance data;

[0042] Conduct dynamic power distribution based on the initial current parameters on the forefoot side and the arch side to obtain a power balance equation for the forefoot side and the arch side;

[0043] Calculate the first humidity change rate according to the foot contact surface humidity data in the foot environment parameter matrix, and adjust the coefficients of the power balance equation based on the first humidity change rate to obtain a current distribution coefficient;

[0044] Calculate the current output of each electrode unit according to the current distribution coefficient and the micropore activity distribution vector diagram to obtain current distribution output parameters.

[0045] Specifically, contact resistance data and contact surface temperature data are extracted from the foot environmental parameter matrix and used as physiological state input variables to construct a micro-pore activity index calculation model. In this model, local skin conductance characteristics are reflected by contact resistance, and the intensity of sweat gland metabolic activity is represented by temperature. The two are combined to establish a multi-point feature response evaluation system, and then a point-by-point activity score is given to each electrode position in the two regions of the forefoot and the arch. A lower resistance indicates sufficient electrolyte coverage and unobstructed ion channels, while an increase in temperature means capillary dilation and enhanced metabolic activity. After these signals are quantitatively combined, they can be mapped to a numerical index of the micro-pore activity level in this region, forming a set of micro-pore activity indexes for the forefoot region and the arch region. According to the set of micro-pore activity indexes for the forefoot region and the arch region, combined with the plantar anatomical characteristics, regional structure modeling and analysis are carried out. Considering that the sweat glands are densely distributed in the forefoot region and the number of micro-pores per unit area is much higher than that in the arch region, a structure density weight is introduced in the process of processing the activity index, and the original activity indexes of different regions are amplified or compressed according to the regional characteristics to form a micro-pore activity distribution vector map with structure sensitivity. This vector map retains the response ability of each electrode unit in the original physiological state and structurally compensates for the physiological differences between regions through weighting. The micro-pore activity distribution vector map is combined with the standardized contact resistance data to estimate the initial current parameters. Fitting calculations are carried out based on the relationship between the activity score of each electrode unit and the skin conductivity at that position. The system calculates an initial total current value for each region, that is, the initial current parameters for the forefoot side and the initial current parameters for the arch side. These two parameters are the core inputs for subsequent power control. During the calculation process, factors such as electrode layout, regional area, and battery power constraints need to be considered to ensure that the current output meets physiological requirements without exceeding the device safety limit. To achieve reasonable allocation and dynamic control of power resources between regions, a dynamic power distribution model is constructed for the forefoot side and the arch side based on the above initial current parameters. The initial current value of the forefoot region is multiplied by its regional average resistance to obtain the theoretical power value of this region, and the same is done for the arch region, thereby constructing a power balance equation. In this balance equation, the system stipulates that the sum of the powers of the two regions is equal to the total system output power, and an adjustment coefficient is reserved for subsequent flexible adjustment to achieve dynamic power consumption control and load balancing. To enhance the system's response ability to changes in the user's sweating rhythm, a humidity data sequence is extracted from the foot environmental parameter matrix, and the first humidity change rate of the forefoot and arch regions is calculated at a fixed sampling time interval, that is, the speed value of humidity increase per unit time, as a key indicator for judging the changing trend of sweat gland activity. When the humidity change rate exceeds the set threshold, the system will judge it as a high sweating state and must increase the current stimulation density. When the humidity growth tends to slow down or even decrease, the system believes that the sweat gland activity is slowing down and should appropriately reduce the stimulation intensity to avoid causing skin interference.Based on the analysis results, the regional coefficients in the power balance equation are dynamically adjusted. By introducing a coefficient mapping function with the humidity change rate as the input, the current distribution ratio between the forefoot and arch regions in the current distribution strategy is reconstructed to form a current distribution coefficient. Based on the updated distribution coefficient and the activity distribution vector map, the current output calculation is performed. At this stage, the control algorithm performs a current value weighting operation for each electrode unit. According to its score value in the vector map and the corresponding current distribution ratio of the region where it is located, the overall current resource is split into each electrode as needed, forming a current output map refined to spatial points. This map is the final expression form of the current distribution output parameters, where each element represents the current intensity required for an electrode unit to output in the current control cycle. The system uses this distribution result as the actual control instruction for the current cycle and issues it to the drive module to achieve a refined electrostimulation control mechanism based on physiological perception, dynamic response, and structural adaptation.

[0046] In a specific embodiment, the process of performing the step of calculating the initial current parameters on the forefoot side and the arch side based on the micropore activity distribution vector map and the contact resistance data may specifically include the following steps:

[0047] Perform regional division on the micropore activity distribution vector map to obtain the regional division result of the electrode units;

[0048] Perform regional matching calculation on the contact resistance data based on the regional division result of the electrode units to obtain the average contact resistance value of the forefoot region and the average contact resistance value of the arch region;

[0049] According to the average contact resistance value of the forefoot region and the average contact resistance value of the arch region, establish current equation models for the forefoot side and the arch side for the forefoot region and the arch region respectively;

[0050] Perform dynamic adjustment calculation on the forefoot region voltage and the arch region voltage based on the micropore activity distribution vector map to obtain the voltage parameters of the forefoot side and the arch side;

[0051] Substitute the voltage parameters of the forefoot side and the arch side into the current equation model for calculation to obtain the initial current values of the forefoot side and the arch side;

[0052] Perform electrode unit allocation calculation on the initial current values of the forefoot side and the arch side to obtain the initial current parameters of the forefoot side and the arch side.

[0053] Specifically, taking the micro-pore activity distribution vector map as the spatial recognition basis, the electrode units on the intelligent insole are regionally divided. This vector map is composed of the micro-activity scores of multiple electrode nodes, representing the physiological activity distribution characteristics of sweat glands in different parts of the sole. According to the insole electrode layout structure, the system divides the vector map into two parts: the forefoot control area and the arch control area according to geometric positions. The forefoot area includes the electrode units from the toe root to the start of the arch, and the arch area includes the electrode units extending from the middle of the arch to the front of the heel. Each unit is numbered according to its physical position for subsequent data matching and model construction. After completing the electrode area division, the area division result is corresponded to the contact resistance subset in the environmental parameter matrix to ensure that the electrode units in each divided area can be paired with their corresponding skin resistance data. By aggregating the contact resistance data of all electrodes in each area and taking their weighted average, the average contact resistance values of the forefoot area and the arch area are obtained respectively. After obtaining the regional average resistance, independent current estimation models are established for the forefoot side and the arch side respectively. The input of the model is the voltage parameter and the contact resistance parameter, and the output of the model is the initial regional current value, which represents the actual current magnitude that may be generated in a unit area under the current conduction state. Due to the different contact characteristics and electrode coverage areas of the two areas, the current models maintain a unified form in the expression structure but need to be separately defined in parameter calibration. The forefoot side is biased towards high-intensity current output modeling, while the arch side is biased towards balanced control. Once again, using the maximum value, mean value, and gradient distribution of the scores in each area of the micro-pore activity distribution vector map, the driving voltages of the forefoot side and the arch side are dynamically adjusted. This voltage adjustment is differentially weighted according to the local sweat gland activity level. When the activity score of a certain area is significantly higher than the average level, the system appropriately increases the voltage of this area to enhance the iontophoresis effect, while in the low-activity area, the basic voltage is maintained or decreased to control the power load and improve the stimulation efficiency. Such voltage parameter adjustments are realized through digital potentiometers or programmable voltage sources at the circuit level and are dynamically generated by the microcontroller combined with the vector map score data at the logic control level. The voltage parameters of the above-mentioned forefoot side and arch side are respectively input into the previously constructed current model, substituting the corresponding average contact resistance values, and the initial current values of the two areas are obtained through the calculation logic embedded in the model. This calculation result represents the total current quota that each large area should obtain in the entire current control cycle under the current conductance state and stimulation requirements. The total initial current value at the regional level is refined into the output parameters of each electrode unit. The control algorithm takes the activity score of the electrode unit as the core index, combines the current quota of the area where the unit is located and its relative activity in the area, and calculates the current proportion it should obtain. That is, within each area, the electrode unit with a higher score will undertake more current output in this stimulation cycle, while the unit with a lower score will undertake less current output.The distributed current distribution strategy based on physiological differences and position hotspots enables the smart insole to achieve electroosmotic intervention with higher spatial resolution, allowing the treatment behavior to accurately target the active sweat gland areas that truly need stimulation, avoiding resource waste and over-intervention in non-target areas. Through the above steps, an initial current parameter set for the forefoot and arch regions is finally output. Each electrode unit has its independent current setting value, and the matrix composed of all values constitutes the basic current drive sequence executed by the system during this period.

[0054] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0055] Obtain the real-time contact surface humidity data and calculate the second humidity change rate, and calculate the humidity gradient ratio between the forefoot region and the arch region based on the second humidity change rate;

[0056] Generate a regional coordination control coefficient according to the humidity gradient ratio, and dynamically adjust the reference current value in the current distribution output parameters based on the regional coordination control coefficient to obtain the real-time current values on the forefoot side and the arch side;

[0057] Perform differential regulation on the real-time current values on the forefoot side and the arch side and set the current change rates on the forefoot side and the arch side according to the second humidity change rate to obtain dynamic current regulation parameters;

[0058] Perform current balancing and temperature compensation based on the dynamic current regulation parameters, the real-time micropore activity index, and the real-time contact surface temperature value to obtain a real-time current control parameter matrix.

[0059] Specifically, the contact surface humidity data of the forefoot area and the arch area are collected in real time from the sensor array. After being filtered and denoised, these data are stored in a buffer in the form of a continuous time series for rate-of-change analysis. Between two consecutive sampling periods, the current humidity value in the same area is compared with the humidity value at the previous time point to calculate the instantaneous rate of change of humidity, that is, the second humidity change rate. This index not only reflects the current secretion intensity of sweat glands but also can reveal in real time whether the sweat generation trend is accelerating or slowing down. After completing the rate-of-change calculation, the humidity growth rates of the forefoot area and the arch area are obtained respectively, and the ratio of the two is calculated to generate a humidity gradient ratio, which is used to reveal the degree of sweating difference between the two main areas of the current sole. If the humidity growth rate of the forefoot area is significantly higher than that of the arch area, it indicates that the sweat gland activity at the forefoot is more intense, and the current output density of this area is preferentially increased to enhance the intervention intensity. Therefore, a set of regional coordination control coefficients is generated according to this humidity gradient ratio. Based on the standard value, when the ratio is greater than 1, the current intensity is increased towards the forefoot area, and when the ratio is less than 1, the output is enhanced towards the arch area, thereby achieving the goal of dynamically responding to the sweat gland activity status of different areas. Taking the regional coordination control coefficient as an adjustment factor, the reference current value in the generated current distribution output parameters is dynamically updated, that is, instead of using a statically set current distribution structure, it is adjusted in real time based on the current environmental state. The adjusted current values are respectively assigned to the forefoot side and the arch side to form new real-time current values. These current values reflect the response intensity of the system to external sweat change signals, and at the same time ensure the continuity and gradualness of the front and rear control behaviors, avoiding the discomfort caused by sudden changes or control oscillations. Since the current output is a continuous time process, to ensure a smooth transition of the current output of each electrode unit at the next moment, differential adjustment processing is performed on the real-time current values of the forefoot and the arch, that is, the rate of change between the current current and the current in the previous control period is calculated, and the upper limit of the current change rate is set according to the known second humidity change rate. When a significant humidity change is detected, the system allows the current value to increase rapidly, but still needs to be controlled within the set safe change range; while when the humidity is stable or decreasing, the system controls the current value to gradually decrease to avoid over-stimulation, thereby forming dynamic current adjustment parameters. The dynamic adjustment parameters are fused with the micro-pore activity index and the contact surface temperature data obtained in real time to achieve a higher-precision current balance control and temperature compensation mechanism. In terms of current balance, by comparing the deviation degree between the micro-pore activity of each electrode unit and the regional average value, it is judged whether the unit is in a high-activity or low-activity area, and local current fine-tuning is performed accordingly, so that the units with high activity bear more current output in the current cycle, thereby improving the spatial targeting of current distribution.Meanwhile, considering that local temperature rise is likely to occur in the active area due to current output, continuously monitor the skin surface temperature obtained in real time. When it is found that the contact surface temperature corresponding to a certain electrode unit exceeds the set safety threshold, reduce the current output at this position, and reduce its current intensity through a predefined linear or exponential temperature compensation function to avoid skin burning or damage caused by excessive temperature. After the above-mentioned data fusion and parameter dynamic adjustment, a real-time current control parameter matrix is finally generated. This matrix contains the current intensity values that each electrode unit needs to output in the current control cycle, and simultaneously encodes the adjustment rate identification, activity correction coefficient, and temperature constraint label of each unit in the matrix dimension, so as to achieve a comprehensive adaptation of the electrical stimulation behavior in terms of spatial distribution and temporal rhythm.

[0060] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0061] Input the real-time current control parameter matrix into an adaptive fuzzy control model, and the adaptive fuzzy control model includes a five-layer fuzzy inference network structure;

[0062] Perform forward calculation on the real-time current control parameter matrix through the five-layer fuzzy inference network structure, and use the root mean square propagation algorithm to optimize the network parameters of the adaptive fuzzy control model by gradient descent to obtain the updated network parameter values;

[0063] Based on the updated network parameter values, perform adaptive step size control and range limiting processing, and generate the optimal current intensity parameter and the optimal pulse frequency parameter through the adaptive fuzzy control model.

[0064] Specifically, the real-time current control parameter matrix is input into a structured fuzzy neural network control system. This control system consists of a five-layer neural network structure and has the ability of fuzzy logic reasoning, so as to perform non-linear feature mapping, fuzzy rule matching and output defuzzification processing on continuous input parameters. The input matrix contains multi-dimensional features such as the current intensity, current change rate, temperature influence factor, micropore activity, and humidity change rate of each electrode unit in the current control cycle. These multi-dimensional input information are recombined into a unified input vector stream through dimension expansion, and are calculated one by one in parallel according to each electrode unit to make full use of the network's expression ability in the spatial domain. After entering the adaptive fuzzy control model, it first passes through the input layer, which mainly completes the weighted integration and format normalization of the input vector and provides a basis for the calculation of the activation functions of subsequent hidden layers. Then comes the first hidden layer, which extracts the linear trend and interval amplitude relationship in the input features through the ReLU activation function and outputs the basic response structure of the humidity change rate, current trend fluctuation and activity non-uniformity; then it enters the second hidden layer, which completes the compression and bidirectional mapping of the intermediate state through the Tanh function, so that the signal can be effectively encoded in both high and low active states. The third hidden layer uses the Sigmoid activation function to enhance the output continuity of the model during the edge state change process, thereby avoiding drastic jumps in the output current and frequency. The fourth layer is the fuzzy rule layer, where several fuzzy control rules in the form of IF-THEN structures are embedded in the system. These rules are encoded as network parameters through the weight matrix and fuzzy subset mapping, and are matched one by one with the activation states of the hidden layer neurons to complete the recognition of fuzzy conditions and the quantitative calculation of the response intensity. The fifth layer is the output layer, which outputs two main continuous control variables, namely the optimal current intensity parameter and the optimal pulse frequency parameter, through a linear transformation function, and calculates the loss function according to the difference between the optimal output target and the current output result in the historical samples, providing a feedback basis for subsequent gradient optimization. To improve the stability and robustness of the model during training, the root mean square propagation algorithm is used as the optimization strategy to perform gradient descent updates on all weight parameters and bias vectors in the above five-layer network. In this process, the first-order gradient of the loss function is calculated, and then the weighted average is performed according to the square mean of this gradient in each round of iteration to obtain the mean square gradient estimate value of each parameter. The system uses this estimate value to update the learning rate, so that the update step size is automatically reduced in the region with severe gradient oscillation and a higher update speed is maintained in the stable descent interval, thereby improving the overall convergence efficiency. After each parameter update, the new network weights are fed back into the forward inference path and a forward calculation is performed again until the optimization task of the current cycle is completed and the error tolerance threshold is reached. After optimization, an adaptive step size control mechanism is implemented on the updated network parameter values.This mechanism determines whether the current network training tends to saturate by detecting the change amplitude of the loss function between consecutive rounds of parameter updates. If the loss reduction amplitude is lower than the threshold in three consecutive iterations, the learning rate is automatically reduced to avoid overfitting. If the loss reduction trend is still obvious, the current learning rate is maintained for stable update. To avoid the output current and frequency exceeding the physiological and device allowable ranges, a range-limiting processing strategy is introduced, that is, physical limitation is imposed on the final output result. When the current intensity is higher than 1.5 mA or lower than 0.5 mA, threshold truncation is automatically performed. Similarly, when the frequency is higher than 0.5 Hz or lower than 0.1 Hz, it is also automatically corrected to the nearest available value to ensure that the final output parameters are practically executable. Based on the output layer results after optimization and range-limiting processing, the optimal current intensity parameter and the optimal pulse frequency parameter in the current state are extracted from the five-layer fuzzy inference network. The current intensity parameter indicates the voltage or current value that each electrode unit should output, while the pulse frequency parameter controls the rhythm and intermittent structure of the stimulation signal. These two parameters together constitute the core configuration indicators of the electrical stimulation signal. Bind these two control quantities to the drive instructions within the current control cycle, and complete the final physical signal generation through underlying modules such as PWM modulation, voltage conversion, and channel control, thereby effectively converting the high-dimensional inference results of the fuzzy control model into precise stimulation behaviors on the insole platform and realizing closed-loop adaptive current regulation.

[0065] In a specific embodiment, the control method of the smart insole further includes the following steps:

[0066] Establish an intermittent pulse power limit model, and calculate the actual power value of each electrode unit according to the optimal current intensity parameter;

[0067] Based on the walking state data collected by the acceleration sensor, dynamically adjust the actual power value of each electrode unit to obtain a dynamic duty cycle control parameter;

[0068] Perform power self-adaptation adjustment on the dynamic duty cycle control parameter according to the remaining battery capacity to obtain a target duty cycle control value;

[0069] Design a linear extended state observer to estimate the interference factors generated during walking in real time, and combine the optimal current intensity parameter to construct an anti-disturbance control law to obtain the current control input value for anti-walking interference;

[0070] Combine the current control input value for anti-walking interference to optimize the pulse waveform of the optimal pulse frequency parameter to obtain optimized current pulse waveform parameters;

[0071] Combine the optimized current pulse waveform parameters with the target duty cycle control value, and allocate current intensity, frequency, duty cycle, and waveform parameters for each electrode unit to obtain a target current control sequence.

[0072] Specifically, an intermittent pulse power limit model for wearable electrical stimulation devices is established. The purpose of this model is to maximize the control of the average power output per unit time while meeting the treatment efficacy, so as to extend the device's battery life and reduce the risk of skin irritation. The model takes the optimal current intensity parameter as input, combines the actual contact resistance values of each electrode unit, calculates the instantaneous power consumption level of each electrode unit in the current cycle, compares this power consumption with the defined threshold, and sets the duty cycle structure for each electrode, that is, determines the distribution relationship between the working duration and the pause duration within every 30 seconds. In this way, the system assigns an initial working cycle ratio to each electrode. Based on the built-in triaxial acceleration sensor, the gait information of the user is continuously collected, including the acceleration amplitude of the foot, the direction change rate, and the stride frequency characteristics. The data is input into the motion state analysis module at a sampling rate of 100 times per second. The module classifies the walking state into four levels: stationary, mildly active, moderately active, and vigorously active, and each level corresponds to different degrees of interference and current stability requirements. When it is detected that the walking state is at a level above moderate, the system believes that the sole contact and liftoff are frequent, and there is a significant risk of intermittent stimulation interference. The originally set working cycle ratio is appropriately reduced to enhance the anti-interference ability and comfort of the system in a high-dynamic state. On the contrary, in the stationary or mildly active state, the working time is appropriately extended to improve the treatment efficiency. Therefore, each electrode unit adjusts its own working rhythm according to the motion level it is in, forming structured dynamic working cycle control parameters. Obtain the remaining capacity of the stored power in the current battery module, and compare it with the set capacity threshold range. If the remaining power is lower than the set lower limit, the power generation adaptive control strategy is triggered. This strategy further compresses and adjusts the generated dynamic working cycle control parameters, such as further shortening the working time or extending the pause duration, or even reducing the current peak, so as to achieve the purpose of extending the battery life and ensuring that the user can complete a full treatment course. Combining the above adjustment results, a set of target working cycle control values are generated. These values have been adjusted according to both the user's behavior state and the energy state, and can adapt to the complex changes in the actual wearing environment. To enhance the ability of the electrical stimulation control system to suppress the sole perturbation noise, a linear extended state observer is introduced as a dynamic interference estimation module. This module takes the current output as the observation object. When it detects that the sole current signal deviates due to external pace impacts or posture changes, it automatically extracts the non-target perturbation component, forms a difference between its estimated value and the target current input, and infers the perturbation trend. Furthermore, through the perturbation gain compensation mechanism inside the observer, an anti-interference control law is constructed, so that the current control signal in the next cycle can pre-avoid the characteristic path of the identified interference, thereby realizing feedforward anti-interference regulation. On this basis, the result of the anti-interference control law action is combined and superimposed with the optimal current intensity parameter to obtain an anti-interference corrected current control input value. The anti-interference control input value is used to adjust the optimal pulse frequency parameter generation mechanism, that is, to optimize the time rhythm of the current stimulation at the waveform level.Although traditional square-wave current stimulation is easy to implement, it is prone to sudden nerve shock sensations under high-dynamic conditions. Therefore, the original square-wave structure is replaced with a trapezoidal wave or adjustable gradient waveform structure. By setting the time parameters of the pulse rising edge and falling edge, the current gradually increases and decreases, avoiding skin irritation caused by steep voltage edges. At the same time, a suitable frequency range is set in combination with the peak current intensity and working duration, so that the stimulation rhythm is neither too frequent nor too long-spaced to cause a weakened stimulation effect. After this process, an optimized set of current pulse waveform parameters is generated. The above-mentioned optimized pulse waveform parameters are combined with the previously generated target duty cycle control value, and four types of parameters, namely current intensity, pulse frequency, duty cycle ratio, and output waveform structure, are accurately allocated to each electrode unit according to its current position number, activity level, and contact resistance state. These four types of parameters constitute the target current control sequence. The control sequence is sent to the drive control module through the SPI or I2C bus, and the per-channel current output is executed according to this sequence to ensure that each electrode meets the personalized treatment requirements within the current control cycle and also complies with the requirements of energy management, anti-interference, and comfort.

[0073] The control method of the smart insole in the embodiment of the present invention is described above. Next, the smart insole in the embodiment of the present invention is described. An embodiment of the smart insole in the embodiment of the present invention includes:

[0074] An acquisition module, configured to acquire temperature data, humidity data, and contact resistance data in the forefoot area and the arch area of the smart insole, and form a foot environment parameter matrix;

[0075] A distribution analysis module, configured to perform micropore activity calculation and distribution analysis on the foot environment parameter matrix to obtain current distribution output parameters;

[0076] A coordination module, configured to calculate the humidity gradient ratio between the forefoot area and the arch area according to the current distribution output parameters and implement a coordinated direct current control strategy to obtain a real-time current control parameter matrix;

[0077] An optimization module, configured to input the real-time current control parameter matrix into an adaptive fuzzy control model for gradient descent optimization processing to obtain an optimal current intensity parameter and an optimal pulse frequency parameter.

[0078] Through the cooperation of the above-mentioned components, a modular multi-electrode array structure layout is adopted, 5 humidity sensors and 3 temperature sensors are set in the forefoot area, 3 humidity sensors and 2 temperature sensors are set in the arch area, and the contact resistance is measured by the four-point probe method, which realizes high-density and multi-dimensional monitoring of foot environmental parameters, and greatly improves the spatial resolution and parameter acquisition accuracy of environmental feature identification. Based on micropore density analysis and activity distribution vector diagram technology, the present invention can accurately identify the micropore activity difference between the forefoot area and the arch area, calculate the current distribution parameters in a targeted manner, and realize precise regulation of different areas, effectively solving the problem of uneven current distribution in traditional technologies. By constructing a five-layer fuzzy reasoning network structure and a fuzzy rule base of 25 IF-THEN rules, the present invention realizes adaptive optimization of current intensity and pulse frequency. The system can automatically adjust the control parameters according to environmental changes, greatly improving the accuracy and intelligence level of environmental regulation. The present invention designs a linear expansion state observer and an anti-interference control law, which can effectively suppress random interference during walking, ensure control stability in a dynamic environment, and significantly improve the adaptability of the system in actual application scenarios. By adopting an intermittent pulse power limiting model with an initial working cycle ratio of 30 seconds working / 120 seconds pause, combined with an adaptive adjustment mechanism for the remaining battery capacity, the present invention achieves the long-term working ability of the smart insole, extends the single-charge usage time compared to traditional technologies, and improves the practicality of the device. By replacing the traditional square wave with a trapezoidal wave with a rise time and a fall time of 50ms, the present invention reduces the impact of current pulses on the electrodes and contact surfaces, improves the comfort and safety of current regulation, and at the same time reduces the aging rate of electrode materials, extending the service life of the equipment. By real-time acquisition and analysis of the change rate of contact surface humidity, the electrode array is intelligently triggered to power on when the regional humidity exceeds the threshold, thereby achieving on-demand regulation of the foot environment, avoiding unnecessary energy consumption, and ensuring the stability of the contact surface humidity.

[0079] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0080] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a XXXX device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0081] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A control method for an intelligent insole, characterized in that, Including: Collecting temperature data, humidity data, and contact resistance data in the forefoot area and the arch area of the intelligent insole to form a foot environmental parameter matrix; Performing micropore activity calculation and distribution analysis on the foot environmental parameter matrix to obtain current distribution output parameters; Calculating the humidity gradient ratio between the forefoot area and the arch area according to the current distribution output parameters and implementing a coordinated direct current control strategy to obtain a real-time current control parameter matrix; Inputting the real-time current control parameter matrix into an adaptive fuzzy control model for gradient descent optimization to obtain optimal current intensity parameters and optimal pulse frequency parameters.

2. The control method of the intelligent insole according to claim 1, wherein The collecting temperature data, humidity data, and contact resistance data in the forefoot area and the arch area of the intelligent insole to form a foot environmental parameter matrix includes: Setting E humidity sensors and F temperature sensors in the forefoot area of the intelligent insole and M humidity sensors and N temperature sensors in the arch area to form a sensor array; Collecting foot contact surface humidity data through the humidity sensors in the sensor array and performing digital low-pass filtering on the foot contact surface humidity data to obtain a humidity data matrix; Collecting foot contact surface temperature data through the temperature sensors in the sensor array and performing digital low-pass filtering on the foot contact surface temperature data to obtain a temperature data matrix; Measuring the contact resistance data in the forefoot area and the arch area by the four-point probe method to obtain a contact resistance data matrix, and performing normalization processing on the humidity data matrix, the temperature data matrix, and the contact resistance data matrix to obtain a foot environmental parameter matrix.

3. The control method of the intelligent insole according to claim 2, characterized in that The performing micropore activity calculation and distribution analysis on the foot environmental parameter matrix to obtain current distribution output parameters includes: Calculating the micropore activity index in the forefoot area and the arch area based on the contact resistance data and the foot contact surface temperature data in the foot environmental parameter matrix to obtain the micropore activity index in the forefoot area and the arch area; Performing regional weight analysis on the micropore activity index according to the structural characteristics of the micropore density in the forefoot area and the arch area to obtain a micropore activity distribution vector diagram; Calculating the initial current parameters on the forefoot side and the arch side based on the micropore activity distribution vector diagram and the contact resistance data; Performing dynamic power distribution based on the initial current parameters on the forefoot side and the arch side to obtain a power balance equation on the forefoot side and the arch side; Calculating the first humidity change rate according to the foot contact surface humidity data in the foot environmental parameter matrix, and adjusting the coefficients of the power balance equation based on the first humidity change rate to obtain current distribution coefficients; Calculating the current output of each electrode unit according to the current distribution coefficients and the micropore activity distribution vector diagram to obtain current distribution output parameters.

4. The control method of the intelligent insole according to claim 3, wherein The calculating the initial current parameters on the forefoot side and the arch side based on the micropore activity distribution vector diagram and the contact resistance data includes: Performing regional division on the micropore activity distribution vector diagram to obtain the regional division result of the electrode unit; Based on the result of the electrode unit area division, perform area matching calculation on the contact resistance data to obtain the average contact resistance value of the forefoot area and the average contact resistance value of the arch area; According to the average contact resistance value of the forefoot area and the average contact resistance value of the arch area, establish current equation models for the forefoot side and the arch side of the forefoot area and the arch area respectively; Based on the micropore activity distribution vector diagram, perform dynamic adjustment calculation on the forefoot area voltage and the arch area voltage to obtain the voltage parameters of the forefoot side and the arch side; Substitute the voltage parameters of the forefoot side and the arch side into the current equation model for calculation to obtain the initial current values of the forefoot side and the arch side; Perform electrode unit allocation calculation on the initial current values of the forefoot side and the arch side to obtain the initial current parameters of the forefoot side and the arch side.

5. The control method of the intelligent insole according to claim 1, characterized in that, Calculating the humidity gradient ratio between the forefoot area and the arch area according to the current distribution output parameters and implementing a coordinated direct current control strategy to obtain a real-time current control parameter matrix, including: Obtain real-time contact surface humidity data and calculate the second humidity change rate, and calculate the humidity gradient ratio between the forefoot area and the arch area based on the second humidity change rate; Generate a regional coordination control coefficient according to the humidity gradient ratio, and dynamically adjust the reference current value in the current distribution output parameters based on the regional coordination control coefficient to obtain the real-time current values of the forefoot side and the arch side; Perform differential adjustment on the real-time current values of the forefoot side and the arch side and set the current change rate of the forefoot side and the arch side according to the second humidity change rate to obtain dynamic current adjustment parameters; Based on the dynamic current adjustment parameters, the real-time micropore activity index and the real-time contact surface temperature value, perform current balance and temperature compensation to obtain a real-time current control parameter matrix.

6. The control method of the intelligent insole according to claim 1, wherein, Inputting the real-time current control parameter matrix into an adaptive fuzzy control model for gradient descent optimization to obtain the optimal current intensity parameter and the optimal pulse frequency parameter, including: Input the real-time current control parameter matrix into an adaptive fuzzy control model, and the adaptive fuzzy control model includes a five-layer fuzzy inference network structure; Perform forward calculation on the real-time current control parameter matrix through the five-layer fuzzy inference network structure, and use the root mean square propagation algorithm to optimize the network parameters of the adaptive fuzzy control model by gradient descent to obtain the updated network parameter values; Based on the updated network parameter values, perform adaptive step size control and range limitation processing, and generate the optimal current intensity parameter and the optimal pulse frequency parameter through the adaptive fuzzy control model.

7. The control method of the intelligent insole according to claim 1, characterized in that, The control method of the intelligent insole further includes: Establish an intermittent pulse power limit model, and calculate the actual power value of each electrode unit according to the optimal current intensity parameter; Based on the walking state data collected by the acceleration sensor, dynamically adjust the actual power value of each electrode unit to obtain dynamic duty cycle control parameters; Perform power self-adaptive adjustment on the dynamic duty cycle control parameters according to the remaining battery capacity to obtain the target duty cycle control value; Design a linear extended state observer to estimate the interference factors generated during walking in real time, and construct a disturbance rejection control law in combination with the optimal current intensity parameter to obtain the current control input value for rejecting walking interference; Optimize the pulse waveform of the optimal pulse frequency parameter in combination with the current control input value for rejecting walking interference to obtain the optimized current pulse waveform parameter; Combine the optimized current pulse waveform parameter with the target duty cycle control value, and allocate current intensity, frequency, duty cycle, and waveform parameters to each electrode unit to obtain the target current control sequence.

8. An intelligent insole, characterized in that, A control method for implementing the intelligent insole according to any one of claims 1-7, wherein the intelligent insole comprises: An acquisition module, configured to acquire temperature data, humidity data, and contact resistance data in the forefoot area and the arch area of the intelligent insole to form a foot environment parameter matrix; A distribution analysis module, configured to calculate the micropore activity and perform distribution analysis on the foot environment parameter matrix to obtain current distribution output parameters; A coordination module, configured to calculate the humidity gradient ratio between the forefoot area and the arch area according to the current distribution output parameters and implement a coordinated direct current control strategy to obtain a real-time current control parameter matrix; An optimization module, configured to input the real-time current control parameter matrix into an adaptive fuzzy control model for gradient descent optimization processing to obtain an optimal current intensity parameter and an optimal pulse frequency parameter.

Images