Method, device and equipment for improving massage strength accuracy of neck massage instrument and medium

Abstract

The invention discloses a method, a device, equipment and a medium for improving massage strength accuracy of a neck massager, and the method comprises the following steps: in the massage process of the neck massager, acquiring neck pressure feedback data collected by pressure sensors arranged in different partitions of the neck massager; according to the neck pressure feedback data and the preset pressure threshold range of each subarea, determining a target comprehensive error of each subarea and a single error pressure value of each pressure sensor; according to the historical target comprehensive error corresponding to each partition, integration is carried out on the target comprehensive error of each partition to obtain a steady-state error, and error change rate calculation is carried out on the target comprehensive error of each partition to obtain an error change rate; performing proportional differential integral operation according to the target comprehensive error, the steady state error and the error change rate of each partition and the single error pressure value of each pressure sensor, and determining a required massage force parameter of the massage head; and adjusting the massage head pressure of the massage head by taking the required massage force parameter as a target.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent massage equipment, and in particular to a method, a device, a equipment and a medium for improving the massage force accuracy of a neck massager. Background Art

[0002] As a common health care device, neck massagers are widely used to relieve neck fatigue, improve blood circulation and relax muscles. However, existing neck massagers have the following problems in terms of massage intensity control: Inaccurate force control: Traditional neck massagers usually use fixed massage modes and cannot dynamically adjust the massage force according to individual differences of users (such as body shape and muscle condition). Too much force may cause discomfort or injury, while too little force cannot achieve the ideal massage effect.

[0003] Lack of real-time feedback mechanism: Existing devices lack the ability to perceive the user's real-time status and cannot dynamically adjust the massage intensity based on the user's feedback. For example, when the user's muscles are tense, the device cannot automatically increase the intensity; when the user feels uncomfortable, the device cannot reduce the intensity in time. Summary of the invention

[0004] The purpose of the present invention is to provide a method, device, equipment and medium for improving the accuracy of massage intensity of a neck massager, so as to solve the technical problem that the neck massager cannot adjust the massage intensity according to individual differences and cannot dynamically adjust the massage intensity according to user feedback.

[0005] The purpose of the present invention can be achieved through the following technical solutions: In a first aspect, a method for improving the massage force accuracy of a neck massager is provided, comprising: During the massage process of the neck massager, neck pressure feedback data collected by pressure sensors arranged in different partitions of the neck massager are obtained in real time, and the pressure sensors are arranged in a one-to-one correspondence with the massage heads of the massager; Based on the cross attention mechanism, according to the neck pressure feedback data corresponding to the different pressure sensors and the preset pressure threshold range of the partitions corresponding to the pressure sensors, the target comprehensive error of the pressure sensors in each partition and the single error pressure value of each pressure sensor are determined; According to the historical target comprehensive errors corresponding to each of the partitions, the target comprehensive errors of each of the partitions are integrated to obtain the steady-state errors of each of the partitions, and the error change rate of the target comprehensive errors of each of the partitions is calculated to obtain the error change rate of each of the partitions; Perform proportional differential integral calculations based on the target comprehensive error, the steady-state error, the error change rate, and the single error pressure value of each pressure sensor of each partition to determine the required massage intensity parameter for the massage head where each pressure sensor is located; The massage head pressure of the massage head where each pressure sensor in the neck massager is located is adjusted based on the required massage intensity parameter.

[0006] In a second aspect, a device for improving the massage force accuracy of a neck massager is provided, comprising: an acquisition module, configured to acquire, in real time during the massage process of the neck massager, neck pressure feedback data collected by pressure sensors arranged in different partitions of the neck massager, wherein the pressure sensors are arranged in one-to-one correspondence with the massage heads of the massager; A first determination module is configured to determine, based on a cross-attention mechanism, a target comprehensive error of the pressure sensors in each partition and a single error pressure value of each pressure sensor according to the neck pressure feedback data corresponding to different pressure sensors and a preset pressure threshold range of each partition corresponding to each pressure sensor; a calculation module configured to integrate the target comprehensive error of each partition according to the historical target comprehensive error corresponding to each partition to obtain the steady-state error of each partition, and to calculate the error change rate of the target comprehensive error of each partition to obtain the error change rate of each partition; The second determination module is configured to perform proportional differential integral operations according to the target comprehensive error of each partition, the steady-state error, the error change rate and the single error pressure value of each pressure sensor to determine the required massage intensity parameter for the massage head where each pressure sensor is located; The adjustment module is configured to adjust the massage head pressure of the massage head where each pressure sensor in the neck massager is located based on the required massage intensity parameter.

[0007] According to a third aspect, an electronic device is provided, including: processor; a memory for storing processor-executable instructions; The processor is configured to execute the executable instructions stored in the memory to implement any method described in the first aspect.

[0008] According to a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of any one of the methods described in the first aspect are implemented.

[0009] Compared with the existing solutions, the present invention achieves the following beneficial effects: By acquiring the data of pressure sensors set in different partitions of the neck massager in real time, the present invention can accurately capture the real-time pressure feedback of each area of ​​the user's neck. It ensures that the massage intensity can closely fit the individual differences of users, whether it is the difference in the tension of the neck muscles or the difference in personal preference for massage intensity, it can be accurately responded and adjusted. The introduction of the cross-attention mechanism not only takes into account the real-time data of the current pressure sensor, but also combines the preset pressure threshold range, so that it can intelligently identify and process the target comprehensive error and single error pressure value in each partition, so that the massager can dynamically and intelligently adjust the massage intensity to ensure that every massage is both effective and comfortable.

[0010] Furthermore, by integrating the target comprehensive error of each partition to obtain the steady-state error and calculating the error change rate, dual control of massage intensity adjustment is achieved - taking into account both the long-term stable massage demand (steady-state error) and the instantaneous user feedback (error change rate), enhancing the adaptability and flexibility of the massager. Based on proportional differential integral operations, according to the target comprehensive error, steady-state error, error change rate and single error pressure value of each partition, the required massage intensity parameters of each massage head are accurately calculated, ensuring that the adjustment of massage intensity is both accurate and efficient, and can maximize the personalized needs of users. It not only improves the accuracy and effectiveness of massage, but also significantly enhances the overall user experience. Users can enjoy a more personalized, comfortable and effective neck massage service based on their own feelings, reducing discomfort or ineffective massage that may be caused by improper massage intensity.

[0011] In summary, by integrating advanced pressure sensing technology, cross-attention mechanism, steady-state and dynamic error control strategies, and PID operation method, the technical difficulties of neck massagers in adjusting massage intensity are solved, providing users with a more personalized, intelligent and efficient neck massage experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The present invention will be further described below in conjunction with the accompanying drawings.

[0013] Figure 1 It is a flow chart of a method for improving the massage intensity accuracy of a neck massager according to an embodiment of the specification.

[0014] Figure 2 It is a block diagram of a device for improving the massage force accuracy of a neck massager according to an embodiment of the specification.

[0015] Figure 3 It is a block diagram of another device for improving the massage intensity accuracy of a neck massager according to an embodiment of the specification. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] like Figure 1 The flowchart of a method for improving the massage force accuracy of a neck massager of the present invention is shown, comprising: In step S11, during the massage process of the neck massager, neck pressure feedback data collected by pressure sensors arranged in different partitions of the neck massager are obtained in real time, and the pressure sensors are arranged in a one-to-one correspondence with the massage heads of the massager; Among them, a plurality of pressure sensors are set inside the neck massager, and these sensors correspond to the massage heads of the massager one by one. During the massage process, each pressure sensor will collect the neck pressure data of its area in real time and send these data to the control system for processing.

[0018] For example, suppose a neck massager has 40 massage heads, and a pressure sensor is set under each massage head. When the massager starts working, these 40 pressure sensors will simultaneously collect pressure data of various areas of the neck.

[0019] In step S12, based on the cross-attention mechanism, according to the neck pressure feedback data corresponding to the different pressure sensors and the preset pressure threshold range of the partitions corresponding to the pressure sensors, the target comprehensive error of the pressure sensors in each partition and the single error pressure value of each pressure sensor are determined; In the disclosed embodiment, the cross-attention mechanism is used to process the data of each pressure sensor, and combined with the preset pressure threshold range, the target comprehensive error of each partition and the single error pressure value of each pressure sensor are calculated. This process helps to identify the pressure distribution of each area of ​​the neck and the direction in which the massage head needs to be adjusted.

[0020] For example, if the preset pressure threshold range of a certain partition is 10-20N, and the data collected by a pressure sensor in the partition is 15N, then the single error pressure value of this sensor is 0 (because 15N is within the preset range). But if the data collected by another sensor is 25N, then its single error pressure value is 5N (exceeding the preset range of 5N). The target comprehensive error of the partition can be calculated based on these data, for example, the sum of all single error pressure values ​​in the partition is calculated to obtain the target comprehensive error.

[0021] In step S13, according to the historical target comprehensive errors corresponding to each of the partitions, the target comprehensive errors of each of the partitions are integrated to obtain the steady-state errors of each of the partitions, and the error change rate of the target comprehensive errors of each of the partitions is calculated to obtain the error change rate of each of the partitions; Among them, the steady-state error is used to measure the stable deviation between the actual output and the expected output of the neck massager after long-term operation. The error change rate is used to indicate how fast the output error of the neck massager changes over time.

[0022] In the disclosed embodiment, the target comprehensive error of each partition is integrated according to the historical target comprehensive error data to obtain the steady-state error. At the same time, the target comprehensive error of each partition is differentiated with time to obtain the error change rate. These two parameters together reflect the current state and change trend of the system.

[0023] For example: Assume that the target comprehensive error of a partition has been fluctuating around 2N over the past period of time. By integrating these historical data, we can get the steady-state error of the partition to be 2N. If the target comprehensive error has gradually decreased over the recent period of time, the error change rate is a negative value, indicating that it is converging to the desired state.

[0024] In step S14, proportional differential integral operation is performed according to the target comprehensive error, the steady-state error, the error change rate and the single error pressure value of each pressure sensor of each partition to determine the required massage intensity parameter for the massage head where each pressure sensor is located; In the disclosed embodiment, PID operation is performed based on the target comprehensive error, steady-state error, error change rate of each partition and the single error pressure value of each pressure sensor. The PID operation calculates the massage intensity parameters required for each massage head by adjusting the three parameters of proportion, differential and integral. These parameters will be used to guide the massage head to adjust its pressure output.

[0025] For example: Assume that the target comprehensive error of a certain partition is 2N, the steady-state error is 1N, the error change rate is -0.5N / s, and the single error pressure value of a pressure sensor in the partition is 0.5N. The control system will input these data into the PID operator, and after calculation, the massage intensity parameters required by the massage head will be obtained. If the result of the PID operation is to increase the pressure output, the pressure of the massage head will be adjusted accordingly.

[0026] In step S15, the massage head pressure of the massage head where each pressure sensor in the neck massager is located is adjusted based on the required massage intensity parameter.

[0027] In the disclosed embodiment, the pressure output of the massage head is changed by adjusting the motor or other actuator inside the massager according to the determined massage intensity parameter. This process realizes the dynamic adjustment and personalized customization of the massage intensity of the neck massager. For example: suppose the massage intensity parameter determined by a massage head is to increase the pressure by 1N. The motor output of the massage head is adjusted accordingly to increase the pressure by 1N. In this way, the user can feel the massage intensity that better suits his or her neck needs.

[0028] The above technical solution acquires the data of pressure sensors set in different partitions of the neck massager in real time, and the present invention can accurately capture the real-time pressure feedback of each area of ​​the user's neck. It ensures that the massage intensity can closely fit the individual differences of users, whether it is the difference in the tension of the neck muscles or the difference in personal preference for massage intensity, it can get accurate response and adjustment. The introduction of the cross-attention mechanism not only takes into account the real-time data of the current pressure sensor, but also combines the preset pressure threshold range, so that it can intelligently identify and process the target comprehensive error and single error pressure value in each partition, so that the massager can dynamically and intelligently adjust the massage intensity to ensure that every massage is both effective and comfortable.

[0029] Furthermore, by integrating the target comprehensive error of each partition to obtain the steady-state error and calculating the error change rate, dual control of massage intensity adjustment is achieved - taking into account both the long-term stable massage demand (steady-state error) and the instantaneous user feedback (error change rate), enhancing the adaptability and flexibility of the massager. Based on proportional differential integral operations, according to the target comprehensive error, steady-state error, error change rate and single error pressure value of each partition, the required massage intensity parameters of each massage head are accurately calculated, ensuring that the adjustment of massage intensity is both accurate and efficient, and can maximize the personalized needs of users. It not only improves the accuracy and effectiveness of massage, but also significantly enhances the overall user experience. Users can enjoy a more personalized, comfortable and effective neck massage service based on their own feelings, reducing discomfort or ineffective massage that may be caused by improper massage intensity.

[0030] In summary, by integrating advanced pressure sensing technology, cross-attention mechanism, steady-state and dynamic error control strategies, and PID operation method, the technical difficulties of neck massagers in adjusting massage intensity are solved, providing users with a more personalized, intelligent and efficient neck massage experience.

[0031] In a possible implementation, in step S12, based on the cross-attention mechanism, according to the neck pressure feedback data corresponding to different pressure sensors and the preset pressure threshold range of the partitions corresponding to the pressure sensors, the target comprehensive error of the pressure sensors in each partition and the single error pressure value of each pressure sensor are determined, including: In step S121, according to the neck pressure feedback data corresponding to each pressure sensor and the preset pressure threshold range of the partition corresponding to the pressure sensor, the error pressure value of each pressure sensor exceeding the preset pressure threshold range is determined; The error pressure value is used to indicate the deviation value caused by the neck pressure feedback data collected by the pressure sensor exceeding or falling below the preset pressure threshold range of the corresponding partition.

[0032] In the disclosed embodiment, for each pressure sensor, the collected neck pressure feedback data is compared with the preset pressure threshold range of the corresponding partition. If the collected data exceeds the preset range, the absolute value of the excess portion is calculated as the error pressure value. This helps to identify which pressure sensor areas need to adjust the pressure.

[0033] For example, suppose the preset pressure threshold range of a certain zone is 15-20N, and the data collected by a certain pressure sensor is 22N. Then, the error pressure value of the sensor is 2N (22N-20N=2N), indicating that the pressure in this area is too high.

[0034] In step S122, a total error pressure value of each partition is calculated according to the error pressure value of each pressure sensor in each partition; The total error pressure value is used to represent the sum of the error pressure values ​​of all pressure sensors in the partition, reflecting the degree to which the overall pressure of the partition deviates from the preset threshold range.

[0035] In the disclosed embodiment, for each partition, the error pressure values ​​of all pressure sensors in the partition are accumulated to obtain the total error pressure value of the partition. For example, suppose there are 3 pressure sensors in a partition, and their error pressure values ​​are 1N, 2N and 0.5N respectively. Then, the total error pressure value of the partition is 3.5N (1N+2N+0.5N=3.5N).

[0036] In step S123, a target comprehensive error of the pressure sensor in each partition is determined according to the error pressure value of each pressure sensor, the total error pressure value corresponding to the partition where the pressure sensor is located, and a preset weight of each pressure sensor, wherein the preset weight is negatively correlated with the distance of the pressure sensor in the corresponding partition relative to the pressure sensor at the center of the partition; The target comprehensive error is used to represent the index reflecting the comprehensive deviation of each partition pressure sensor calculated by combining the error pressure value of each pressure sensor, the total error pressure value of the partition, and the preset weight. The preset weight is a weight set according to the relative position of the pressure sensor in the partition (such as the distance from the center of the partition) and is used to adjust the contribution of pressure sensors at different positions to the target comprehensive error.

[0037] In the disclosed embodiment, for each partition, the target comprehensive error of each pressure sensor is calculated based on the error pressure value of each pressure sensor, the total error pressure value of the partition, and the preset weight. The preset weight is usually negatively correlated with the distance of the pressure sensor from the center of the partition, that is, the farther the sensor is from the center, the smaller its weight is, and the lower its contribution to the target comprehensive error is. In this way, the comprehensive deviation of the pressure sensors in each partition can be reflected more accurately.

[0038] For example: Assume that there are 3 pressure sensors in a certain partition, and their error pressure values ​​are 1N, 2N and 0.5N respectively, and the preset weights are 0.8 (close to the center), 0.6 (middle distance) and 0.4 (far from the center). Then, the target comprehensive error of a pressure sensor in the partition can be calculated by the following formula: Target comprehensive error = (error pressure value ÷ total error pressure value) × preset weight × total error pressure value of the partition. Taking the first sensor as an example, its target comprehensive error is (1N ÷ 3.5N) × 0.8 × 3.5N = 0.8N. The target comprehensive error of each pressure sensor in each partition can be accurately calculated.

[0039] In step S124, the pressure sensors adjacent to each of the pressure sensors in the partition are used as first cross attention sensors, and the pressure sensors adjacent to each of the pressure sensors in the adjacent partitions of the partition where the pressure sensors are located are used as second cross attention sensors; In the disclosed embodiment, the adjacent sensors of each pressure sensor in its partition are first identified (first cross-attention sensors), and then the sensors in the adjacent partitions of the partition corresponding to or close to the target pressure sensor are further identified (second cross-attention sensors). The purpose of this step is to build a cross-attention network to improve the accuracy of pressure distribution assessment by integrating local and cross-region pressure information.

[0040] For example: suppose there is a 3x3 partition grid, and each partition has multiple pressure sensors. For a pressure sensor located in the middle partition (assuming it is sensor A), its first cross-attention sensor will be the four adjacent sensors in the partition (assuming it is sensors B, C, D, E), and there may be no second cross-attention sensor. The second cross-attention sensor is a sensor that sensor A does not have in the same zone in a certain direction, and there are other partitions in that direction. If there are no adjacent partitions in that direction, then sensor A will not have a second cross-attention sensor. The second cross-attention sensor may be any of the four adjacent partitions, the upper, lower, left, and right, that are adjacent to the partition where sensor A is located, and the sensor position corresponds to or is close to sensor A (assuming it is sensors F, G, H, I, but it depends on the partition layout and sensor position). In other words, it is possible that sensor A has no adjacent sensors in the same partition on the right, so sensor A may have only three adjacent sensors in the same zone, and one adjacent sensor in a different partition.

[0041] In step S125, a single error pressure value of each of the pressure sensors is determined based on the error pressure value of each of the pressure sensors, the error pressure value of the corresponding first cross attention sensor, and the error pressure value of the corresponding second cross attention sensor.

[0042] Among them, the single error pressure value refers to a single indicator reflecting the comprehensive error of the pressure sensor calculated by comprehensively considering the error pressure value of the target pressure sensor itself, the error pressure value of its first cross-attention sensor and the error pressure value of the second cross-attention sensor.

[0043] In the disclosed embodiment, a single error pressure value of each pressure sensor is calculated based on the error pressure value of each pressure sensor and the error pressure values ​​of the first cross-attention sensor and the second cross-attention sensor adjacent thereto through a certain weighted or average algorithm. The purpose of this step is to integrate the local and cross-regional pressure information into a single indicator so as to more intuitively evaluate the error of each pressure sensor.

[0044] For example: Assume that the error pressure value of sensor A is 2N, the error pressure values ​​of its first cross-attention sensors B, C, and D are 1N, 1.5N, and 0.5N respectively, and the error pressure values ​​of the second cross-attention sensors F, and F are 1.2N respectively. Based on these values, a single error pressure value of sensor A can be calculated by weighted average or other algorithms. For example, the error pressure values ​​of all relevant sensors can be simply added and divided by the total number of sensors (including target sensors and cross-attention sensors) to obtain the average error pressure value, and then the single error pressure value can be obtained by weighted summing with the error pressure value of sensor A of 2N. However, more complex algorithms may consider the weights or distance factors of different sensors to more accurately reflect the actual situation.

[0045] In a possible implementation, in step S123, determining a target comprehensive error of the pressure sensor in each partition according to the error pressure value of each pressure sensor, the total error pressure value corresponding to the partition where the pressure sensor is located, and the preset weight of each pressure sensor includes: In step S1231, the product of the error pressure value of each pressure sensor and the preset weight corresponding to the pressure sensor is calculated to obtain a relative error pressure value corresponding to each pressure sensor; Among them, the relative error pressure value refers to the product of the error pressure value of the pressure sensor and its preset weight, which reflects the relative size of the sensor error relative to other sensors after considering the weight. For each pressure sensor, its error pressure value is multiplied by the preset weight to obtain the relative error pressure value. The purpose of this step is to adjust the error value of the sensor according to factors such as the location and importance of the sensor, so that the final error assessment is more accurate. For example: Assuming that the error pressure value of a sensor is 3N and the preset weight is 0.6, the relative error pressure value is 3N × 0.6 = 1.8N.

[0046] In step S1232, the ratio between the relative error pressure value corresponding to each of the pressure sensors and the total error pressure value corresponding to the partition where the pressure sensor is located is calculated to determine the error ratio of each of the pressure sensors in each partition; Among them, the error ratio value refers to the ratio between the relative error pressure value of the pressure sensor and the total error pressure value of the partition where it is located, which reflects the proportion of the sensor error in the total error of the partition. For each pressure sensor, its relative error pressure value is divided by the total error pressure value of the partition where it is located to obtain the error ratio value. The purpose of this step is to quantify the contribution of each sensor error to the overall error of the partition. For example: assuming that the total error pressure value of a partition is 10N, and the relative error pressure value of a sensor is 2N, then the error ratio value is 2N / 10N = 0.2.

[0047] In step S1233, based on the preset k value in the ksigma algorithm, according to the standard deviation and mean of the error proportion value of each pressure sensor in each partition, the K times standard deviation interval of the partition is calculated; Among them, the K times standard deviation interval is an interval calculated based on the standard deviation and mean of the error share value based on the ksigma algorithm, which is used to identify abnormal or deviating error share values. For each partition, the standard deviation and mean of the error share values ​​of all pressure sensors are calculated, and then the K times standard deviation interval is calculated based on the preset k value (such as k=3 in the 3sigma principle). The purpose of this step is to identify pressure sensors with abnormal error share values. For example: Assuming that the mean of the error share value of a partition is 0.1, the standard deviation is 0.05, and the preset k value is 3, then the K times standard deviation interval is [0.1 -30.05, 0.1 + 30.05] = [0.05, 0.25].

[0048] In step S1234, the pressure sensor in each partition whose error proportion value is within the corresponding K times standard deviation interval is used as the target pressure sensor corresponding to the partition, and the sum of the error pressure values ​​of the target pressure sensors corresponding to the partition is calculated to obtain the target comprehensive error of the pressure sensors in each partition.

[0049] Among them, target pressure sensor: refers to the pressure sensor whose error ratio is within the K times standard deviation interval. These sensors are regarded as the main error source or key sensor of the partition. Target comprehensive error: refers to the sum of the error pressure values ​​of all target pressure sensors in the partition, reflecting the comprehensive situation of the overall error of the partition.

[0050] For each partition, the pressure sensors whose error ratios are within the K-times standard deviation interval are taken as target pressure sensors, and the sum of the error pressure values ​​of these sensors is calculated to obtain the target comprehensive error. The purpose of this step is to identify the main error sources within the partition and quantify their overall impact.

[0051] For example: Assume that there are three pressure sensors in a certain partition, and their error percentage values ​​are 0.15, 0.22, and 0.08 respectively, and the K-fold standard deviation interval of the partition is [0.05, 0.25]. Then the target pressure sensors are two sensors with error percentage values ​​of 0.15 and 0.22, and their error pressure values ​​are 3N and 4N respectively, so the target comprehensive error is 3N + 4N = 7N.

[0052] In a possible implementation, in step S125, determining a single error pressure value of each of the pressure sensors according to the error pressure value of each of the pressure sensors, the error pressure value of the corresponding first cross attention sensor, and the error pressure value of the corresponding second cross attention sensor includes: In step S1251, the product of the error pressure value of each of the pressure sensors and the error pressure value of the corresponding first cross attention sensor is calculated to obtain a first calculation parameter, and the product of the error pressure value of each of the pressure sensors and the error pressure value of the corresponding second cross attention sensor is calculated to obtain a second calculation parameter; The first calculation parameter refers to the product of the error pressure value of the pressure sensor and the error pressure value of the corresponding first cross-attention sensor, reflecting the degree of correlation between the errors of adjacent sensors in the local area.

[0053] The second calculation parameter refers to the product of the error pressure value of the pressure sensor and the error pressure value of the corresponding second cross-attention sensor, which reflects the degree of correlation between the errors of adjacent sensors across regions.

[0054] First, the error pressure value of each pressure sensor is multiplied by the error pressure value of the first cross-attention sensor adjacent to it to obtain the first calculation parameter. At the same time, the error pressure value of each pressure sensor is multiplied by the error pressure value of the second cross-attention sensor adjacent to the partition where it is located to obtain the second calculation parameter. The purpose of this step is to quantify the mutual influence between adjacent sensor errors by calculating the product. For example: Assuming that the error pressure value of a pressure sensor is 2N, and the error pressure value of the corresponding first cross-attention sensor is 1.5N, the first calculation parameter is 2 × 1.5 = 3.

[0055] In step S1252, a logarithmic function value of the first calculation parameter corresponding to each of the pressure sensors is calculated to obtain a third calculation parameter, and a logarithmic function value of the second calculation parameter corresponding to each of the pressure sensors is calculated to obtain a fourth calculation parameter; The third calculation parameter refers to the logarithmic function value of the first calculation parameter, and the logarithmic function is used to further amplify or reduce the degree of correlation between the errors of adjacent sensors in the local area. The fourth calculation parameter refers to the logarithmic function value of the second calculation parameter, and the logarithmic function is used to further amplify or reduce the degree of correlation between the errors of adjacent sensors across regions.

[0056] The logarithmic function is applied to the first calculation parameter and the second calculation parameter to obtain the third calculation parameter and the fourth calculation parameter. The purpose of this step is to further emphasize or weaken the mutual influence between the errors of adjacent sensors through the nonlinear characteristics of the logarithmic function, so as to more accurately reflect the comprehensive error of each pressure sensor.

[0057] In step S1253, the sum of the first calculation parameter and the corresponding second calculation parameter corresponding to each of the pressure sensors is calculated to obtain a fifth calculation parameter, and the logarithmic function value of the fifth calculation parameter corresponding to each of the pressure sensors is calculated to obtain a sixth calculation parameter; Among them, the fifth calculation parameter refers to the sum of the first calculation parameter and the second calculation parameter corresponding to the pressure sensor, reflecting the comprehensive influence of the errors of adjacent sensors in the local area and across regions. The sixth calculation parameter refers to the logarithmic function value of the fifth calculation parameter, which further emphasizes or weakens the influence of the comprehensive error through the logarithmic function.

[0058] First, the sum of the first calculation parameter and the second calculation parameter of each pressure sensor is calculated to obtain the fifth calculation parameter. The purpose of this step is to combine the error effects of adjacent sensors within the local area and across regions into a comprehensive indicator. Then, a logarithmic function is applied to the fifth calculation parameter to obtain the sixth calculation parameter. The purpose of this step is to further emphasize or weaken the impact of the comprehensive error through the nonlinear characteristics of the logarithmic function, so as to more accurately evaluate the error of each pressure sensor in the subsequent steps.

[0059] For example: Assuming that the first calculation parameter of a pressure sensor is 3 and the second calculation parameter is 2, the fifth calculation parameter is 3 + 2 = 5. If the base of the logarithmic function is 2, the sixth calculation parameter is log25.

[0060] In step S1254, a ratio of the third calculation parameter corresponding to each pressure sensor to the corresponding sixth calculation parameter is calculated to obtain a first cross-attention weight corresponding to each pressure sensor, and a ratio of the fourth calculation parameter corresponding to each pressure sensor to the corresponding sixth calculation parameter is calculated to obtain a second cross-attention weight corresponding to each pressure sensor; Among them, the first cross-attention weight refers to the ratio of the third calculation parameter corresponding to the pressure sensor to the sixth calculation parameter, which reflects the relative influence of the adjacent sensor errors in the local area on the target sensor error.

[0061] The second cross-attention weight refers to the ratio of the fourth calculation parameter to the sixth calculation parameter corresponding to the pressure sensor, which reflects the relative influence of the cross-region adjacent sensor error on the target sensor error.

[0062] The ratio of the third calculation parameter to the sixth calculation parameter of each pressure sensor is calculated to obtain the first cross attention weight. At the same time, the ratio of the fourth calculation parameter to the sixth calculation parameter of each pressure sensor is calculated to obtain the second cross attention weight. By calculating the ratio, the relative influence of the errors of adjacent sensors within and across the local area on the error of the target sensor is quantified.

[0063] In step S1255, the product of the first cross-attention weight corresponding to each pressure sensor and the error pressure value of the corresponding first cross-attention sensor is subtracted from the error pressure value corresponding to each pressure sensor, and the product of the second cross-attention weight corresponding to each pressure sensor and the error pressure value of the corresponding second cross-attention sensor is added to determine a single error pressure value of each pressure sensor.

[0064] Among them, according to the error pressure value of each pressure sensor, the first cross attention weight, the error pressure value of the first cross attention sensor, the second cross attention weight and the error pressure value of the second cross attention sensor, a single error pressure value of each pressure sensor is calculated by a certain algorithm. The purpose of this step is to comprehensively consider the influence of the error of each pressure sensor itself and the error of adjacent sensors, so as to obtain a more accurate error index.

[0065] For example: Assuming that the error pressure value of a pressure sensor is P, the first cross attention weight is W1, the error pressure value of the first cross attention sensor is P1, the second cross attention weight is W2, and the error pressure value of the second cross attention sensor is P2. Then the single error pressure value of the pressure sensor can be calculated as follows: Single error pressure value = P - W1 × P1 + W2 × P2. It reflects the interaction between the target sensor error and the adjacent sensor error, and is adjusted by the weight coefficient.

[0066] In a possible implementation, in step S14, the proportional differential integral operation is performed according to the target comprehensive error of each partition, the steady-state error, the error change rate, and the single error pressure value of each pressure sensor to determine the required massage intensity parameter for the massage head where each pressure sensor is located, including: In step S141, the comprehensive proportional-integral-differential parameters of each partition are determined according to a preset proportional coefficient, a preset integral coefficient, a preset differential coefficient, the target comprehensive error of each partition, the steady-state error and the error change rate; In the disclosed embodiment, the comprehensive proportional-integral-differential parameters of each partition are calculated by the PID control algorithm according to the preset proportional coefficient, integral coefficient and differential coefficient, as well as the target comprehensive error, steady-state error and error change rate of each partition. The purpose of this step is to calculate the appropriate control amount adjustment value through the PID control strategy according to the current system state (i.e., the error and its change rate) to improve the control effect of the system.

[0067] In step S142, the product of the preset weight of each pressure sensor and the inverse of the single error pressure value corresponding to the pressure sensor is determined as the relative corrected error pressure value of each pressure sensor; The relative corrected error pressure value refers to the opposite number of the single error pressure value of the pressure sensor after weight adjustment, and is used to indicate the magnitude of the error pressure that needs to be corrected for the massage head where the sensor is located.

[0068] First, the preset weight of each pressure sensor is obtained, and then the weight is multiplied by the opposite number of the single error pressure value of the corresponding pressure sensor to obtain the relative corrected error pressure value. The purpose of this step is to calculate the error pressure value that needs to be corrected for each massage head based on the error size of the pressure sensor and its importance (i.e., weight).

[0069] For example: Assuming that the preset weight of a pressure sensor is w and the single error pressure value is P, the relative corrected error pressure value of the pressure sensor can be calculated by the following formula: Relative corrected error pressure value = -w × P. The negative sign here indicates that the correction direction is opposite to the error direction.

[0070] In step S143, the sum of the relative corrected error pressure value of each pressure sensor and the comprehensive proportional-integral-differential parameter corresponding to the partition is determined as the required massage intensity parameter of the massage head where each pressure sensor is located.

[0071] The relative corrected error pressure value of each pressure sensor is added to the comprehensive proportional integral differential parameter corresponding to the partition where the pressure sensor is located to obtain the required massage intensity parameter of the massage head where the pressure sensor is located. The purpose of this step is to comprehensively consider the control amount adjustment value calculated by the PID control strategy and the error pressure value that needs to be corrected for each massage head to determine the final required massage intensity parameter.

[0072] For example: Assuming that the comprehensive proportional integral differential parameter of the zone where a pressure sensor is located is PID, and the relative corrected error pressure value of the pressure sensor is ΔP, the required massage intensity parameter of the massage head can be calculated by the following formula: Required massage intensity parameter = PID + ΔP. This parameter will be used to guide the actual massage intensity adjustment of the massage head.

[0073] In a possible implementation, in step S141, determining the comprehensive proportional-integral-differential parameter of each partition according to the preset proportional coefficient, the preset integral coefficient, the preset differential coefficient, the target comprehensive error of each partition, the steady-state error and the error change rate includes determining the comprehensive proportional-integral-differential parameter of each partition by the following formula: Among them, K p is the preset proportionality coefficient, K i is the preset integral coefficient, K d is the preset differential coefficient, E(t) is the target comprehensive error of each partition, E(τ) is the steady-state error of each partition, u t The comprehensive proportional integral derivative parameters of each of the partitions.

[0074] In a possible implementation, in step S13, calculating the error change rate of the target comprehensive error of each partition to obtain the error change rate of each partition includes: In step S131, the difference between two adjacent target comprehensive errors of the partition is determined according to the historical target comprehensive errors corresponding to each partition; Among them, historical target comprehensive error refers to the comprehensive error value calculated for a specific partition at one or more time points in the past. These values ​​are used to evaluate the performance of the model or system.

[0075] Two adjacent times: refers to two time points or iteration cycles that are closely connected in the time series.

[0076] Difference: The absolute difference between two values, which here refers to the change between two consecutive target comprehensive errors.

[0077] By comparing the target comprehensive errors of the same partition at different time points (such as two consecutive iterations), the difference between them can be calculated. This difference reflects a direct measure of the change in error over time. For example: Assume that the target comprehensive error of partition A in the first iteration is 10, and the target comprehensive error in the second iteration is 8. The difference between the two consecutive iterations is 10-8=2, indicating that the error has decreased by 2.

[0078] In step S132, the difference values ​​and the ratio of the target comprehensive errors of the two adjacent target comprehensive errors are calculated to obtain the error change rate of each partition.

[0079] Among them, the error change rate refers to the ratio of the error change (difference) to a certain benchmark value (such as the previous target comprehensive error), which is used to measure the relative speed or degree of error change.

[0080] The error change rate can be obtained by dividing the difference between two consecutive target comprehensive errors by the previous target comprehensive error. This ratio not only reflects the absolute value change of the error, but also takes into account the relative size of the error change, which helps to more accurately evaluate the stability of system performance.

[0081] For example: Continuing with the above example, the target comprehensive error of partition A in the first iteration is 10, and the difference between the two consecutive iterations is 2, then the error change rate is 2 / 10=0.2 (or 20%). This means that from the first iteration to the second iteration, the target comprehensive error of partition A has decreased by 20% relatively.

[0082] By calculating the difference between two consecutive target comprehensive errors of each partition, and further calculating the ratio between these differences and the previous target comprehensive error, the error change rate of each partition is obtained. This helps to identify which partitions have rapidly decreasing or increasing errors.

[0083] The disclosed embodiment also provides a device for improving the massage force accuracy of a neck massager, see Figure 2 As shown, including: The acquisition module 210 is configured to acquire, in real time, neck pressure feedback data collected by pressure sensors disposed in different partitions of the neck massager during the massage process of the neck massager, wherein the pressure sensors are disposed in one-to-one correspondence with the massage heads of the massager; The first determination module 220 is configured to determine the target comprehensive error of the pressure sensors in each partition and the single error pressure value of each pressure sensor based on the neck pressure feedback data corresponding to the different pressure sensors and the preset pressure threshold range of the partition corresponding to each pressure sensor based on the cross attention mechanism; The calculation module 230 is configured to integrate the target comprehensive error of each partition according to the historical target comprehensive error corresponding to each partition to obtain the steady-state error of each partition, and calculate the error change rate of the target comprehensive error of each partition to obtain the error change rate of each partition; The second determination module 240 is configured to perform proportional differential integral operations according to the target comprehensive error, the steady-state error, the error change rate and the single error pressure value of each pressure sensor of each partition to determine the required massage intensity parameter for the massage head where each pressure sensor is located; The adjustment module 250 is configured to adjust the massage head pressure of the massage head where each pressure sensor in the neck massager is located based on the required massage intensity parameter.

[0084] In a possible implementation, the first determining module 220 is configured to: Determine, according to the neck pressure feedback data corresponding to each of the pressure sensors and the preset pressure threshold range of the partition corresponding to the pressure sensor, the error pressure value of each of the pressure sensors exceeding the preset pressure threshold range; Calculating a total error pressure value of each of the partitions according to the error pressure value of each of the pressure sensors in each of the partitions; Determine a target comprehensive error of the pressure sensor in each partition according to the error pressure value of each pressure sensor, the total error pressure value corresponding to the partition where the pressure sensor is located, and a preset weight of each pressure sensor, wherein the preset weight is negatively correlated with the distance of the pressure sensor in the corresponding partition relative to the pressure sensor at the center of the partition; The pressure sensors adjacent to each of the pressure sensors in the partition are used as first cross attention sensors, and the pressure sensors adjacent to each of the pressure sensors in the partition adjacent to the partition where the pressure sensors are located are used as second cross attention sensors; A single error pressure value of each of the pressure sensors is determined based on the error pressure value of each of the pressure sensors, the error pressure value of the corresponding first cross-attention sensor, and the error pressure value of the corresponding second cross-attention sensor.

[0085] In a possible implementation, the first determining module 220 is configured to: Calculating the product of the error pressure value of each pressure sensor and the preset weight corresponding to the pressure sensor to obtain a relative error pressure value corresponding to each pressure sensor; Calculate the ratio between the relative error pressure value corresponding to each of the pressure sensors and the total error pressure value corresponding to the partition where the pressure sensor is located, and determine the error ratio of each of the pressure sensors in each partition; Based on the preset k value in the ksigma algorithm, according to the standard deviation and mean of the error proportion value of each pressure sensor in each partition, calculate the K times standard deviation interval of the partition; The pressure sensor whose error proportion value in each partition is within the corresponding K times standard deviation interval is used as the target pressure sensor corresponding to the partition, and the sum of the error pressure values ​​of the target pressure sensors corresponding to the partition is calculated to obtain the target comprehensive error of the pressure sensor in each partition.

[0086] In a possible implementation, the first determining module 220 is configured to: Calculating the product of the error pressure value of each of the pressure sensors and the error pressure value of the corresponding first cross-attention sensor to obtain a first calculation parameter, and calculating the product of the error pressure value of each of the pressure sensors and the error pressure value of the corresponding second cross-attention sensor to obtain a second calculation parameter; Calculating the logarithmic function value of the first calculation parameter corresponding to each of the pressure sensors to obtain a third calculation parameter, and calculating the logarithmic function value of the second calculation parameter corresponding to each of the pressure sensors to obtain a fourth calculation parameter; Calculate the sum of the first calculation parameter and the corresponding second calculation parameter corresponding to each of the pressure sensors to obtain a fifth calculation parameter, and calculate the logarithmic function value of the fifth calculation parameter corresponding to each of the pressure sensors to obtain a sixth calculation parameter; Calculating the ratio of the third calculation parameter corresponding to each of the pressure sensors to the corresponding sixth calculation parameter to obtain a first cross-attention weight corresponding to each of the pressure sensors, and calculating the ratio of the fourth calculation parameter corresponding to each of the pressure sensors to the corresponding sixth calculation parameter to obtain a second cross-attention weight corresponding to each of the pressure sensors; The product of the first cross-attention weight corresponding to each pressure sensor and the error pressure value of the corresponding first cross-attention sensor is subtracted from the error pressure value corresponding to each pressure sensor, and the product of the second cross-attention weight corresponding to each pressure sensor and the error pressure value of the corresponding second cross-attention sensor is added to determine a single error pressure value of each pressure sensor.

[0087] In a possible implementation, the second determining module 240 is configured to: Determine the comprehensive proportional-integral-differential parameters of each of the partitions according to a preset proportional coefficient, a preset integral coefficient, a preset differential coefficient, the target comprehensive error of each of the partitions, the steady-state error and the error change rate; Determine the relative corrected error pressure value of each pressure sensor by multiplying the preset weight of each pressure sensor by the inverse of the single error pressure value corresponding to the pressure sensor; The sum of the relative corrected error pressure value of each pressure sensor and the comprehensive proportional-integral-differential parameter corresponding to the partition where the pressure sensor is located is determined as the required massage intensity parameter of the massage head where each pressure sensor is located.

[0088] In a possible implementation, the second determination module 240 is configured to determine the comprehensive proportional-integral-differential parameter of each partition by using the following formula: Among them, Kp is the preset proportionality coefficient, K i is the preset integral coefficient, K d is the preset differential coefficient, E(t) is the target comprehensive error of each partition, E(τ) is the steady-state error of each partition, u t The comprehensive proportional integral derivative parameters of each of the partitions.

[0089] In a possible implementation, the calculation module 230 is configured to: Determine the difference between two adjacent target comprehensive errors of the partition according to the historical target comprehensive errors corresponding to each of the partitions; The difference values ​​and the ratio of the target comprehensive errors of two adjacent target comprehensive errors are calculated traversally to obtain the error change rate of each partition.

[0090] The present disclosure also provides an electronic device, including: processor; a memory for storing processor-executable instructions; The processor is configured to execute the executable instructions stored in the memory to implement the method described in any one of the aforementioned embodiments.

[0091] The embodiments of the present disclosure further provide a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the method described in any one of the aforementioned embodiments are implemented.

[0092] Figure 3 The tobacco question-and-answer device 100 based on the multimodal large model shown includes: a processor 1001 and a memory 1003. Among them, the processor 1001 and the memory 1003 are connected, such as through a bus 1002. Optionally, the tobacco question-and-answer device 100 based on the multimodal large model may also include a communication component 1004, which can be used for data interaction between the device 100 and other devices, such as data transmission and / or data reception. It should be noted that in actual scheduling, the communication component 1004 is not limited to one, and the structure of the tobacco question-and-answer device 100 based on the multimodal large model does not constitute a limitation on the embodiments of the present application.

[0093] Processor 1001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. Processor 1001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0094] The bus 1002 may include a path to transmit information between the above components. The bus 1002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 1002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0095] The memory 1003 may be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), magnetic disk storage medium, other magnetic storage devices, or any other medium that can be used to carry or store program code and can be read by a computer, without limitation herein.

[0096] The memory 1003 is used to store program codes for executing the embodiments of the present disclosure, and the execution is controlled by the processor 1001. The processor 1001 is used to execute the program codes stored in the memory 1003 to implement the steps shown in the above-mentioned method embodiment for improving the massage force accuracy of the neck massager.

[0097] The embodiment of the present disclosure also provides a computer-readable storage medium having program code stored thereon. When the program code is executed by a processor, the steps and corresponding contents of the aforementioned method embodiment for improving the massage force accuracy of a neck massager can be implemented.

[0098] The preferred embodiments of the present disclosure are described in detail above in conjunction with the accompanying drawings; however, the present disclosure is not limited to the specific details in the above embodiments; within the technical concept of the present disclosure, various changes, modifications, substitutions and variations may be made to these embodiments, and these changes, modifications, substitutions and variations all fall within the protection scope of the present disclosure.

[0099] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any appropriate manner without contradiction, and they should also be regarded as the contents disclosed in this disclosure. In order to avoid unnecessary repetition, this disclosure will not further describe various possible combinations. The technical scope of this application is not limited to the contents in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A method for improving the massage intensity accuracy of a neck massager, characterized in that: include: During the massage process of the neck massager, neck pressure feedback data collected by pressure sensors arranged in different partitions of the neck massager are obtained in real time, and the pressure sensors are arranged in a one-to-one correspondence with the massage heads of the massager; Based on the cross attention mechanism, according to the neck pressure feedback data corresponding to the different pressure sensors and the preset pressure threshold range of the partitions corresponding to the pressure sensors, the target comprehensive error of the pressure sensors in each partition and the single error pressure value of each pressure sensor are determined; According to the historical target comprehensive errors corresponding to each of the partitions, the target comprehensive errors of each of the partitions are integrated to obtain the steady-state errors of each of the partitions, and the error change rate of the target comprehensive errors of each of the partitions is calculated to obtain the error change rate of each of the partitions; Perform proportional differential integral calculations based on the target comprehensive error, the steady-state error, the error change rate, and the single error pressure value of each pressure sensor of each partition to determine the required massage intensity parameter for the massage head where each pressure sensor is located; The massage head pressure of the massage head where each pressure sensor in the neck massager is located is adjusted based on the required massage intensity parameter.

2. The method for improving the massage intensity accuracy of a neck massager according to claim 1, characterized in that: The cross-attention mechanism is based on the neck pressure feedback data corresponding to different pressure sensors and the preset pressure threshold range of the pressure sensor corresponding to the partition, and determines the target comprehensive error of the pressure sensor in each partition and the single error pressure value of each pressure sensor, including: Determine, according to the neck pressure feedback data corresponding to each of the pressure sensors and the preset pressure threshold range of the partition corresponding to the pressure sensor, the error pressure value of each of the pressure sensors exceeding the preset pressure threshold range; Calculating a total error pressure value of each of the partitions according to the error pressure value of each of the pressure sensors in each of the partitions; Determine a target comprehensive error of the pressure sensor in each partition according to the error pressure value of each pressure sensor, the total error pressure value corresponding to the partition where the pressure sensor is located, and a preset weight of each pressure sensor, wherein the preset weight is negatively correlated with the distance of the pressure sensor in the corresponding partition relative to the pressure sensor at the center of the partition; The pressure sensors adjacent to each of the pressure sensors in the partition are used as first cross attention sensors, and the pressure sensors adjacent to each of the pressure sensors in the partition adjacent to the partition where the pressure sensors are located are used as second cross attention sensors; A single error pressure value of each of the pressure sensors is determined based on the error pressure value of each of the pressure sensors, the error pressure value of the corresponding first cross-attention sensor, and the error pressure value of the corresponding second cross-attention sensor.

3. The method for improving the massage intensity accuracy of a neck massager according to claim 2, characterized in that: Determining the target comprehensive error of the pressure sensor in each partition according to the error pressure value of each pressure sensor, the total error pressure value corresponding to the partition where the pressure sensor is located, and the preset weight of each pressure sensor includes: Calculating the product of the error pressure value of each pressure sensor and the preset weight corresponding to the pressure sensor to obtain a relative error pressure value corresponding to each pressure sensor; Calculate the ratio between the relative error pressure value corresponding to each of the pressure sensors and the total error pressure value corresponding to the partition where the pressure sensor is located, and determine the error ratio of each of the pressure sensors in each partition; Based on the preset k value in the ksigma algorithm, according to the standard deviation and mean of the error proportion value of each pressure sensor in each partition, calculate the K times standard deviation interval of the partition; The pressure sensor whose error proportion value in each partition is within the corresponding K times standard deviation interval is used as the target pressure sensor corresponding to the partition, and the sum of the error pressure values ​​of the target pressure sensors corresponding to the partition is calculated to obtain the target comprehensive error of the pressure sensor in each partition.

4. The method for improving the massage intensity accuracy of a neck massager according to claim 2, characterized in that: Determining a single error pressure value of each of the pressure sensors according to the error pressure value of each of the pressure sensors, the error pressure value of the corresponding first cross-attention sensor, and the error pressure value of the corresponding second cross-attention sensor includes: Calculating the product of the error pressure value of each of the pressure sensors and the error pressure value of the corresponding first cross-attention sensor to obtain a first calculation parameter, and calculating the product of the error pressure value of each of the pressure sensors and the error pressure value of the corresponding second cross-attention sensor to obtain a second calculation parameter; Calculating the logarithmic function value of the first calculation parameter corresponding to each of the pressure sensors to obtain a third calculation parameter, and calculating the logarithmic function value of the second calculation parameter corresponding to each of the pressure sensors to obtain a fourth calculation parameter; Calculate the sum of the first calculation parameter and the corresponding second calculation parameter corresponding to each of the pressure sensors to obtain a fifth calculation parameter, and calculate the logarithmic function value of the fifth calculation parameter corresponding to each of the pressure sensors to obtain a sixth calculation parameter; Calculating the ratio of the third calculation parameter corresponding to each of the pressure sensors to the corresponding sixth calculation parameter to obtain a first cross-attention weight corresponding to each of the pressure sensors, and calculating the ratio of the fourth calculation parameter corresponding to each of the pressure sensors to the corresponding sixth calculation parameter to obtain a second cross-attention weight corresponding to each of the pressure sensors; The product of the first cross-attention weight corresponding to each pressure sensor and the error pressure value of the corresponding first cross-attention sensor is subtracted from the error pressure value corresponding to each pressure sensor, and the product of the second cross-attention weight corresponding to each pressure sensor and the error pressure value of the corresponding second cross-attention sensor is added to determine a single error pressure value of each pressure sensor.

5. The method for improving the massage intensity accuracy of a neck massager according to claim 1, characterized in that: The method of performing proportional differential integral operation according to the target comprehensive error, the steady-state error, the error change rate and the single error pressure value of each pressure sensor of each partition to determine the required massage intensity parameter for the massage head where each pressure sensor is located includes: Determine the comprehensive proportional-integral-differential parameters of each of the partitions according to a preset proportional coefficient, a preset integral coefficient, a preset differential coefficient, the target comprehensive error of each of the partitions, the steady-state error and the error change rate; Determine the relative corrected error pressure value of each pressure sensor by multiplying the preset weight of each pressure sensor by the inverse of the single error pressure value corresponding to the pressure sensor; The sum of the relative corrected error pressure value of each pressure sensor and the comprehensive proportional-integral-differential parameter corresponding to the partition where the pressure sensor is located is determined as the required massage intensity parameter of the massage head where each pressure sensor is located.

6. The method for improving the massage force accuracy of a neck massager according to claim 5, characterized in that: Determining the comprehensive proportional-integral-differential parameters of each partition according to the preset proportional coefficient, the preset integral coefficient, the preset differential coefficient, the target comprehensive error of each partition, the steady-state error and the error change rate includes determining the comprehensive proportional-integral-differential parameters of each partition by the following formula: Among them, K p is the preset proportionality coefficient, K i is the preset integral coefficient, K d is the preset differential coefficient, E(t) is the target comprehensive error of each partition, E(τ) is the steady-state error of each partition, u t The comprehensive proportional integral derivative parameters of each of the partitions.

7. The method for improving the massage intensity accuracy of a neck massager according to any one of claims 1 to 6, characterized in that: Calculating the error change rate of the target comprehensive error of each partition to obtain the error change rate of each partition includes: Determine the difference between two adjacent target comprehensive errors of the partition according to the historical target comprehensive errors corresponding to each of the partitions; The difference values ​​and the ratio of the target comprehensive errors of two adjacent target comprehensive errors are calculated traversally to obtain the error change rate of each partition.

8. A device for improving the massage force accuracy of a neck massager, characterized in that: include: An acquisition module is configured to acquire, in real time during the massage process of the neck massager, neck pressure feedback data collected by pressure sensors arranged in different partitions of the neck massager, wherein the pressure sensors are arranged in a one-to-one correspondence with the massage heads of the massager; A first determination module is configured to determine, based on a cross-attention mechanism, a target comprehensive error of the pressure sensors in each partition and a single error pressure value of each pressure sensor according to neck pressure feedback data corresponding to different pressure sensors and a preset pressure threshold range of each partition corresponding to each pressure sensor; a calculation module configured to integrate the target comprehensive error of each partition according to the historical target comprehensive error corresponding to each partition to obtain the steady-state error of each partition, and to calculate the error change rate of the target comprehensive error of each partition to obtain the error change rate of each partition; The second determination module is configured to perform proportional differential integral operations according to the target comprehensive error, the steady-state error, the error change rate and the single error pressure value of each pressure sensor of each partition to determine the required massage intensity parameter for the massage head where each pressure sensor is located; The adjustment module is configured to adjust the massage head pressure of the massage head where each pressure sensor in the neck massager is located based on the required massage intensity parameter.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to execute the executable instructions stored in the memory to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images