Debugging device and method of AI human body sensor and AI human body sensor

Abstract

The invention provides a debugging device and method of an AI human body sensor and the AI human body sensor. Belongs to the technical field of intelligent sensing. The method comprises the following steps: constructing a special debugging device, and constructing a comprehensive debugging environment; the AI human body sensor is placed in a comprehensive debugging environment, preliminary hardware parameter setting is carried out, and initial hardware parameter configuration data is obtained; a physiological signal reproduction module and an interference environment simulation module are started, and original output data of the sensor in the comprehensive debugging environment are collected; performing preliminary analysis on the original output data to generate a preliminary performance deviation analysis report; through the automatic batch debugging process design, tedious operations such as manual repeated plugging and visual waveform judgment are reduced, the debugging efficiency is improved, and the consistency of sensor debugging on a large-scale production line can be ensured.

Description

TECHNICAL FIELD

[0001] The application provides an AI human body sensor debugging device and method and an AI human body sensor, and belongs to the technical field of intelligent sensing. BACKGROUND

[0002] In the field of AI human body sensors, such as PPG photoelectric sensors, MEMS accelerometers, thermoelectric infrared temperature sensors, and biological impedance electrodes, the debugging link before mass production and deployment is crucial, but current debugging techniques have many outstanding problems.

[0003] Traditional debugging is mostly carried out in static laboratories, relying only on standard signal sources or simple analog circuits, and cannot reproduce the nonlinearity, time variability, and individual differences of real human physiological signals. Key factors such as skin color affecting PPG signals and subcutaneous fat thickness affecting biological impedance are not considered, resulting in a significant decline in performance of factory parameters when actually worn.

[0004] At the same time, the lack of controllable, quantifiable, and composite injection mechanisms for typical interference sources makes it impossible to fully verify the sensor's anti-interference ability. Moreover, independent debugging of each hardware module parameters does not consider system-level coupling effects, resulting in suboptimal overall performance.

[0005] In addition, the debugging process relies on manual repeated plugging and unplugging and visual waveform judgment, which is not only time-consuming and inconsistent, but also difficult to meet the needs of large-scale production lines. Existing public information and manufacturer documents mostly describe vague concepts and do not start from the physical layer of the sensor hardware. Therefore, there is an urgent need for a new type of debugging method that focuses on the performance of sensor hardware. SUMMARY

[0006] The application provides an AI human body sensor debugging device, method, and AI human body sensor to solve the problems mentioned in the background art:

[0007] The application provides an AI human body sensor debugging method, which includes:

[0008] S1, a special debugging device is constructed, and a comprehensive debugging environment is built;

[0009] S2, the AI human body sensor is placed in the comprehensive debugging environment, preliminary hardware parameter settings are made, initial hardware parameter configuration data is obtained, the physiological signal reproduction module and the interference environment simulation module are started, and the original output data of the sensor in the comprehensive debugging environment is collected; the original output data is preliminarily analyzed, and a preliminary performance deviation analysis report is generated;

[0010] S3, according to the preliminary performance deviation analysis report, entering the hardware parameter collaborative debugging stage; through the intelligent algorithm, a hardware parameter collaborative adjustment scheme is generated, the sensor hardware parameters are collaboratively adjusted, and the sensor output data is collected again, and the collaboratively adjusted output data is generated;

[0011] S4, the collaboratively adjusted output data is deeply analyzed, and performance evaluation is performed, and a performance improvement evaluation report is generated; if the performance improvement does not reach the preset target, the hardware parameter collaborative adjustment scheme is further optimized according to the performance improvement evaluation report, and S3 and S4 are repeated until the sensor performance reaches the preset standard, and the final hardware parameter optimization data is generated;

[0012] S5, based on the final hardware parameter optimization data, an automatic batch debugging process design is performed on the AI human body sensor; an automatic batch debugging process scheme is generated, and the AI human body sensor on a large-scale production line is batch debugged, and batch debugging completion data is generated; the batch debugging completion data is sampled and rechecked, and a final debugging qualified report is generated.

[0013] The debugging device of the AI human body sensor provided by the application comprises a memory, a processor and a computer program stored on the memory and capable of running on the memory, and the processor executes the program to realize the debugging method of the AI human body sensor as described in any one of the above.

[0014] The AI human body sensor provided by the application has a computer program stored thereon, and the program is executed by the processor to realize the debugging method of the AI human body sensor as described in any one of the above.

[0015] The application has the following beneficial effects: by constructing a special debugging device to reproduce real human physiological signals and typical interference environments, the physiological authenticity and interference controllability of sensor debugging are greatly improved, so that the sensor can adapt to complex and changeable actual application scenes before leaving the factory. In the debugging process, a three-stage collaborative debugging process is adopted, the system-level coupling effect is fully considered, the key parameters of the sensor hardware are accurately optimized, the sensitivity, recognition accuracy and anti-interference ability of the sensor are significantly enhanced, and the overall performance suboptimal problem caused by parameter isolated optimization is effectively reduced. At the same time, the automatic batch debugging process design reduces the tedious operations such as manual repeated plugging, visual waveform judgment, etc., not only improves the debugging efficiency, but also ensures the consistency of sensor debugging on a large-scale production line. In addition, the method strictly focuses on sensor hardware debugging, avoids the problem of software covering up hardware defects, can collaboratively optimize the core indicators of the sensor from the hardware physical layer, and can ensure that the sensor has stable and reliable performance in actual application, thereby providing users with better use experience. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1The method steps of the present application are shown in the following. DETAILED DESCRIPTION

[0017] The preferred embodiments of the present application are described below in conjunction with the accompanying drawings, it should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and are not used to limit the present application.

[0018] An embodiment of the present application, as shown in Figure 1 The debugging method of the AI human sensor, the method comprises:

[0019] S1, a special debugging device is constructed, the device includes a physiological signal reproduction module and an interference environment simulation module; the physiological signal reproduction module reproduces the nonlinearity, time variability and individual difference characteristics of real human physiological signals through a multi-parameter adjustable physiological excitation source, generates simulated real human physiological signal data; the interference environment simulation module can controllably, quantitatively and complexly inject typical interference sources, including 50Hz power frequency electromagnetic interference, motion artifact, environmental light mutation, temperature and humidity fluctuation, to generate simulated interference environment data; based on the simulated real human physiological signal data and the simulated interference environment data, a comprehensive debugging environment is built;

[0020] S2, place the AI human sensor in the comprehensive debugging environment, perform preliminary hardware parameter setting, and obtain initial hardware parameter configuration data; start the physiological signal reproduction module and the interference environment simulation module according to the initial hardware parameter configuration data, and simultaneously collect the original output data of the sensor in the comprehensive debugging environment; preliminarily analyze the original output data, identify the performance deviation problems caused by debugging environment distortion or missing interference simulation, and generate a preliminary performance deviation analysis report;

[0021] S3, according to the preliminary performance deviation analysis report, enter the hardware parameter collaborative debugging stage; in this stage, instead of optimizing the parameters of each hardware module (pre-amplification, filtering, ADC) independently, the system-level coupling effect is considered, the sensitivity, accuracy and anti-interference performance are taken as the optimization targets, and the hardware parameter collaborative adjustment scheme is generated through intelligent algorithm; according to the hardware parameter collaborative adjustment scheme, the hardware parameters of the sensor are collaboratively adjusted, and the sensor output data is collected again to generate the collaborative adjustment output data;

[0022] S4, perform deep analysis on the collaborative adjustment output data, evaluate the performance improvement of the sensor in terms of sensitivity, identification accuracy and anti-interference ability, and generate a performance improvement evaluation report; if the performance improvement does not reach the preset target, further optimize the hardware parameter collaborative adjustment scheme according to the performance improvement evaluation report, and repeat steps S3 and S4 until the performance of the sensor reaches the preset standard, and generate the final hardware parameter optimization data;

[0023] S5, based on the final hardware parameter optimization data, the AI human body sensor is subjected to automatic batch debugging process design; the process is cooperated with the special debugging device and the automatic control software, automatic loading of the sensor hardware parameters, automatic switching of the debugging environment, automatic evaluation and feedback adjustment of the performance are realized, and an automatic batch debugging process scheme is generated; according to the automatic batch debugging process scheme, the AI human body sensor on a large-scale production line is subjected to batch debugging, and batch debugging completion data is generated; the batch debugging completion data is subjected to sampling inspection and review, it is ensured that each batch of sensors reaches the preset performance standard, and a final debugging qualified report is generated.

[0024] The working principle and effects of the above technical solutions are as follows:

[0025] The debugging method can accurately reproduce physiological signal characteristics and complex interference by constructing a comprehensive debugging environment close to the actual scene, greatly improves the debugging accuracy, avoids performance misjudgment caused by environment simulation distortion; the hardware parameter cooperative debugging mode breaks the limitation of isolated optimization, fully considers the coupling effect between modules, enhances the parameter adaptability, reduces the time cost of repeated trial and error, significantly improves the sensor sensitivity, precision and anti-interference performance and other core performances; the performance is continuously checked through the iterative optimization mechanism, it is ensured that each sensor can reach the preset standard, and the product qualification rate is effectively improved; the automatic batch debugging process is matched with the special device and software, the manual debugging cost is greatly reduced, the human operation error is reduced, and the production line debugging efficiency is significantly improved; the sampling inspection and review link after batch debugging can accurately intercept unqualified products, avoid the flow of defective products into the market, and enhance the product reliability.

[0026] In an embodiment of the present application, the S1 comprises:

[0027] S11, based on the application scene (such as medical monitoring, sports fitness) of the AI human body sensor, the physiological signal type (such as heart rate, blood oxygen, muscle electricity, brain electricity, etc.) and the interference scene to be reproduced are determined, and the real physiological signal data of different ages and physical conditions are obtained from the existing information library, a physiological signal feature database containing 100,000+ samples is constructed, and the quantization parameter range of the typical interference (such as 50Hz power frequency interference intensity 0-10V / m) is sorted out;

[0028] S12, a special debugging device integrating a physiological signal reproduction module, an interference environment simulation module, a signal fusion module and a data acquisition module is built, wherein the physiological signal reproduction module is configured with a 0-5V adjustable excitation source and a signal waveform synthesis unit, the interference environment simulation module is equipped with an electromagnetic interference generator, a motion artifact simulation mechanical arm, an environment light controller and a temperature and humidity adjusting box;

[0029] S13, zero point calibration and gain calibration are performed on the physiological signal reproduction module, non-linear time-varying signals containing circadian rhythm changes and motion state switching are generated through a waveform synthesis unit based on a physiological signal feature database, individual difference parameters (such as the heart rate fluctuation range of the elderly) are superimposed, and a multi-scene simulated real human physiological signal data set is generated;

[0030] S14, precision calibration is performed on each interference unit (such as an ambient light mutation amplitude error of less than or equal to 5%), and single interference independent output or multiple interference composite output according to a preset ratio (such as a power frequency interference + motion artifact composite scene) is supported, and a simulated interference environment data set covering 20+ typical interference scenes is generated;

[0031] S15, millisecond-level synchronous output of physiological signals and interference signals is performed through a signal fusion module, and a comprehensive debugging environment is built; 3 standard calibration sensors are selected to access the environment, output data and theoretical values are compared, and it is ensured that the signal distortion degree is less than or equal to 3%, and the environment parameters are verified after the environment effectiveness is locked.

[0032] The working principle and effects of the above technical solutions are as follows:

[0033] By constructing a physiological signal feature database of 100,000+ samples, covering real data of different ages and physical conditions, and clearly defining the quantitative parameter range of typical interference, the authenticity of physiological signal and interference simulation is greatly improved, and the one-sided problem of simulation scene caused by single sample or ambiguous interference parameters is avoided. The special debugging device built integrates multiple modules, which eliminates the trouble of scattered collocation of multiple devices, enhances the integration and convenience of debugging, and reduces the risk of device compatibility failure; the precise calibration of physiological signals and interference units makes the simulation signal distortion degree controllable, such as an ambient light mutation amplitude error of less than or equal to 5%, which significantly improves the accuracy of simulation data and avoids the inaccuracy of debugging benchmark caused by module deviation. Single or composite interference output is supported, 20+ typical scene data is generated, and the comprehensive debugging environment built is closer to the actual application scene, which enhances the debugging pertinence; the environment effectiveness is verified through standard sensors, and it is ensured that the signal distortion degree is less than or equal to 3%, which prevents misjudgment caused by unreliable environment in subsequent debugging.

[0034] In an embodiment of the present application, the S13 comprises:

[0035] According to the specification of the physiological signal reproduction module, a zero point and gain calibration process is developed, a standard signal source is used for calibration operation, and after completion, a calibration qualified confirmation report is generated;

[0036] Based on the calibration qualified confirmation report, the module is started, the physiological signal feature database is called, the basic amplitude and frequency parameters of the related signals (such as heart rate and blood oxygen) are extracted, and a basic physiological signal waveform set is generated;

[0037] Input the basic physiological signal waveform set into the waveform synthesis unit, load the circadian rhythm time curve and the motion state switching threshold, and superimpose to generate a nonlinear time-varying signal data set;

[0038] Retrieve the individual difference parameter library (including age, body constitution correlation coefficient) from the physiological signal feature database, match the time-varying signal one by one and superimpose the parameters to generate signal data with individual characteristics;

[0039] The signal data with individual characteristics is classified and labeled according to the scene, and after removing abnormal data, a multi-scene simulated real human physiological signal data set is generated.

[0040] The working principle and effect of the above technical solution are:

[0041] By calibrating the standard signal source according to the process and generating a qualified report, the output accuracy of the physiological signal reproduction module is ensured from the source, and the distortion problem of the basic signal caused by module zero drift and gain deviation is avoided, which lays a solid foundation for subsequent signal generation. The core parameters such as heart rate and blood oxygen are extracted from the database to generate basic waveforms, which makes the signal fit the real physiological characteristics and reduces the blindness of generating signals in the air, improving the authenticity of the basic signal. Load the circadian rhythm and motion state threshold and superimpose them to make the signal show the real nonlinear time-varying characteristics of the human body, breaking the limitations of fixed waveforms. Then match the individual parameters such as age and constitution to make the signal have exclusive characteristics, enhance the adaptability to different groups of people, and avoid the problem that a single signal cannot simulate multiple physiological states. Finally, classify and label the abnormal data to further improve the reliability of the data set and ensure that the output multi-scene simulation signal can accurately match the actual debugging requirements.

[0042] In one embodiment of the present application, the S2 comprises:

[0043] S21, appearance inspection and power-on self-test are performed on the AI human sensor to be debugged, and the residual interference source on the surface of the sensor is removed; the sensor is fixed to a standard test station in a comprehensive debugging environment, and the distance between the sensor and the signal transmitting end is calibrated by a laser positioner (error ≤ 2mm);

[0044] S22, load the sensor factory default parameters through the debugging host computer to complete the basic parameter setting, the basic parameters include sampling frequency (500Hz-1kHz), preamplifier gain (100-1000 times), low-pass filter cutoff frequency (50-100Hz); then generate an initial hardware parameter configuration data set according to the historical debugging basic optimization parameters matched with the sensor model;

[0045] S23, according to the initial hardware parameter configuration data, start three typical scenes of the comprehensive debugging environment in turn, the three typical scenes include static physiological signal + no interference, dynamic physiological signal + single interference and dynamic physiological signal + composite interference, 20 minutes of original output data are collected continuously in each scene, environment parameters are recorded synchronously, and a multi-scene initial output data set is generated;

[0046] S24, denoising (excluding abnormal values, smoothing processing) is performed on the multi-scene initial output data set, time domain (peak value, period) and frequency domain (power spectrum density) features are extracted, standard templates in a physiological signal feature database are compared one by one, signal core indexes are calculated, and the core indexes include similarity and signal-to-noise ratio.

[0047] S25, based on the index comparison result, a deviation type (such as signal-to-noise ratio < 30 dB under power frequency interference, peak value recognition error > 5% under dynamic signal) is identified, a root cause (such as unreasonable filter cutoff frequency, insufficient amplification gain) is located, and a preliminary performance deviation analysis report containing deviation details and root cause analysis is generated after classification and summarization.

[0048] The working principle and effects of the above technical solutions are as follows:

[0049] Through appearance inspection and power-on self-checking, residual interference is removed, the distance error between the sensor and the transmitting end is controlled within 2mm through laser positioning, installation deviation and initial state abnormality interference on the test are excluded from the source, and the stability of subsequent data acquisition is improved; after loading the factory default parameters, the initial configuration is generated by matching the historical optimization parameters, which is more suitable for the actual characteristics of the sensor than using only the factory parameters, reduces the invalid debugging caused by unreasonable initial parameters, and improves the adaptability of parameter configuration; three typical scenes are started in turn to collect 20 minutes of data and record environment parameters, cover the core scenes from static no interference to dynamic composite interference, avoid the situation of missing performance problems due to single test scene, and enhance the representativeness of the initial output data; after denoising the data, the time domain and frequency domain features are extracted and compared with the standard templates, the indexes such as similarity and signal-to-noise ratio are accurately calculated, the subjective error of manual analysis is reduced, and the accuracy of data interpretation is improved; and the deviation analysis report finally generated clearly indicates the deviation type and root cause, such as the root cause of insufficient signal-to-noise ratio under power frequency interference is unreasonable filter parameters, avoids subsequent debugging blind trial and error, and provides a clear guide for parameter optimization.

[0050] In an embodiment of the application, the S3 comprises:

[0051] S31, analyze the preliminary performance deviation analysis report, and decompose the core deviation indexes (sensitivity, accuracy, and anti-interference performance) into associated parameters of each hardware module (for example, the preamplifier gain affects the sensitivity, and the filtering parameter affects the anti-interference performance) ; in combination with the sensor hardware specification book, the adjustable range of each parameter (for example, the ADC sampling bit number is 12-16 bits) is defined, and a parameter and index correlation model is established;

[0052] S32, an improved particle swarm optimization algorithm (introducing an inertia weight self-adaptive adjustment mechanism) is selected, the parameter and index correlation model is used as an algorithm constraint condition, the sensitivity is greater than or equal to 95%, the accuracy error is less than or equal to 2%, and the anti-interference signal-to-noise ratio is greater than or equal to 40 dB are used as optimization objectives, and a hardware parameter collaborative optimization model is constructed.

[0053] S33, input the preliminary deviation data into the optimization model, and generate 10 groups of hardware parameter collaborative adjustment schemes through 100 times of iteration calculation; the analytic hierarchy process is used to evaluate the schemes (weight: anti-interference performance 40%, accuracy 35%, and sensitivity 25%), and the three groups of candidate schemes with the highest comprehensive scores are selected.

[0054] S34, three sensors of the same type are selected, the three groups of candidate schemes are used for parameter adjustment, and output data is collected in a composite interference scene; the performance indexes of the three groups of data are compared, and the hardware parameter collaborative adjustment scheme with the optimal comprehensive performance is determined.

[0055] S35, the parameters of the optimal scheme are synchronously loaded into a sensor to be debugged through the debugging host computer, the comprehensive debugging environment is restarted, and full-scene output data is collected, and the output data set after the collaborative adjustment is generated.

[0056] The working principle and effects of the above technical scheme are as follows:

[0057] The core indexes are decomposed into hardware parameters and correlation is established in the deviation report, the adjustable range of each parameter is defined in combination with the specification book, the problem that isolated parameter adjustment ignores module coupling is effectively avoided, parameter optimization is more targeted, the improved particle swarm algorithm with inertia weight self-adaptation is used, the model is constructed with the clear target of sensitivity greater than or equal to 95% and parameter constraints, the optimization process is more accurate, invalid iteration is reduced, and the optimization efficiency is greatly improved; after 10 groups of schemes are generated, the three groups of candidate schemes are selected by using the analytic hierarchy process according to the weights of anti-interference performance, accuracy and the like, the one-sidedness of a single scheme is avoided, and the reliability of the scheme is enhanced; the optimal scheme is not only theoretically feasible, but also more suitable for actual use requirements through the comparison of the actual measurement of the three sensors of the same type in the composite interference scene, and the practicability of the scheme is improved; finally, the optimal parameters are loaded and full-scene data is collected, it is ensured that the adjusted sensor meets the requirements in various scenes, and the problem that the local performance is qualified but the overall performance is insufficient is avoided.

[0058] In an embodiment of the present application, the S33 comprises:

[0059] Call preliminary performance deviation data and parameter-index correlation model, determine the input boundary of the optimization model, configure the initial value and adjustment threshold of the inertia weight of 100 iterations, and generate the model input configuration table;

[0060] Import the input configuration table into the hardware parameter collaborative optimization model, start iteration calculation, output intermediate parameter combination every 20 iterations, and finally generate 10 complete hardware parameter collaborative adjustment schemes;

[0061] Determine the analytic hierarchy process evaluation dimension combined with the sensor application priority, clear the weight allocation rule of 40% anti-interference, 35% accuracy and 25% sensitivity, and build a scheme evaluation framework;

[0062] Substitute the 10 adjustment schemes into the evaluation framework, calculate the single score and comprehensive weighted score of each group in the anti-interference, accuracy and sensitivity dimensions, and generate a scheme score ranking table;

[0063] According to the score ranking table, select the top 3 schemes, check whether the parameters of each scheme meet the hardware specification book limit, and finally determine 3 candidate adjustment schemes.

[0064] The working principle and effect of the above technical scheme are:

[0065] The deviation data and the correlation model are called to determine the input boundary, the iteration parameters are configured to generate the input configuration table, the model input is avoided to be chaotic or the parameter setting is avoided to be blind, the optimization calculation has a clear basis, and the input accuracy is significantly improved;After importing the configuration table into the model, 100 iterations are performed and intermediate results are output every 20 times, and finally 10 schemes are generated, which can monitor the iteration process in real time and troubleshoot exceptions in time, and can also ensure the diversity of the schemes and avoid the limitations of a single scheme;The weight of the analytic hierarchy process is determined combined with the application priority, and the evaluation framework is built with clear rules such as 40% anti-interference, 35% accuracy, etc., so that the scheme evaluation is closely related to the actual demand, the selected scheme deviates from the core demand due to the imbalance of the weight, and the evaluation is targeted;10 groups of schemes are substituted into the framework for quantitative calculation and scoring, and the data support comparison is used to reduce the subjective error of manual screening and improve the fairness and accuracy of scheme screening;After screening the top 3, check whether the parameters meet the hardware specification book, avoid the selected scheme from being unable to land due to parameter out-of-range, and ensure the feasibility of the candidate scheme.

[0066] In one embodiment of the present application, the S34 comprises:

[0067] Select three AI human body sensors of the same type with consistent performance (factory error ≤3%) for appearance cleaning and power self-checking, mark the abnormal devices as test samples 1-3 after removing the abnormal devices, and generate a test sample confirmation list;

[0068] According to the parameter configuration requirements of the 3 candidate schemes, the upper computer is debugged to load the corresponding parameters for the 3 test samples in turn, and after each loading is completed, the parameter writing success rate (≥99%) is verified, and a parameter configuration confirmation table is generated;

[0069] Start the composite interference scene of the comprehensive debugging environment (preset power frequency interference + motion artifact + temperature and humidity fluctuation combination), control the stable output of the scene parameters, synchronously collect the output data of the 3 test samples for 30 minutes, and generate a grouped test data set;

[0070] Extract the sensitivity, accuracy and anti-interference signal-to-noise ratio core indicators in the grouped test data set, calculate the comprehensive performance score of each group of schemes according to the preset weight, and generate a multi-scheme performance comparison and analysis table;

[0071] According to the comparison and analysis table and the actual application scene requirement, the scheme with the highest comprehensive score is selected as the optimal hardware parameter coordination adjustment scheme.

[0072] The working principle and effect of the above technical scheme are:

[0073] Select the same type of sensors with an out-of-factory error of ≤3%, mark and generate a confirmation list after cleaning and self-checking, effectively eliminate the interference of sample performance difference on the test result, lay a fair foundation for scheme comparison, and improve the reliability of subsequent test; after loading the parameters according to the candidate scheme, verify that the writing success rate is ≥99% and generate a confirmation table, avoid test failure caused by parameter writing error, and reduce the invalid workload of repeated debugging; start the composite interference scene of power frequency interference, motion artifact and the like, and synchronously collect data for 30 minutes, which is closer to the actual use of the complex scene, and can more accurately test the comprehensive adaptability of the scheme than single interference test, and the generated grouped data set makes the scheme performance comparison more intuitive. Extract the core indicators and calculate the comprehensive score according to the preset weight, replace the subjective judgment with quantitative data, and generate a comparison and analysis table that clearly presents the advantages and disadvantages of each scheme, greatly reducing the human error in scheme selection; combined with the actual application requirement, the anti-interference performance of the selected scheme is guaranteed, and the scheme with the highest comprehensive score is selected, which avoids selecting a theoretically optimal but actually poorly adapted scheme, and ensures that the finally determined optimal scheme can truly fit the use scene of the sensor.

[0074] In one embodiment of the present application, the S4 comprises:

[0075] S41, according to the sensor application scene requirement, refine the performance evaluation standard (such as medical grade sensor precision error ≤1%, motion scene anti-interference signal-to-noise ratio ≥45dB); take the output data of the standard calibrated sensor as the reference, establish the evaluation error allowable range (such as ≤2%);

[0076] S42, multi-dimension analysis is made on the output data set after the coordinated adjustment, sensitivity is quantified by using signal response time (≤100 ms is qualified), physiological feature recognition accuracy is calculated by using a confusion matrix, anti-interference capability is evaluated by using an interference suppression ratio (≥45 dB is qualified), reliability is verified by using long-term stability test (output fluctuation is ≤1% for 2 hours continuously), and a performance improvement evaluation report is generated;

[0077] S43, the evaluation report is compared with the refined standard, if there is an unqualified index (such as motion artifact + power frequency interference accuracy 89% < 95%), the influence of each hardware parameter on the index is tested one by one by using the control variable method, and the root cause (such as the filter parameter not matching the motion artifact frequency) is located.

[0078] S44, based on the root cause positioning result, the target weight of the particle swarm optimization algorithm is modified (the motion artifact suppression weight is increased to 50%), a hardware parameter coordinated adjustment scheme is regenerated, and a modified parameter set is generated by fine-tuning key parameters (such as filter cutoff frequency).

[0079] S45, the parameter loading and data acquisition of S35 and the analysis and evaluation process of the present step are repeated until all performance indicators meet the refined standard, the final hardware parameter is locked, and the final hardware parameter optimization data is generated.

[0080] The working principle and effects of the above technical solution are as follows:

[0081] According to the scene refined evaluation standard, for example, the medical level precision error is ≤1%, and the motion scene anti-interference signal-to-noise ratio is ≥45 dB, the standard is used to calibrate the sensor data as a reference, the performance evaluation is more in line with the actual use requirements, and the evaluation pertinence is improved; multi-dimension analysis is made on the adjusted data, the sensitivity is quantified by using the response time, the accuracy is calculated by using the confusion matrix, and the anti-interference capability is evaluated by using the interference suppression ratio, and the 2-hour stability test is added, so that the generated evaluation report can fully present the sensor performance, the performance misjudgment caused by single index evaluation is avoided, and the evaluation integrity is enhanced; when the unqualified index is found, the influence of each parameter is tested one by one by using the control variable method, the root cause such as the filter parameter not matching the motion artifact frequency is accurately located, the invalid trial and error caused by blind parameter adjustment is avoided, and the optimization time consumption is greatly reduced; based on the root cause, the weight of the algorithm is modified, the scheme is regenerated, and the key parameters are fine-tuned, so that the optimization is more focused on the problem core, and the parameter optimization precision is improved; the parameter loading, collection and evaluation process are repeated until the final parameter is locked, and it is ensured that all performance indicators of the sensor meet the scene standard, and the use failure caused by “partial compliance” is avoided.

[0082] In an embodiment of the present application, the S44 comprises:

[0083] Extract the key influencing factors in the root cause positioning results (such as motion artifact frequency and filter parameter mismatch), determine the core direction of algorithm correction, and generate an algorithm adjustment requirement list;

[0084] Based on the requirement list, adjust the target weight of the improved particle swarm optimization algorithm (such as the motion artifact suppression weight to 50%), reset the iteration constraint condition, and run the algorithm to generate a new hardware parameter collaborative adjustment scheme;

[0085] From the new adjustment scheme, screen the key parameters (such as filter cutoff frequency, amplification gain), determine the parameter fine-tuning interval (such as filter cutoff frequency ± 5Hz) based on the root cause analysis results, and generate a parameter fine-tuning scheme;

[0086] According to the fine-tuning scheme, adjust the key parameters one by one, record the parameter values before and after adjustment, check whether the parameters meet the requirements of the hardware specification book, and generate a preliminary draft of the corrected parameter set;

[0087] Simulate and test the preliminary draft of the parameter set to verify the improvement effect on the unqualified indicators (such as precision), and lock the corrected parameter set after confirming its effectiveness.

[0088] The working principle and effect of the above technical solution are:

[0089] Extracting the key factors of root cause positioning and generating an algorithm adjustment requirement list can accurately grasp the core of the unqualified problem, such as the mismatch between motion artifact frequency and filter parameter, avoiding the deviation of algorithm correction direction, allowing the optimization to have a clear target, and improving the adjustment specificity; based on the demand adjustment algorithm target weight, such as increasing the motion artifact suppression weight to 50%, and then resetting the constraints to generate a new scheme, allowing the optimization to focus on the short board indicators, reducing the invalid iteration of irrelevant parameters, and greatly improving the efficiency and accuracy of scheme generation; from the new scheme, screen the filter cutoff frequency and other key parameters, determine the fine-tuning interval such as ± 5Hz based on root cause analysis, avoid the complexity and risk of full parameter adjustment, accurately attack the problem point, and enhance the controllability of parameter adjustment; after adjustment, check whether the parameters meet the hardware specification book, generate a preliminary draft, and then do simulation test to verify the improvement effect on the unqualified indicators, ensure that the parameters are effective and feasible, avoid repeated adjustment due to parameter out-of-range or insufficient effect, and reduce the trial and error cost.

[0090] One embodiment of the present application, the S5, comprises:

[0091] S51, based on the final hardware parameter optimization data, comb the core links of batch debugging (parameter configuration, scene debugging, performance detection, deviation correction); write an automatic debugging script through Python, integrate sensor batch communication interface (RS485 bus), environment scene switching instruction, performance index calculation function, and generate an automatic batch debugging process scheme

[0092] S52, select 10 sensors for small batch debugging, verify the success rate of script parameter loading (≥99%), scene switching stability; according to the common problems of the production line, expand the debugging scene library (add extreme scenes such as low temperature environment and high humidity environment), optimize the script fault tolerance mechanism;

[0093] S53, deploy the optimized automatic batch debugging process scheme to the production line debugging system, realize the automatic connection of sensor feeding, debugging and discharging through the linkage interface of the special debugging device and the production line; batch parameter configuration, multi-scene debugging for large-scale production line sensors, real-time recording of each sensor debugging data, and generation of batch debugging completion dataset;

[0094] S54, adopt the hierarchical sampling mode of preliminary inspection + re-inspection, the preliminary inspection extracts 5% samples from the batch data, and tests the core indicators in the production line self-inspection environment; the re-inspection extracts 30% from the preliminary inspection qualified samples, and performs full-scene repeated test in the independent third-party debugging environment to verify the performance consistency;

[0095] S55, summarize the batch debugging data and sampling results, generate the final debugging qualified report containing batch information, performance statistics and sampling report; and enter the debugging data of each sensor into the blockchain system.

[0096] The working principle and effect of the above technical scheme are:

[0097] Based on the optimal parameter, the core link is combed and the automatic script is written, the RS485 batch communication, scene switching and other functions are integrated to generate the process scheme, the manual repetitive operation is replaced by automation, the batch debugging efficiency is greatly improved, the error of manual parameter configuration is reduced, and the debugging failure caused by manual operation failure is avoided; The script success rate (≥99%) is verified by small batch debugging, the fault tolerance mechanism is optimized, the low temperature scene library is expanded, the script vulnerabilities or scene missing during large-scale deployment are avoided, and the process stability is enhanced; After deployment to the production line, realize the automatic connection of feeding, debugging and discharging, complete batch parameter configuration and multi-scene debugging, and record the data in real time, reduce the labor cost of the production line, make the adaptability of large-scale production more sufficient; The hierarchical sampling mode of preliminary inspection + re-inspection is very practical, the preliminary inspection is rapid screening, the re-inspection is full-scene verification in the third-party environment, the performance consistency is ensured, the single sampling mode is avoided, the defective products are not missed, and the product qualified rate is significantly improved; After generating the qualified report by summarizing the data, the data of each sensor is recorded into the blockchain traceability, the parameter, performance and other information can be checked throughout the life cycle, the subsequent fault cannot be traced, and the credibility of product quality can be enhanced.

[0098] An embodiment of the present application is an AI human sensor debugging device, characterized by comprising a memory, a processor, and a computer program stored on the memory and executable on the memory, the processor executing the program to implement the AI human sensor debugging method of any one of the above.

[0099] An embodiment of the present application is an AI human sensor, characterized by having a computer program stored thereon, the program being executable by a processor to implement the AI human sensor debugging method of any one of the above.

[0100] Obviously, various modifications and changes can be made to the present application by those skilled in the art without departing from the spirit and scope of the present application. Thus, it is intended that the present application cover the modifications and changes as they come within the scope of the claims and their equivalents.

Claims

1. A method for debugging an AI human body sensor, characterized in that, The method includes: S1. Construct a dedicated debugging device and set up a comprehensive debugging environment; S2. Place the AI ​​human body sensor in the comprehensive debugging environment, perform preliminary hardware parameter settings, and obtain initial hardware parameter configuration data; and start the physiological signal reproduction module and interference environment simulation module, while collecting the sensor's raw output data in the comprehensive debugging environment; perform preliminary analysis on the raw output data and generate a preliminary performance deviation analysis report; S3. Based on the preliminary performance deviation analysis report, enter the hardware parameter collaborative debugging stage; generate a hardware parameter collaborative adjustment scheme through intelligent algorithms, collaboratively adjust the sensor hardware parameters, and collect sensor output data again to generate collaboratively adjusted output data; S4. Perform in-depth analysis on the output data after collaborative adjustment and conduct performance evaluation to generate a performance improvement evaluation report. If the performance improvement does not meet the preset target, further optimize the hardware parameter collaborative adjustment scheme according to the performance improvement evaluation report, and repeat S3 and S4 until the sensor performance reaches the preset standard and generate the final hardware parameter optimization data. S5. Based on the final hardware parameter optimization data, design an automated batch debugging process for AI human body sensors; generate an automated batch debugging process plan, perform batch debugging of AI human body sensors on large-scale production lines, and generate batch debugging completion data; conduct random checks and verifications on the batch debugging completion data, and generate a final debugging qualification report.

2. The debugging method for the AI ​​human body sensor according to claim 1, characterized in that, S1 includes: S11. Based on the application scenarios of AI human body sensors, determine the types of physiological signals that need to be reproduced and the interference scenarios, and obtain real physiological signal data of different age groups and physical conditions from the existing information database to construct a physiological signal feature database, while sorting out the quantitative parameter range of typical interferences. S12. Build a dedicated debugging device that integrates a physiological signal reproduction module, an interference environment simulation module, a signal fusion module, and a data acquisition module; S13. Perform zero-point calibration and gain calibration on the physiological signal reproduction module. Based on the physiological signal feature database, generate nonlinear time-varying signals containing diurnal rhythm changes and motion state switching through the waveform synthesis unit. Superimpose individual difference parameters to generate a multi-scenario simulated real human physiological signal dataset. S14. Perform precision calibration on each interference unit, and support independent output of a single interference or composite output of multiple interferences according to a preset ratio to generate a simulated interference environment dataset. S15. The physiological signal and interference signal are synchronously output at the millisecond level through the signal fusion module to build a comprehensive debugging environment; three standard calibration sensors are selected to be connected to the environment, the output data is collected and compared with the theoretical value, and the environmental parameters are locked to verify the effectiveness of the environment.

3. The debugging method for the AI ​​human body sensor according to claim 2, characterized in that, S13 includes: Based on the specifications of the physiological signal reproduction module, a zero-point and gain calibration process was developed. A standard signal source was used for calibration, and a calibration confirmation report was generated upon completion. Based on the calibration qualification confirmation report, the module is started, and the physiological signal feature database is called to extract the basic amplitude and frequency parameters of relevant signals and generate a basic physiological signal waveform set. The basic physiological signal waveform set is input into the waveform synthesis unit, loaded with the diurnal rhythm time sequence curve and the motion state switching threshold, and superimposed to generate a nonlinear time-varying signal dataset. The individual difference parameter library is retrieved from the physiological signal feature database, and time-varying signals are matched one by one and parameters are superimposed to generate signal data with individual characteristics. Signal data with individual characteristics are classified and labeled according to scenarios. After removing abnormal data, they are integrated to generate a dataset of real human physiological signals simulating multiple scenarios.

4. The debugging method for the AI ​​human body sensor according to claim 1, characterized in that, S2 includes: S21. Perform visual inspection and power-on self-test on the AI ​​human body sensor to be debugged, and remove any residual interference sources on the sensor surface; fix the sensor in the standard test station of the integrated debugging environment, and calibrate the distance between the sensor and the signal transmitter using a laser positioning device; S22. Load the sensor's factory default parameters into the host computer to complete the basic parameter settings, and then match the basic optimization parameters from the historical debugging process according to the sensor model to generate the initial hardware parameter configuration dataset. S23. Based on the initial hardware parameter configuration data, start the three typical scenarios of the integrated debugging environment in sequence, continuously collect raw output data for 20 minutes for each scenario, record environmental parameters synchronously, and generate initial output datasets for multiple scenarios. S24. Denoise the initial output dataset of multiple scenarios, extract time-domain and frequency-domain features, compare it one by one with the standard template in the physiological signal feature database, and calculate the core indicators of the signal. S25. Based on the index comparison results, identify the types of deviations, locate the root causes, and generate a preliminary performance deviation analysis report after classification and summarization.

5. The debugging method for the AI ​​human body sensor according to claim 1, characterized in that, The S3 includes: S31. Analyze the preliminary performance deviation analysis report, break down the core deviation indicators into the associated parameters of each hardware module; combine the sensor hardware specifications to define the adjustable range of each parameter and establish a parameter and indicator correlation model. S32. Select the improved particle swarm optimization algorithm, use the parameter and index correlation model as the algorithm constraint, set the sensitivity ≥95%, accuracy error ≤2%, and anti-interference signal-to-noise ratio ≥40dB as the optimization objectives, and construct a hardware parameter collaborative optimization model. S33. Input the initial deviation data into the optimization model, and generate 10 sets of hardware parameter coordinated adjustment schemes through 100 iterations; use the analytic hierarchy process to evaluate the schemes and select the 3 candidate schemes with the highest comprehensive scores. S34. Select three sensors of the same model and adjust their parameters using three candidate schemes respectively. Collect and output data under a complex interference scenario. Compare the performance indicators of the three sets of data and determine the hardware parameter collaborative adjustment scheme with the best overall performance. S35. By debugging the host computer, the parameters of the optimal solution are synchronously loaded into the sensor to be debugged, the integrated debugging environment is restarted and the output data of the whole scene is collected to generate the output dataset after collaborative adjustment.

6. The debugging method for the AI ​​human body sensor according to claim 5, characterized in that, S33 includes: Retrieve preliminary performance deviation data and parameter-index correlation model, determine the input boundary of the optimization model, configure the initial value and adjustment threshold of the inertia weight for 100 iterations, and generate the model input configuration table; Import the input configuration table into the hardware parameter co-optimization model, start iterative calculation, output intermediate parameter combinations every 20 iterations, and finally generate 10 complete hardware parameter co-adjustment schemes. Based on the priority of sensor applications, the evaluation dimensions of the analytic hierarchy process (AHP) were determined, and the weight allocation rules of 40% for immunity, 35% for accuracy, and 25% for sensitivity were clarified to build a solution evaluation framework. Substitute the 10 adjustment schemes into the evaluation framework, calculate the individual scores and comprehensive weighted scores of each group in the dimensions of anti-interference, accuracy, and sensitivity, and generate a scheme score ranking table; Based on the score ranking table, the top 3 solutions were selected, and the parameters of each solution were checked to see if they met the hardware specification limits. Finally, 3 candidate adjustment solutions were determined.

7. The debugging method for the AI ​​human body sensor according to claim 1, characterized in that, The S4 includes: S41. Based on the requirements of sensor application scenarios, refine the performance evaluation standards; establish the allowable range of evaluation error using the output data of standard calibrated sensors as a benchmark. S42. Perform multi-dimensional analysis on the output dataset after collaborative adjustment, quantify sensitivity using signal response time, calculate physiological feature recognition accuracy using confusion matrix, evaluate anti-interference capability using interference suppression ratio, verify reliability using long-term stability test, and generate a performance improvement evaluation report. S43. Compare the evaluation report with the detailed standards. If there are any indicators that do not meet the standards, use the controlled variable method to test the impact of each hardware parameter on the indicator one by one to locate the root cause. S44. Based on the root cause localization results, correct the target weights of the particle swarm optimization algorithm and regenerate the hardware parameter coordination adjustment scheme; fine-tune the key parameters and generate the corrected parameter set. S45, repeat the parameter loading and data acquisition process of S35, and the analysis and evaluation process of this step, until all performance indicators meet the refined standards, lock in the final hardware parameters, and generate the final hardware parameter optimization data.

8. The debugging method for the AI ​​human body sensor according to claim 1, characterized in that, The S5 includes: S51. Based on the final hardware parameters, optimize the data and identify the core steps of batch debugging; use Python to write automated debugging scripts, integrating sensor batch communication interfaces, environmental scene switching commands, and performance indicator calculation functions to generate an automated batch debugging process solution. S52. Select 10 sensors for small-batch trial debugging to verify the script's parameter loading success rate and scene switching stability; expand the debugging scene library and optimize the script's fault tolerance mechanism based on common production line problems. S53. Deploy the optimized automated batch debugging process to the production line debugging system. Through the linkage interface between the dedicated debugging device and the production line, realize the automated connection of sensor loading, debugging and unloading. Perform batch parameter configuration and multi-scenario debugging for large-scale production line sensors, record the debugging data of each sensor in real time, and generate a batch debugging completion dataset. S54. A stratified sampling inspection model of initial inspection + re-inspection is adopted. The initial inspection extracts 5% of the samples from the batch data and tests the core indicators in the production line self-inspection environment. The re-inspection extracts 30% of the samples that passed the initial inspection and conducts full-scenario repeated testing in an independent third-party debugging environment to verify performance consistency. S55. Summarize the batch debugging data and sampling results to generate the final debugging qualification report; and enter the debugging data of each sensor into the blockchain system.

9. An adjustment device for an AI human body sensor, characterized in that, It includes a memory, a processor, and a computer program stored on and executable on the memory, wherein the processor executes the program to implement the debugging method of the AI ​​human body sensor as described in any one of claims 1 to 8.

10. An AI human body sensor, having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the debugging method of the AI ​​human body sensor as described in any one of claims 1 to 8.

Images