A multi-channel AD data processing method for electrocardio 12-lead

Abstract

The application discloses a kind of for electrocardiogram 12 lead multipath AD data processing method, the application relates to biological electric signal acquisition and analog-digital conversion technical field, comprising the following steps: S1: constructing 12 lead signal acquisition channel mapping model, the model is clear 12 electrocardiogram lead electrode configuration, each lead corresponding analog signal input channel, the hardware architecture of multichannel AD conversion system and the signal transmission path characteristics of each channel.The for electrocardiogram 12 lead multipath AD data processing method, by constructing accurate channel mapping model, and combining dynamic signal feature analysis, realize the individualized precision configuration of each lead AD conversion parameter.The method is verified by model consistency test and physiological rule, and the dual protection mechanism of mechanism, effectively improve the accuracy and reliability of each channel signal acquisition, while ensuring the strict time sequence synchronization between multichannel, provide high-quality, high-consistency electrocardiogram data basis for clinical diagnosis.

Description

Technical Field

[0001] This invention relates to the field of bioelectrical signal acquisition and analog-to-digital conversion technology, specifically a multi-channel AD data processing method for 12-lead electrocardiogram. Background Technology

[0002] In the field of medical diagnostics, 12-lead electrocardiography (ECG) is a core tool for monitoring cardiac electrical activity and diagnosing heart diseases. The accuracy and efficiency of its signal acquisition directly affect clinical diagnostic results. Multi-channel analog-to-digital (AD) conversion is a crucial step in 12-lead signal acquisition, requiring the simultaneous processing of weak bioelectrical potential signals from 12 leads. However, current multi-channel AD conversion processes face significant technical bottlenecks: the amplitude of 12-lead signals is weak and susceptible to power frequency interference, electromyographic noise, etc., and traditional AD conversion methods have insufficient noise suppression capabilities, easily leading to signal distortion; high requirements for multi-channel synchronous acquisition, and timing deviations between channels can easily disrupt signal correlation, affecting the overall analysis of cardiac electrical activity; to meet real-time monitoring needs, high sampling rates often come with high power consumption, posing a challenge to the battery life of portable and wearable 12-lead devices; furthermore, the signal characteristics of different leads vary greatly, and uniform AD conversion parameters are difficult to adapt to all channels, resulting in poor acquisition performance of some channels. Therefore, how to construct an optimized multi-channel AD scheme tailored to the characteristics of 12-lead signals, improving signal acquisition accuracy and achieving multi-channel synchronization while reducing power consumption, has become an urgent technical problem to be solved. Summary of the Invention

[0003] The purpose of this invention is to provide a method for processing multichannel AD data for 12-lead electrocardiograms, so as to solve the problems mentioned in the background art.

[0004] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for multichannel AD data processing for 12-lead electrocardiogram, comprising the following steps:

[0005] S1: Construct a 12-lead signal acquisition channel mapping model. The model clarifies the electrode configuration of the 12 ECG leads, the analog signal input channel corresponding to each lead, the hardware architecture of the multi-channel AD conversion system, and the signal transmission path characteristics of each channel.

[0006] S2: Based on the 12-lead signal acquisition channel mapping model, within the set signal acquisition period, the signal characteristic parameters of each AD conversion channel are dynamically acquired. The signal characteristic parameters include at least the signal amplitude range, noise amplitude ratio, main frequency components and bandwidth requirements of the signal, and channel timing deviation.

[0007] S3: Based on the signal characteristic parameters of each channel obtained dynamically, a set of AD conversion optimization parameters is determined for the channel through a parameter decision mechanism. The AD conversion optimization parameters include at least the gain value of the programmable gain amplifier, the sampling rate, the timing compensation value, and the digital filtering parameters.

[0008] S4: Based on the AD conversion optimization parameters determined for each channel, independently configure and coordinately control the AD conversion process of each channel; after the AD conversion process is executed, perform verification screening on the output multi-channel digital signals based on physiological rules consistency; and dynamically adjust the working mode of each channel to achieve power consumption optimization based on the device power supply status and channel signal stability.

[0009] The parameter decision mechanism includes a model consistency test on the initially calculated AD conversion optimization parameters. This test involves substituting the initially calculated parameters into a virtual forward model based on the signal transmission path to predict the output signal quality after applying the set of parameters. If the prediction result deviates from the preset target by more than an allowable threshold, the optimization parameters are recalculated.

[0010] Furthermore, the verification screening based on physiological rule consistency includes constructing an inter-lead relationship model using the inherent mathematical relationship between the 12-lead ECG signals, and using this model to perform logical consistency verification on the optimized multi-channel digital signals. If it is found that the data of a certain channel seriously violates the physiological law, it is determined that the optimized parameters of that channel may be inaccurate, and parameter rollback or key monitoring measures are taken.

[0011] First, a 12-lead signal acquisition channel mapping model is constructed to fully describe the entire signal path from acquisition to conversion. This model needs to clearly define the analog signal input channels corresponding to each of the twelve standard ECG leads and describe in detail the hardware composition of the multi-channel AD conversion system, including the connection relationships between the analog signal conditioning module, the AD conversion unit, and the timing synchronization control module. Simultaneously, signal attenuation characteristics under different transmission paths are determined through experimental measurements, and these transmission characteristics are integrated into the model to form a complete link description covering electrodes, transmission cables, conditioning circuits, and the AD conversion unit, laying the physical foundation for subsequent targeted optimization.

[0012] Based on the established model, the system enters periodic operation. Within each set acquisition cycle, the system dynamically acquires the real-time signal characteristics of each AD conversion channel. These characteristics include, but are not limited to: the amplitude fluctuation range of the channel signal, the amplitude proportion of noise components in the signal, the main frequency components of the signal determined through spectrum analysis and the required bandwidth, and the deviation between the actual sampling time of the channel and the system's unified timing reference. These dynamically acquired parameters together constitute a quantitative description of the current signal state of the channel.

[0013] Subsequently, the system initiates a parameter decision and dual verification mechanism. Based on the aforementioned real-time signal characteristics, the system initially calculates a set of AD conversion optimization parameters for each channel, including the optimal gain setting of the programmable gain amplifier, a sampling frequency that satisfies the Nyquist theorem and considers synchronization requirements, a time offset for timing deviation compensation, and targeted digital filtering parameters. This initial parameter set is not applied immediately but first undergoes a model consistency check. This check substitutes the initial parameters into a built-in virtual forward model constructed based on the aforementioned channel mapping model, simulating the processed signal state and predicting its signal-to-noise ratio improvement and potential saturation risk. If the prediction results do not meet the preset performance improvement target or have an excessively high saturation risk, the parameter set is deemed unreliable, and the system will trigger a parameter recalculation process, which may involve adjusting the gain calculation strategy or selecting a more complex filtering scheme.

[0014] The parameter set that passes the model consistency check is allowed to be configured on the hardware. Each channel performs signal conditioning, synchronous acquisition, and AD conversion according to its specific parameters. After conversion, the system performs a second verification, namely, screening based on physiological rule consistency. The system constructs an inter-lead relationship model using the inherent and known mathematical relationships between ECG leads; among which, the known mathematical relationships include the voltage relationships between limb leads. This model performs a logical consistency check on the multi-channel digital signals after preliminary optimization and filtering, calculating the degree of conformity between the theoretical relationship values ​​and the measured values ​​between specific lead pairs. If the data of a certain channel is found to deviate significantly from the expected range based on physiological rules, it is highly suspected that the optimized parameters of that channel are inaccurate due to transient strong interference. At this time, the system will abandon the parameters calculated for that channel in the current cycle, automatically revert to the parameter configuration that has been verified to be effective in the previous cycle, and mark that channel as an object requiring key monitoring, strengthening the observation of its signal quality in subsequent cycles.

[0015] Finally, the system introduces an adaptive power management strategy. This strategy is not simply frequency reduction when the battery is low, but is deeply coupled with signal quality monitoring. Based on the rate of change of recent characteristic parameters, the system continuously evaluates the stability of signals in each channel and monitors the device's battery level. Different strategies are adopted in different battery ranges: at high battery levels, the system operates with all parameters; at medium battery levels, the sampling rate of channels with stable signals is appropriately reduced and non-core processing modules are shut down; at low battery levels, based on the analysis of the inter-lead relationship model, a key subset of leads is intelligently selected to maintain normal acquisition, while the sampling rate of the remaining leads is significantly reduced, ensuring that clinically valuable ECG information can still be reconstructed through algorithms under power-constrained conditions.

[0016] The entire process embodies a complete closed loop from model building, dynamic perception, intelligent decision-making, dual verification to adaptive optimization, ensuring a dynamic balance between processing accuracy, timing synchronization and system power consumption for multi-channel AD data.

[0017] Furthermore, the process of constructing the 12-lead signal acquisition channel mapping model specifically includes:

[0018] The electrode configuration of 12 ECG leads is defined, including limb leads and chest leads, with each lead corresponding to an independent analog signal input channel. A hardware architecture model of a multi-channel AD conversion system is constructed, which includes an analog signal conditioning module and an AD conversion unit corresponding to the number of analog signal input channels, as well as a timing synchronization control module. Each analog signal input channel is sequentially connected to the corresponding analog signal conditioning module and AD conversion unit, and the timing synchronization control module is connected to each AD conversion unit to coordinate the sampling timing.

[0019] The signal attenuation characteristics of each channel signal under specific transmission cable lengths and wiring methods are determined by experimental measurement, and a signal transmission attenuation model is established. The corresponding connection relationship between the analog signal input channel and the AD conversion unit, as well as the signal transmission attenuation model, are integrated to generate the 12-lead signal acquisition channel mapping model. This model is used to characterize the complete path from electrode signal acquisition through transmission to AD conversion and the key parameters of each link.

[0020] Furthermore, the process of dynamically acquiring the signal characteristic parameters of each AD conversion channel specifically includes:

[0021] The signal acquisition cycle is set according to the frequency characteristics of the electrocardiogram signal and the real-time requirements of clinical diagnosis. In each acquisition cycle, the analog signal input of each channel is acquired through the sampling circuit of the analog signal conditioning module, and the signal amplitude range of that channel in the current cycle is statistically obtained. The noise component contained in the signal of each channel is detected by the noise analysis module, and the proportion of noise amplitude to signal amplitude is calculated. The frequency distribution of the signal of each channel is analyzed by the spectrum analysis tool to determine the main frequency components of the signal and the required signal bandwidth. The timing synchronization control module monitors and records the difference between the actual acquisition time and the theoretical synchronization time of the signal of each channel to obtain the timing deviation of that channel.

[0022] Furthermore, the process of determining the AD conversion optimization parameters for each channel based on the aforementioned signal characteristic parameters specifically includes:

[0023] Based on the signal amplitude range and the noise amplitude ratio, combined with the full-scale input range of the AD conversion unit, the optimal gain value of the programmable gain amplifier for this channel is calculated and determined. The calculation principle is to ensure that the amplified signal is located in the middle region of the linear operating range of the AD conversion unit. Based on the main frequency components and bandwidth requirements of the signal, the minimum sampling rate for this channel is determined according to the Nyquist sampling theorem and considering anti-aliasing requirements. The sampling rate is then fine-tuned to reduce timing deviation. Based on the frequency distribution and noise characteristics of the signal, a digital filter type and its parameters suitable for this channel are designed. Based on the timing deviation, the required timing compensation value for this channel is calculated to adjust the start time of the AD conversion.

[0024] Furthermore, the model consistency check performed on the initially calculated AD conversion optimization parameters is specifically implemented as follows:

[0025] The virtual forward model integrates the attenuation characteristics of the signal transmission path and the response characteristics of the hardware module in the 12-lead signal acquisition channel mapping model. The preliminarily calculated gain value, sampling rate, and filtering parameters are input into the virtual forward model to simulate the state of the signal after processing by the model. The model outputs the predicted signal-to-noise ratio (SNR) improvement degree and the signal amplitude saturation risk index. The predicted SNR improvement degree is compared with the preset expected improvement target, and the saturation risk is evaluated to see if it is below the allowable limit. If the SNR improvement does not reach the target or the saturation risk is too high, the preliminary parameters are determined to be unsatisfactory, triggering a parameter recalculation process. The parameter recalculation process may include adjusting the gain calculation weights or selecting a more complex filter structure.

[0026] Furthermore, the inter-lead relationship model used in the verification screening based on physiological rule consistency is constructed and applied as follows:

[0027] The inter-lead relationship model is established based on the known mathematical relationships between limb leads. After obtaining the multi-channel digital signals after preliminary optimization and filtering, the signal amplitude or waveform characteristics of specific lead pairs are extracted. The theoretical relationship values ​​between these lead pairs are calculated using the inter-lead relationship model and compared with the relationship values ​​obtained from actual measurements. A reasonable deviation tolerance range is set. If the measured relationship value of a certain channel continuously exceeds the deviation tolerance range, it is determined that the current optimized parameters of that channel may be inapplicable due to transient interference. The system will abandon the optimized parameters calculated for that channel in this acquisition cycle and restore the parameter configuration that was verified to be effective in the previous acquisition cycle. At the same time, the channel will be marked so that its signal quality can be monitored in subsequent acquisition cycles.

[0028] Furthermore, the process of dynamically adjusting the operating mode of each channel based on the equipment power supply status and channel signal stability includes:

[0029] The system monitors the battery level of the device in real time; it evaluates the stability index of each channel signal, which is calculated based on the rate of change of signal characteristic parameters of the channel over several recent consecutive acquisition cycles; when the battery level is higher than a first threshold, all channels operate normally according to their respective optimized parameters; when the battery level drops to between the first and second thresholds, for channels whose signal stability index is higher than the set stability threshold, the sampling rate is reduced and unnecessary functional modules associated with that channel are shut down; when the battery level is lower than the second threshold, a low-power operation mode is activated. In this mode, the system selects a key subset of leads that can reconstruct the information of other leads to the greatest extent possible through the algorithm, based on the inter-lead relationship model, and maintains only the normal or near-normal sampling rate of the channels within this key subset of leads, while reducing the sampling rate of the remaining channels to a lower level that meets basic monitoring requirements.

[0030] Furthermore, the selection principle for the key lead subset in the low-power operation mode is:

[0031] Based on the aforementioned inter-lead relationship model, the impact of different lead combinations on the ability to represent complete ECG information is analyzed. A subset of leads is selected such that, with only the channel data of this subset retained, the main morphological features of the ECG signal represented by the closed or slowed channels can be recovered by using known inter-lead transformation relationships and computational reconstruction algorithms. This significantly reduces power consumption while maintaining the clinical usability of ECG monitoring information as much as possible.

[0032] Furthermore, the method also includes a continuous learning and adaptive optimization phase, including the following:

[0033] The system records the signal characteristic parameters of each channel and the corresponding optimized parameter configuration effects within the historical acquisition period; establishes a parameter configuration effect evaluation database for each channel; when the same or similar signal characteristics reappear, the system prioritizes calling the optimized parameter combinations that have been verified to have good effects in the historical database for trial, and makes fine adjustments based on the current actual effect, thereby accelerating the parameter optimization process and improving the accuracy of parameter configuration.

[0034] Furthermore, in the hardware architecture of the multi-channel AD conversion system, the timing synchronization control module is implemented by a field-programmable gate array (FPGA). The FPGA communicates with each AD conversion unit through a serial peripheral interface to precisely control the start-up timing of each channel's AD conversion. The analog signal conditioning module includes an instrumentation amplifier and an anti-aliasing filter. The AD conversion unit adopts a high-resolution analog-to-digital converter based on Sigma-Delta modulation technology.

[0035] This invention provides a method for multichannel AD data processing for 12-lead electrocardiograms. It has the following beneficial effects:

[0036] This multi-channel AD data processing method for 12-lead ECG achieves personalized and precise configuration of AD conversion parameters for each lead by constructing an accurate channel mapping model and combining it with dynamic signal feature analysis. Through a dual guarantee mechanism of model consistency verification and physiological rule validation, this method effectively improves the accuracy and reliability of signal acquisition for each channel, while ensuring strict temporal synchronization between multiple channels, providing a high-quality, highly consistent ECG data foundation for clinical diagnosis.

[0037] This multi-channel AD data processing method for 12-lead ECG achieves an intelligent balance between system performance and power consumption in complex application scenarios by introducing an adaptive power management strategy deeply coupled with signal quality and a continuous learning optimization mechanism based on historical experience. This method not only dynamically adjusts the operating mode to extend device battery life under different power conditions, but also improves the long-term stability and adaptability of the system through parameter self-learning and self-verification, effectively ensuring the practicality and reliability of portable ECG monitoring devices. Attached Figure Description

[0038] Figure 1 This is a flowchart illustrating a multichannel AD data processing method for a 12-lead electrocardiogram according to the present invention.

[0039] Figure 2 This is a system architecture diagram of a multichannel AD data processing method for 12-lead electrocardiogram according to the present invention;

[0040] Figure 3 This is a parameter decision flowchart for a multichannel AD data processing method for 12-lead electrocardiogram according to the present invention;

[0041] Figure 4 This is a lead relationship verification diagram for a multi-channel AD data processing method for 12-lead electrocardiogram according to the present invention. Detailed Implementation

[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] Please see Figures 1 to 4 This invention provides a technical solution: a method for processing multichannel AD data for 12-lead electrocardiogram, comprising the following steps:

[0044] S1: Construct a 12-lead signal acquisition channel mapping model. The model clarifies the electrode configuration of the 12 ECG leads, the analog signal input channel corresponding to each lead, the hardware architecture of the multi-channel AD conversion system, and the signal transmission path characteristics of each channel.

[0045] S2: Based on the 12-lead signal acquisition channel mapping model, within the set signal acquisition period, the signal characteristic parameters of each AD conversion channel are dynamically acquired. The signal characteristic parameters include at least the signal amplitude range, noise amplitude ratio, main frequency components and bandwidth requirements of the signal, and channel timing deviation.

[0046] S3: Based on the signal characteristic parameters of each channel dynamically acquired, a set of AD conversion optimization parameters is determined for the channel through a parameter decision mechanism. The AD conversion optimization parameters include at least the gain value of the programmable gain amplifier, the sampling rate, the timing compensation value, and the digital filtering parameters.

[0047] S4: Based on the AD conversion optimization parameters determined for each channel, independently configure and coordinate the AD conversion process of each channel; after the AD conversion process is executed, verify and screen the output multi-channel digital signals based on physiological rules consistency; and dynamically adjust the working mode of each channel to achieve power consumption optimization based on the power supply status of the device and the stability of the channel signals.

[0048] The parameter decision mechanism includes a model consistency test on the initially calculated AD conversion optimization parameters. This test involves substituting the initially calculated parameters into a virtual forward model based on the signal transmission path to predict the output signal quality after applying the set of parameters. If the prediction result deviates from the preset target by more than the allowable threshold, the optimization parameters are recalculated.

[0049] Furthermore, the verification screening based on physiological rule consistency includes constructing an inter-lead relationship model using the inherent mathematical relationship between 12-lead ECG signals, and using this model to perform logical consistency verification on the optimized multi-channel digital signals. If it is found that the data of a certain channel seriously violates physiological rules, it is determined that the optimized parameters of that channel may be inaccurate, and parameter rollback or key monitoring measures are taken.

[0050] The process of constructing the 12-lead signal acquisition channel mapping model specifically includes:

[0051] The electrode configuration of 12 ECG leads is defined, including limb leads and chest leads, with each lead corresponding to an independent analog signal input channel. A hardware architecture model of a multi-channel AD conversion system is constructed, which includes an analog signal conditioning module and an AD conversion unit corresponding to the number of analog signal input channels, as well as a timing synchronization control module. Each analog signal input channel is sequentially connected to the corresponding analog signal conditioning module and AD conversion unit, and the timing synchronization control module is connected to each AD conversion unit to coordinate the sampling timing.

[0052] The signal attenuation characteristics of each channel signal under specific transmission cable length and wiring method were determined by experimental measurement, and a signal transmission attenuation model was established. The corresponding connection relationship between the analog signal input channel and the AD conversion unit, as well as the signal transmission attenuation model, were integrated to generate a 12-lead signal acquisition channel mapping model. This model is used to characterize the complete path from electrode signal acquisition to AD conversion and the key parameters of each link.

[0053] It should be further explained that, in the specific implementation process, the construction of the 12-lead signal acquisition channel mapping model first requires the precise definition of the physical basis of ECG signal acquisition. The twelve standard ECG leads include six limb leads and six chest leads, each assigned an independent analog signal input channel. The electrodes are made of Ag / AgCl material with stable electrochemical properties to ensure effective acquisition of weak bioelectric potential signals from the human body surface and to minimize interference introduced by the electrode's own characteristics.

[0054] At the hardware architecture level, the system constructs an acquisition network comprising twelve independent analog signal conditioning modules and twelve corresponding AD conversion units. The signal flow path for each analog signal input channel is explicitly defined: it sequentially passes through the corresponding analog signal conditioning module, which typically contains an instrumentation amplifier for initial signal amplification and a pre-filter for initial interference suppression, before being connected to the designated AD conversion unit. A core timing synchronization control module establishes a connection with all AD conversion units through a digital communication interface. Its core function is to send a unified sampling clock reference and a precisely adjustable start signal to each unit, providing a guarantee for multi-channel synchronous acquisition from the hardware level; the digital communication interface is SPI.

[0055] To quantify signal changes during transmission, a systematic experimental model of signal attenuation is needed. Under controlled experimental conditions, such as using a simulated electrocardiogram (ECG) signal source with stable amplitude and frequency characteristics, the attenuation of the signal after passing through cables of different lengths and types is measured. By comparing and analyzing the experimental data, an optimal cable length range that balances signal attenuation with the flexibility of practical equipment deployment is determined. Simultaneously, twisted-pair cable with electromagnetic shielding is chosen as the transmission medium to effectively suppress common interferences such as power frequency interference in the environment. The attenuation at specific cable lengths measured experimentally, as well as the anti-interference performance parameters of the selected shielded cable, will be recorded in detail.

[0056] Finally, all the above elements are systematically integrated: a fixed mapping relationship between each lead number and its corresponding AD conversion unit number is clearly recorded; attenuation models and anti-interference parameters characterizing signal transmission are associated with the corresponding physical channels. The resulting 12-lead signal acquisition channel mapping model is not only a connection diagram, but also a complete information database containing key physical parameters and electrical characteristics throughout the entire process from the signal source, transmission path to the conversion terminal. This provides an indispensable underlying basis and structural framework for personalized and precise parameter optimization and problem diagnosis for each channel in subsequent steps. In this model, the signal source corresponds to the electrode, the transmission path to the cable, and the conversion terminal to the AD unit. The establishment of this model ensures that subsequent signal analysis and optimization are no longer based on assumptions about a general system, but on a deep understanding of the unique physical implementation of each channel.

[0057] The process of dynamically acquiring the signal characteristic parameters of each AD conversion channel specifically includes:

[0058] The signal acquisition cycle is set according to the frequency characteristics of the electrocardiogram signal and the real-time requirements of clinical diagnosis. In each acquisition cycle, the analog signal input of each channel is acquired through the sampling circuit of the analog signal conditioning module, and the signal amplitude range of that channel in the current cycle is statistically obtained. The noise component contained in the signal of each channel is detected by the noise analysis module, and the proportion of noise amplitude to signal amplitude is calculated. The frequency distribution of the signal of each channel is analyzed by the spectrum analysis tool to determine the main frequency components of the signal and the required signal bandwidth. The timing deviation of the channel is obtained by monitoring and recording the difference between the actual acquisition time and the theoretical synchronization time of the signal of each channel through the timing synchronization control module.

[0059] It should be further explained that, in the specific implementation process, the process of dynamically acquiring the signal characteristic parameters of each AD conversion channel begins with setting a unified signal acquisition cycle. The duration of this cycle needs to be determined by comprehensively considering the typical frequency range of the ECG signal and the real-time performance required for clinical diagnosis. After the system starts, each acquisition cycle is assigned a unique sequence number for tracking purposes.

[0060] Within each acquisition cycle, the system performs parallel extraction of multiple signal features. Through a precision sampling circuit integrated into the analog signal conditioning module, the system continuously measures the instantaneous voltage value of the analog signal input to each channel and statistically determines the amplitude fluctuation range of that channel's signal within that cycle, such as recording its maximum positive and negative offsets. Simultaneously, the system invokes a dedicated noise analysis module, which can identify and separate specific noise components mixed in with the signal, such as common power frequency interference or electromyographic noise, and calculate the ratio of the amplitude of these noise components to the amplitude of the useful signal.

[0061] To understand the frequency characteristics of the signal, the system uses spectrum analysis tools to transform the time-domain signal acquired from each channel. By analyzing its spectral distribution, the main frequency regions where signal energy is concentrated are identified, and the bandwidth occupied by the effective components of the signal in that channel is determined accordingly. This step is crucial for the subsequent rational setting of the sampling rate and filter parameters.

[0062] In addition, the system continuously monitors the actual sampling start time of each channel's AD conversion unit through a high-precision timing synchronization control module, compares it with the ideal synchronization time determined by the system's master clock, and calculates the slight difference between the two, which is the timing deviation value of the channel in the current cycle.

[0063] All the above operations—amplitude statistics, noise analysis, spectrum calculation, and timing monitoring—are performed independently and synchronously on each channel within each acquisition cycle. This constructs a dynamically updated set of characteristic parameters for each signal channel in the system, comprehensively reflecting its current signal quality and characteristics. This continuous signal "health check" mechanism is the fundamental prerequisite for achieving subsequent personalized parameter optimization and adaptive control.

[0064] The process of determining the AD conversion optimization parameters for each channel based on signal characteristic parameters specifically includes:

[0065] Based on the signal amplitude range and noise amplitude ratio, and combined with the full-scale input range of the AD conversion unit, the optimal gain value of the programmable gain amplifier for this channel is calculated and determined. The calculation principle is to ensure that the amplified signal is located in the middle region of the linear operating range of the AD conversion unit. According to the main frequency components and bandwidth requirements of the signal, the minimum sampling rate of this channel is determined according to the Nyquist sampling theorem and considering anti-aliasing requirements. The sampling rate is then fine-tuned to reduce timing deviation. Based on the frequency distribution and noise characteristics of the signal, the type and parameters of a digital filter suitable for this channel are designed. Based on the timing deviation, the timing compensation value required for this channel is calculated to adjust the start time of the AD conversion.

[0066] It should be further explained that, in the specific implementation process, after obtaining the dynamic signal characteristic parameters of each channel, the system enters the parameter optimization decision-making stage. First, for the gain value of the programmable gain amplifier of each channel, the system makes a comprehensive judgment based on the measured signal amplitude range and the proportion of noise amplitude for that channel. The basic principle is that when the signal amplitude is small and the noise effect is relatively significant, the gain is appropriately increased to enhance the signal amplitude and allow it to better occupy the linear operating range of the AD conversion unit; conversely, if the signal amplitude is close to the upper limit of the input of the AD conversion unit, the gain needs to be reduced to prevent signal saturation distortion. The specific calculations comprehensively consider the full-scale input voltage of the AD conversion unit, the maximum amplitude of the signal, and the noise level to ensure that the amplified signal is stable in the middle of the linear response region of the converter.

[0067] Secondly, regarding the determination of the sampling rate, the system strictly follows the Nyquist sampling theorem, using the main frequency components of the channel signal and the analyzed bandwidth requirements as a benchmark to calculate the theoretical minimum sampling rate. Based on this, an anti-aliasing reserve coefficient is introduced to provide necessary frequency margin. Simultaneously, the timing deviation of the channel is also taken into consideration; the system fine-tunes the theoretically calculated sampling rate by changing the sampling clock period to indirectly compensate for or reduce inherent timing differences between channels, thereby improving the overall synchronization accuracy of the multi-channel system.

[0068] Next, the system designs dedicated digital filtering parameters for each channel. This design is closely dependent on the frequency distribution characteristics of the channel signal and the identified noise characteristics. For example, for channels where the main ECG waveform components are concentrated, a bandpass filter is configured with a passband range that strictly matches the main frequency range of the channel signal; for channels susceptible to slow baseline drift, an additional high-pass filter is introduced to suppress such low-frequency interference. The filter type, order, and cutoff frequency are all configured independently for each channel.

[0069] Finally, the system directly uses the monitored timing deviation value to determine the timing compensation value. This compensation value is equal in magnitude but opposite in sign to the timing deviation. Its function is to advance or delay the transmission of the AD conversion start signal for that channel in the next acquisition cycle via the timing synchronization control module, thereby offsetting the measured timing deviation and achieving precise alignment of multiple channels at startup. The entire process reflects independent parameter configuration for each channel, ensuring a high degree of matching between optimization measures and the unique signal conditions of each channel.

[0070] The specific implementation method for performing model consistency checks on the initially calculated AD conversion optimization parameters is as follows:

[0071] The virtual forward model integrates the attenuation characteristics of the signal transmission path and the response characteristics of the hardware module from the 12-lead signal acquisition channel mapping model. The preliminarily calculated gain value, sampling rate, and filtering parameters are input into the virtual forward model to simulate the state of the signal after processing by the model. The model outputs the predicted signal-to-noise ratio (SNR) improvement and the signal amplitude saturation risk index. The predicted SNR improvement is compared with the preset expected improvement target, and the saturation risk is assessed to see if it is below the allowable limit. If the SNR improvement does not meet the target or the saturation risk is too high, the preliminary parameters are deemed unsatisfactory, triggering a parameter recalculation process. This recalculation process may include adjusting the gain calculation weights or selecting a more complex filter structure.

[0072] It should be further explained that, in the specific implementation process, after the system initially calculates the gain value, sampling rate, and filtering parameters for a channel, it does not immediately drive the hardware to execute them. Instead, it inputs these parameters into the virtual forward model corresponding to that channel. The model simulates an input signal with typical characteristics of that channel and allows it to go through the entire processing chain based on this parameter set. The key evaluation indicators of the model output mainly include two aspects: first, the degree of improvement in the predicted signal-to-noise ratio compared to before optimization; and second, whether the predicted signal, after amplification, might exceed the linear input range of the AD converter, i.e., the risk of saturation.

[0073] The system compares the signal-to-noise ratio (SNR) improvement predicted by the model with a preset expected improvement target. This preset target is determined in advance based on the basic signal quality requirements for clinical diagnosis. Simultaneously, the system assesses whether the saturation risk predicted by the model is within an acceptable limit. If the prediction results show that the SNR improvement fails to reach the preset target, or the saturation risk exceeds the allowable safety limit, then the system determines that this set of preliminary parameters has failed the model consistency test. Its application to actual hardware may not achieve the expected results, or may even have negative effects.

[0074] Once a parameter fails the test, the system immediately triggers a parameter recalculation process. This process doesn't simply repeat the original calculations; instead, it may adjust the weighting in the gain calculation, or select a more complex but potentially better-performing filter structure from the filter design library, regenerate a new set of candidate parameters, and then re-feed them to the virtual forward model for testing. This continues until a set of parameters with satisfactory model prediction performance is generated, or, after reaching the iteration limit, a backup conservative parameter scheme is activated. This mechanism ensures that every set of parameters ultimately applied to the physical hardware undergoes a predictable performance verification based on the system's own physical model, improving the reliability and effectiveness of parameter configuration.

[0075] The inter-lead relationship model used in the verification screening based on the consistency of physiological rules includes the following construction and application:

[0076] The inter-lead relationship model is established based on the known mathematical relationships between limb leads. After obtaining the multi-channel digital signals that have undergone preliminary optimization and filtering, the signal amplitude or waveform characteristics of specific lead pairs are extracted. The theoretical relationship values ​​between these lead pairs are calculated using the inter-lead relationship model and compared with the relationship values ​​obtained from actual measurements. A reasonable deviation tolerance range is set. If the measured relationship value of a certain channel continuously exceeds the deviation tolerance range, it is determined that the current optimized parameters of that channel may be inapplicable due to transient interference. The system will discard the optimized parameters calculated for that channel in the current acquisition cycle and restore the parameter configuration that was verified to be effective in the previous acquisition cycle. At the same time, the channel will be marked so that its signal quality can be monitored in subsequent acquisition cycles.

[0077] It should be further explained that, in the specific implementation process, the verification screening based on the consistency of physiological rules is achieved by constructing an inter-lead relationship model. This model is based on the fundamental principles of ECG signal acquisition, particularly utilizing the deterministic mathematical relationships that exist between limb leads. After obtaining multi-channel digital signals that have undergone preliminary optimization and filtering, the system extracts key features from the signals of specific lead pairs. These features are typically amplitude or waveform morphology parameters that reflect the spatial vector of cardiac electrical activity.

[0078] The system then uses an inter-lead relationship model to calculate the theoretical expected values ​​between the characteristics of these lead pairs based on known physiological rules. Simultaneously, it measures the corresponding actual values ​​from the acquired and processed digital signals. Next, the system meticulously compares the theoretical calculations with the actual measurements to assess the degree of agreement. This process employs a predefined tolerance range, determined by a combination of permissible errors in clinical ECG diagnosis and the system's own measurement accuracy.

[0079] If the system detects that the measured values ​​of a certain channel consistently deviate from the theoretical expectations beyond the tolerance range, it will not immediately adjust the parameters of that channel. Instead, it will first determine that this phenomenon is an anomaly. The system infers that the optimized parameters calculated for that channel in the current acquisition cycle may no longer be applicable due to unforeseen strong transient interference, or the signal itself may have undergone a temporary pathological change. As a response, the system will abandon the newly calculated set of optimized parameters for that channel in the current cycle and instead revert to the parameter configuration that has been verified as effective and stable in the previous acquisition cycle. At the same time, the system will set a monitoring flag for that channel and conduct more intensive tracking and evaluation of the signal characteristics and parameter effects of that channel in several subsequent acquisition cycles until its signal performance stabilizes or the anomaly is confirmed and eliminated. This mechanism provides the system with robustness against transient interference, preventing system instability caused by continuous parameter adjustments due to a single anomaly.

[0080] The process of dynamically adjusting the operating mode of each channel based on the equipment power supply status and channel signal stability includes:

[0081] The system monitors the device's battery level in real time; it evaluates the stability index of each channel's signal, calculated based on the rate of change of signal characteristic parameters over several recent acquisition cycles; when the battery level is above a first threshold, all channels operate normally according to their respective optimized parameters; when the battery level drops to between the first and second thresholds, for channels with a stability index higher than the set stability threshold, their sampling rate is reduced, and unnecessary functional modules associated with those channels are shut down; when the battery level is below the second threshold, a low-power operation mode is activated. In this mode, the system selects a key subset of leads that can reconstruct other lead information to the greatest extent possible through an algorithm based on the inter-lead relationship model, maintaining only the normal or near-normal sampling rate of channels within this key subset, while reducing the sampling rate of other channels to a lower level that meets basic monitoring requirements. It should be further noted that in the specific implementation process, the system achieves dynamic power management by real-time monitoring of the device's battery level and combining this with the evaluation of the signal stability of each channel. The signal stability evaluation is based on calculating the rate of change of signal characteristic parameters for each channel over several consecutive acquisition cycles, thereby generating a stability index reflecting the smoothness of the channel's signal. When the device battery level is higher than the set first threshold, the system maintains all channels operating normally according to their respective optimized parameters, with the primary goal of ensuring signal acquisition performance.

[0082] When the battery level drops to between the first and second thresholds, the system activates an intermediate power consumption optimization strategy: for channels whose signal stability index is higher than the preset stability threshold, the system appropriately reduces their sampling rate by a certain margin and simultaneously shuts down unnecessary background processing modules related to that channel, such as the continuously running real-time noise analysis unit, and instead adopts a periodic detection mode, thereby reducing system power consumption without significantly affecting the quality of critical signals.

[0083] When the battery level drops further below the second threshold, the system activates a low-power operation mode. In this mode, the system does not simply shut down some channels, but makes intelligent decisions based on the inter-lead relationship model. It analyzes and selects a subset of key leads that can reconstruct complete ECG information to the greatest extent possible through the algorithm, maintains the normal or near-normal sampling rate of the channels within this subset, and uniformly reduces the sampling rate of the remaining non-key channels to a low level that can only meet the basic ECG monitoring needs.

[0084] This tiered power management strategy ensures that the system can achieve a dynamic and intelligent balance between power consumption and the signal quality required for clinical use under different power conditions, thus avoiding the rapid depletion of power and maximizing the clinical value of ECG monitoring.

[0085] The selection principle for the critical lead subset in low-power operation mode is:

[0086] Based on the inter-lead relationship model, this study analyzes the impact of different lead combinations on the ability to represent complete ECG information. A subset of leads is selected such that, while retaining only the channel data of this subset, the main morphological characteristics of the ECG signals represented by closed or slowed channels can be recovered using known inter-lead transformation relationships and a computational reconstruction algorithm. This significantly reduces power consumption while maintaining the clinical usability of ECG monitoring information as much as possible. It should be further noted that, in the specific implementation process, under low-power operation mode, the selection of the key lead subset is an intelligent decision-making process based on the ECG vector principle. The system first calls the established inter-lead relationship model, which encodes the deterministic mathematical transformation relationships between signals in each lead of the standard 12-lead system. The system analyzes all possible lead combination schemes and evaluates the ability of each candidate subset to represent complete ECG information. Specifically, for each candidate subset, the system uses the transformation rules defined in the inter-lead relationship model to simulate and deduce the signal waveforms of those leads not directly acquired by that subset through a computational reconstruction algorithm. The system quantifies the information retention capability of the subset by comparing the morphological similarity and feature point preservation between the reconstructed waveform and the actual waveform acquired under normal conditions.

[0087] The selection principle is to define a subset of leads that, under the condition of satisfying hardware power consumption constraints, can reconstruct ECG signals from this subset using algorithms that maintain a high degree of consistency with the original signals in terms of key waveform features. This ensures that the data obtained after power reduction still has clear clinical diagnostic value. The key waveform features include the morphology, amplitude, and duration of the P wave, QRS complex, and T wave. This selection process is not fixed but model-driven, ensuring that even with the sacrifice of some channel data, the integrity of ECG information can still be maintained to the greatest extent possible through computation.

[0088] The method also includes continuous learning and adaptive optimization phases, including the following:

[0089] The system records the signal characteristic parameters of each channel and the corresponding optimized parameter configuration effects within historical acquisition periods; it establishes a parameter configuration effect evaluation database for each channel; when the same or similar signal characteristics reappear, the system prioritizes trying optimized parameter combinations that have been verified to have good results in the historical database, and makes fine adjustments based on the current actual effect, thereby accelerating the parameter optimization process and improving the accuracy of parameter configuration. It should be further noted that in the specific implementation process, the system has continuous learning and adaptive optimization capabilities, which is achieved by establishing a historical parameter configuration effect database. During operation, the system continuously records the set of signal characteristic parameters for each channel in different acquisition periods, as well as the AD conversion optimized parameter configuration used at that time. It also records the actual effect evaluation results after applying this set of parameters, verified through model consistency checks and physiological rules. This data is stored in a structured manner, forming a dedicated historical experience library for each channel. When the system detects a similar or identical signal characteristic combination in a channel in subsequent operations, it does not immediately start calculating optimized parameters from scratch, but first searches for matching cases in the historical database for that channel.

[0090] The system prioritizes parameter combinations that have historically proven effective, demonstrating significant signal quality improvement and passing subsequent tests, as candidate configurations for the current cycle. Based on these historically optimized parameters, the system makes subtle adjustments to the specific signal performance of the current cycle. This ensures parameter configuration reliability while accelerating the optimization process, reducing computational resource consumption, and, over time, making the parameter configurations provided to each channel increasingly accurate and efficient. This self-learning mechanism allows the system to accumulate operational experience and gradually adapt to changes in signal characteristics among different patients or under different physiological states of the same patient.

[0091] In the hardware architecture of the multi-channel AD conversion system, the timing synchronization control module is implemented by a field-programmable gate array (FPGA). The FPGA communicates with each AD conversion unit through a serial peripheral interface to precisely control the start-up timing of each channel's AD conversion. The analog signal conditioning module includes an instrumentation amplifier and an anti-aliasing filter. The AD conversion unit uses a high-resolution analog-to-digital converter based on Sigma-Delta modulation technology. It should be further noted that, in the specific implementation, the timing synchronization control module in the hardware architecture of the multi-channel AD conversion system is implemented by the FPGA. This FPGA establishes a communication link with each AD conversion unit through a serial peripheral interface. Utilizing its parallel processing capability and programmable characteristics, it generates a high-precision synchronous clock reference and sends a start-up conversion signal with an independently programmable delay to each AD conversion unit, thereby achieving precise coordinated control of the start-up time of the twelve channels' AD conversions and ensuring the timing consistency of signal acquisition from the hardware perspective. The analog signal conditioning module specifically includes an instrumentation amplifier and an anti-aliasing filter. The instrumentation amplifier is responsible for the initial amplification of the weak differential ECG signal acquired by the electrodes with a high common-mode rejection ratio. The anti-aliasing filter is placed after the amplifier to filter out high-frequency noise components in the signal that are higher than the Nyquist frequency, preventing spectral aliasing during the sampling process.

[0092] The AD conversion unit uses a high-resolution analog-to-digital converter based on Sigma-Delta modulation technology. This type of converter pushes quantization noise to the high-frequency band through oversampling and noise shaping technology, and then filters it out after digital filtering, thereby achieving high-resolution signal conversion at a relatively low hardware cost. It is especially suitable for bioelectrical signal acquisition scenarios with high dynamic range requirements, such as electrocardiogram signals.

[0093] It should be further explained that, in the specific implementation process, this method constructs a complete closed-loop processing system from signal acquisition, feature analysis, parameter optimization, real-time verification to adaptive control. The system first needs to establish a precise twelve-lead signal acquisition channel mapping model. This model clearly defines the correspondence between each standard ECG lead and a specific analog signal input channel, and details the complete hardware connection path from electrodes and signal conditioning modules to analog-to-digital converters (ADCs). The signal conditioning module typically includes an instrumentation amplifier for initial amplification of weak ECG signals and an anti-aliasing filter for suppressing high-frequency interference. The ADC is responsible for converting analog signals to digital signals. A key timing synchronization control module, for example implemented using a field-accessible gate array (FGA), connects to all ADCs via a serial peripheral interface and is responsible for coordinating the sampling timing of each channel.

[0094] After the system starts running, it enters a periodic signal acquisition and processing cycle. The duration of each acquisition cycle is determined comprehensively based on the frequency characteristics of the ECG signal and the real-time requirements of clinical monitoring. Within each cycle, the system performs real-time signal feature extraction on all twelve channels in parallel. This includes measuring and statistically analyzing the voltage amplitude fluctuation range of each channel's signal, analyzing the amplitude of noise components in the signal and calculating their relative proportion to the useful signal, determining the main frequency components and required bandwidth of the signal through spectral analysis, and monitoring the deviation between the actual sampling time of each channel and the system's unified timing reference. These dynamically acquired feature parameters together constitute a comprehensive description of the current signal state of each channel.

[0095] Based on the aforementioned real-time characteristic parameters, the system enters the personalized parameter decision-making stage. For each channel, the system independently calculates a set of analog-to-digital conversion optimization parameters. The optimal gain value of the programmable gain amplifier is calculated using the following formula: G i =[V FS ×(1-N ki )] / [2×max(A ki )]; where: G i V represents the gain value of the i-th channel. FS N represents the full-scale input voltage of the analog-to-digital converter unit. ki The maximum value of the i-th channel in the k-th acquisition period represents the proportion of noise amplitude. ki () represents the maximum amplitude of the signal in the i-th channel during the k-th acquisition period. The principle for determining the gain parameter is to make the amplitude of the amplified signal match the linear input range of the analog-to-digital converter, avoiding the signal being too small and overwhelmed by noise or too large and causing saturation distortion.

[0096] The sampling rate must satisfy the Nyquist sampling theorem, cover the main frequency components of the signal, and allow sufficient margin to prevent aliasing. It should also be fine-tuned based on the timing deviation of the channel to improve synchronization. Digital filter parameters are designed specifically for the signal's frequency distribution and noise characteristics; for example, a bandpass filter is configured for channels with concentrated ECG characteristic waveform components. Timing compensation values ​​are directly determined based on the measured timing deviation and are used to adjust the channel's start-up time in the next cycle. Specifically, the sampling rate must satisfy S... i =2.2×max(F ki ), where S i Let max(F) represent the sampling rate of the i-th channel. ki The ) represents the maximum frequency component of the signal in the i-th channel during the k-th acquisition period, and 2.2 is the anti-aliasing reserve coefficient. Timing compensation value τ i Based directly on the measured time series deviation Δt ki Determined, the relationship is τ i =-Δtki ; where τ i : Timing compensation value of the i-th channel, Δt ki : The deviation between the actual sampling time of the i-th channel and the unified timing reference of the system during the k-th acquisition period.

[0097] To ensure the reliability of the optimized parameters, the system employs a dual verification mechanism. The first is a model consistency check. The system maintains a virtual forward model that integrates hardware characteristics from the channel mapping model, such as transmission attenuation and amplifier response. The initially calculated optimized parameters are input into this model to simulate signal processing and predict the signal-to-noise ratio improvement and saturation risk of the output signal. If the prediction results do not meet the preset performance target, the parameters are deemed unreliable, and the system will trigger a parameter recalculation process, potentially adjusting the calculation strategy or selecting a more complex filtering scheme. Only parameters that pass the verification are allowed to be configured in the hardware.

[0098] The second layer of verification is a consistency screening based on physiological rules. The system constructs an inter-lead relationship model using the inherent and known mathematical relationships between ECG leads. After parameter application and analog-to-digital conversion, the system uses this model to perform logical consistency checks on the optimized multi-channel digital signals. For example, it checks whether the relationship I = II - III is satisfied between leads I, II, and III of a specific limb. If the data of a certain channel is found to deviate significantly from theoretical expectations, it is inferred that its current parameters may be inaccurate due to transient interference. At this time, the system will discard the new parameters, revert to the effective parameter configuration of the previous cycle, and mark the channel for key monitoring.

[0099] In terms of power management, the system implements adaptive dynamic adjustments. The strategy considers both the device's battery level and the stability of each channel's signal. Signal stability is assessed by calculating the rate of change of recent characteristic parameters. When the battery is sufficient, all channels operate with optimal parameters. When the battery level drops to a moderate level, the sampling rate of channels with stable signals is appropriately reduced, and unnecessary functional modules are shut down. When the battery is low, the system enters a low-power mode, intelligently selecting a subset of key leads based on the inter-lead relationship model. The selection principle for this subset is to reconstruct the most complete ECG information possible through algorithms. Only channels within this subset maintain normal acquisition, while the sampling rate of other channels is significantly reduced, thereby significantly reducing system energy consumption while ensuring that core clinical information is not lost.

[0100] Furthermore, the system possesses continuous learning capabilities. It records the signal characteristics of each channel in historical acquisition cycles, the applied optimization parameters, and their verification effects, forming an experience database. When encountering similar signal characteristics again, the system prioritizes calling up the parameter combinations that have been verified in the past, and then fine-tunes them, thereby accelerating the optimization process and improving the accuracy of parameter configuration.

[0101] This method, by constructing an accurate physical channel model, realizing dynamic feature perception and parameter decision-making, performing rigorous dual verification, implementing intelligent power consumption control, and incorporating a continuous learning mechanism, forms a highly adaptive and robust closed-loop processing system, effectively solving the technical challenge of balancing high precision, high synchronization, and low power consumption in the acquisition of twelve-lead ECG signals.

[0102] By constructing a precise channel mapping model and combining it with dynamic signal feature analysis, personalized and accurate configuration of AD conversion parameters for each lead was achieved. This method effectively improves the accuracy and reliability of signal acquisition for each channel through a dual guarantee mechanism of model consistency verification and physiological rule validation, while ensuring strict temporal synchronization between multiple channels, providing a high-quality, highly consistent ECG data foundation for clinical diagnosis.

[0103] By introducing an adaptive power management strategy deeply coupled with signal quality, and a continuous learning optimization mechanism based on historical experience, an intelligent balance between system performance and power consumption is achieved in complex application scenarios. This method not only dynamically adjusts the operating mode to extend the device's battery life under different power conditions, but also improves the system's long-term stability and adaptability through parameter self-learning and self-verification, effectively ensuring the practicality and reliability of portable ECG monitoring devices.

[0104] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0105] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for processing multichannel AD data for a 12-lead electrocardiogram, characterized in that, Includes the following steps: S1: Construct a 12-lead signal acquisition channel mapping model. The model clarifies the electrode configuration of the 12 ECG leads, the analog signal input channel corresponding to each lead, the hardware architecture of the multi-channel AD conversion system, and the signal transmission path characteristics of each channel. S2: Based on the 12-lead signal acquisition channel mapping model, within the set signal acquisition period, the signal characteristic parameters of each AD conversion channel are dynamically acquired. The signal characteristic parameters include at least the signal amplitude range, noise amplitude ratio, main frequency components and bandwidth requirements of the signal, and channel timing deviation. S3: Based on the signal characteristic parameters of each channel obtained dynamically, a set of AD conversion optimization parameters is determined for the channel through a parameter decision mechanism. The AD conversion optimization parameters include at least the gain value of the programmable gain amplifier, the sampling rate, the timing compensation value, and the digital filtering parameters. S4: Based on the AD conversion optimization parameters determined for each channel, independently configure and coordinately control the AD conversion process of each channel; after the AD conversion process is executed, perform verification screening on the output multi-channel digital signals based on physiological rules consistency; and dynamically adjust the working mode of each channel for power consumption optimization based on the device power supply status and channel signal stability. The parameter decision-making mechanism includes a model consistency check on the initially calculated AD conversion optimization parameters. This check involves substituting the initially calculated parameters into a virtual forward model based on the signal transmission path to predict the output signal quality after applying the initially calculated AD conversion optimization parameters. If the prediction result deviates from the preset target by more than an allowable threshold, the optimization parameters are recalculated. The virtual forward model integrates the attenuation characteristics of the signal transmission path and the response characteristics of the hardware module in the 12-lead signal acquisition channel mapping model. The initially calculated gain value, sampling rate, and filtering parameters are input into the virtual forward model to simulate the state of the signal after processing by the model. The model outputs the predicted signal-to-noise ratio improvement degree and the signal amplitude saturation risk index. The predicted signal-to-noise ratio improvement degree is compared with the preset expected improvement target, and the saturation risk is assessed to see if it is lower than the allowable limit. If the signal-to-noise ratio improvement does not reach the target or the saturation risk is too high, the initial parameters are determined to be unsatisfactory, triggering a parameter recalculation process. The parameter recalculation process may include adjusting the gain calculation weights or selecting a more complex filter structure. Furthermore, the verification screening based on physiological rule consistency includes constructing an inter-lead relationship model using the inherent mathematical relationship between the 12-lead ECG signals, and using this model to perform logical consistency verification on the optimized multi-channel digital signals. If it is found that the data of a certain channel seriously violates the physiological law, it is determined that the optimized parameters of that channel may be inaccurate, and parameter rollback or key monitoring measures are taken.

2. The method for multichannel AD data processing for 12-lead electrocardiogram according to claim 1, characterized in that: The process of constructing the 12-lead signal acquisition channel mapping model specifically includes: The electrode configuration of 12 ECG leads is defined, including limb leads and chest leads, with each lead corresponding to an independent analog signal input channel. A hardware architecture model of a multi-channel AD conversion system is constructed, which includes an analog signal conditioning module and an AD conversion unit corresponding to the number of analog signal input channels, as well as a timing synchronization control module. Each analog signal input channel is sequentially connected to the corresponding analog signal conditioning module and AD conversion unit, and the timing synchronization control module is connected to each AD conversion unit to coordinate the sampling timing. The signal attenuation characteristics of each channel signal under specific transmission cable lengths and wiring methods are determined by experimental measurement, and a signal transmission attenuation model is established. The corresponding connection relationship between the analog signal input channel and the AD conversion unit, as well as the signal transmission attenuation model, are integrated to generate the 12-lead signal acquisition channel mapping model. This model is used to characterize the complete path from electrode signal acquisition through transmission to AD conversion and the key parameters of each link.

3. The method for multichannel AD data processing for 12-lead ECG as described in claim 2, characterized in that: The process of dynamically acquiring the signal characteristic parameters of each AD conversion channel specifically includes: The signal acquisition cycle is set according to the frequency characteristics of the electrocardiogram signal and the real-time requirements of clinical diagnosis. In each acquisition cycle, the analog signal input of each channel is acquired through the sampling circuit of the analog signal conditioning module, and the signal amplitude range of that channel in the current cycle is statistically obtained. The noise component contained in the signal of each channel is detected by the noise analysis module, and the proportion of noise amplitude to signal amplitude is calculated. The frequency distribution of the signal of each channel is analyzed by the spectrum analysis tool to determine the main frequency components of the signal and the required signal bandwidth. The timing synchronization control module monitors and records the difference between the actual acquisition time and the theoretical synchronization time of the signal of each channel to obtain the timing deviation of that channel.

4. The method for multichannel AD data processing for 12-lead ECG as described in claim 3, characterized in that: The process of determining the AD conversion optimization parameters for each channel based on the aforementioned signal characteristic parameters specifically includes: Based on the signal amplitude range and the noise amplitude ratio, combined with the full-scale input range of the AD conversion unit, the optimal gain value of the programmable gain amplifier for this channel is calculated and determined. The calculation principle is to ensure that the amplified signal is located in the middle region of the linear operating range of the AD conversion unit. Based on the main frequency components and bandwidth requirements of the signal, the minimum sampling rate for this channel is determined according to the Nyquist sampling theorem and considering anti-aliasing requirements. The sampling rate is then fine-tuned to reduce timing deviation. Based on the frequency distribution and noise characteristics of the signal, a digital filter type and its parameters suitable for this channel are designed. Based on the timing deviation, the required timing compensation value for this channel is calculated to adjust the start time of the AD conversion.

5. A method for processing multichannel AD data for 12-lead electrocardiogram according to claim 4, characterized in that: The specific implementation method for the model consistency check performed on the initially calculated AD conversion optimization parameters is as follows: The virtual forward model integrates the attenuation characteristics of the signal transmission path and the response characteristics of the hardware module in the 12-lead signal acquisition channel mapping model. The preliminarily calculated gain value, sampling rate, and filtering parameters are input into the virtual forward model to simulate the state of the signal after processing by the model. The model outputs the predicted signal-to-noise ratio (SNR) improvement degree and the signal amplitude saturation risk index. The predicted SNR improvement degree is compared with the preset expected improvement target, and the saturation risk is evaluated to see if it is below the allowable limit. If the SNR improvement does not reach the target or the saturation risk is too high, the preliminary parameters are determined to be unsatisfactory, triggering a parameter recalculation process. The parameter recalculation process may include adjusting the gain calculation weights or selecting a more complex filter structure.

6. A method for processing multichannel AD data for 12-lead electrocardiogram according to claim 5, characterized in that: The inter-lead relationship model used in the verification screening based on physiological rule consistency includes the following aspects in its construction and application: The inter-lead relationship model is established based on the known mathematical relationships between limb leads. After obtaining the multi-channel digital signals after preliminary optimization and filtering, the signal amplitude or waveform characteristics of specific lead pairs are extracted. The theoretical relationship values ​​between these lead pairs are calculated using the inter-lead relationship model and compared with the relationship values ​​obtained from actual measurements. A reasonable deviation tolerance range is set. If the measured relationship value of a certain channel continuously exceeds the deviation tolerance range, it is determined that the current optimized parameters of that channel may be inapplicable due to transient interference. The system will abandon the optimized parameters calculated for that channel in this acquisition cycle and restore the parameter configuration that was verified to be effective in the previous acquisition cycle. At the same time, the channel will be marked so that its signal quality can be monitored in subsequent acquisition cycles.

7. A method for processing multichannel AD data for 12-lead electrocardiogram according to claim 6, characterized in that: The process of dynamically adjusting the operating mode of each channel based on the equipment power supply status and channel signal stability includes: The system monitors the battery level of the device in real time; it evaluates the stability index of each channel signal, which is calculated based on the rate of change of signal characteristic parameters of the channel over several recent consecutive acquisition cycles; when the battery level is higher than a first threshold, all channels operate normally according to their respective optimized parameters; when the battery level drops to between the first and second thresholds, for channels whose signal stability index is higher than the set stability threshold, the sampling rate is reduced and unnecessary functional modules associated with that channel are shut down; when the battery level is lower than the second threshold, a low-power operation mode is activated. In this mode, the system selects a key subset of leads that can reconstruct the information of other leads to the greatest extent possible through the algorithm, based on the inter-lead relationship model, and maintains only the normal or near-normal sampling rate of the channels within this key subset of leads, while reducing the sampling rate of the remaining channels to a lower level that meets basic monitoring requirements.

8. A method for processing multichannel AD data for 12-lead electrocardiogram according to claim 7, characterized in that: The selection principle for the key lead subset in the low-power operation mode is: Based on the aforementioned inter-lead relationship model, the impact of different lead combinations on the ability to represent complete ECG information is analyzed. A subset of leads is selected such that, with only the channel data of this subset retained, the main morphological features of the ECG signal represented by the closed or slowed channels can be recovered by using known inter-lead transformation relationships and computational reconstruction algorithms. This significantly reduces power consumption while maintaining the clinical usability of ECG monitoring information as much as possible.

9. A method for processing multichannel AD data for 12-lead electrocardiograms according to claim 8, characterized in that: The method also includes continuous learning and adaptive optimization steps, including the following: The system records the signal characteristic parameters of each channel and the corresponding optimized parameter configuration effects within the historical acquisition period; establishes a parameter configuration effect evaluation database for each channel; when the same or similar signal characteristics reappear, the system prioritizes calling the optimized parameter combinations that have been verified to have good effects in the historical database for trial, and makes fine adjustments based on the current actual effect, thereby accelerating the parameter optimization process and improving the accuracy of parameter configuration.

10. A method for processing multichannel AD data for a 12-lead electrocardiogram according to any one of claims 2 to 9, characterized in that: In the hardware architecture of the multi-channel AD conversion system, the timing synchronization control module is implemented by a field-programmable gate array (FPGA). The FPGA communicates with each AD conversion unit through a serial peripheral interface to precisely control the start-up timing of each channel's AD conversion. The analog signal conditioning module includes an instrumentation amplifier and an anti-aliasing filter. The AD conversion unit adopts a high-resolution analog-to-digital converter based on Sigma-Delta modulation technology.

Images