Clock running time error compensation method and system based on multi-source sensing data fusion

Abstract

The application relates to the technical field of electronic timers, and discloses a clock running time error compensation method and system based on multi-source sensing data fusion, which comprises the following steps: performing reliability analysis on the standardized time sequence data of a clock, determining the data quality weight of a multi-source sensor in the clock within a corresponding time window, performing weighted fusion processing on the standardized time sequence data of the multi-source sensor to obtain environmental state fusion data of the clock, performing environmental interaction collaborative calculation on the feature vectors of the environmental state fusion data to obtain running time error comprehensive compensation parameters of the clock, performing confidence degree mapping on the covariance matrix of the environmental state fusion data to determine the confidence degree index of the running time error comprehensive compensation parameters, and adjusting the output frequency of a timing pulse generation unit in the clock based on the running time error comprehensive compensation parameters, so that the accuracy and reliability of the clock running time error compensation in a complex environment are comprehensively improved.

Description

Technical Field

[0001] This invention relates to the field of electronic timekeeping technology, and in particular to a method and system for compensating for timekeeping errors in clocks based on multi-source sensor data fusion. Background Technology

[0002] When electronic clocks operate under complex environmental conditions such as temperature changes, vibration, shock, and magnetic field interference, the frequency of their internal reference oscillation source will drift, causing timekeeping errors to exceed the allowable range. Current technologies generally employ a single sensor for environmental monitoring and compensation to suppress timekeeping errors caused by such environmental factors. For example, a temperature sensor is installed inside the watch movement to compensate for the oscillation frequency of the quartz crystal oscillator using temperature measurement data. The limitation of this single-sensor method is that it can only sense environmental changes in a single dimension and cannot comprehensively reflect the combined impact of multiple physical factors such as vibration and magnetic fields on the watch's timekeeping accuracy. When the watch is in a real-world application scenario where multiple factors coexist, the compensation capability of a single sensor is clearly insufficient.

[0003] Current solutions for integrating multiple environmental sensors into a clock typically send the data collected by each sensor to an independent compensation channel. Each channel calculates its compensation amount separately, and these are then arithmetically summed to generate the frequency correction. This approach has three main drawbacks. First, the sensor data is not subject to quality inspection and reliability screening before entering the compensation calculation. Inferior data from individual sensors, affected by transient shocks or localized saturation interference, directly participates in the compensation calculation, contaminating the final result. Second, the compensation amounts for each environmental factor are treated as independent, unrelated components and superimposed, completely ignoring the objectively existing interactive coupling effects of multiple factors such as temperature, vibration, and magnetic fields at the physical level. High-temperature environments exacerbate the modulation effect of vibration on the crystal oscillator frequency, and the frequency impact of multiple factors acting simultaneously is not a simple summation of their individual effects. Third, the calculated compensation parameters are directly used for adjusting the timing pulse frequency without assessing their confidence level. When the quality of the fused data deteriorates, reducing the reliability of the compensation solution, unreliable compensation parameters are still applied to the timing pulse generation unit, introducing additional timekeeping deviations. Summary of the Invention

[0004] This invention provides a method and system for compensating for timekeeping errors in watches based on multi-source sensor data fusion, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a clock timekeeping error compensation method based on multi-source sensor data fusion, comprising:

[0006] B1: Perform reliability analysis on the standardized time series data of the clock, and determine the data quality weights of the multi-source sensors in the clock within the corresponding time window based on the analysis results;

[0007] B2: Based on the data quality weights, the standardized time-series data from the multi-source sensors are weighted and fused to obtain the environmental state fusion data of the clock.

[0008] B3: Based on the preset environmental factor coupling relationship, the feature vector of the environmental state fusion data is subjected to environmental interaction and collaborative solution to obtain the comprehensive compensation parameters for the timekeeping error of the clock.

[0009] B4: Perform confidence mapping on the covariance matrix of the environmental state fusion data to determine the confidence index of the comprehensive compensation parameter for travel time error;

[0010] B5: When the confidence index meets the timekeeping requirements of the clock, the output frequency of the timekeeping pulse generation unit in the clock is adjusted based on the comprehensive compensation parameters for the timekeeping error.

[0011] In a preferred embodiment, the step of performing reliability analysis on the standardized time-series data of the clock, and determining the data quality weights of the multi-source sensors in the clock within the corresponding time window based on the analysis results, includes:

[0012] Based on the standardized time-series data of the multi-source sensors within the corresponding time window, a data sequence corresponding to the multi-source sensors is constructed.

[0013] Statistical feature extraction is performed on the data sequence to obtain the fluctuation feature values ​​corresponding to the multi-source sensors;

[0014] Based on the normalized time-series data corresponding to the multi-source sensors, a group consensus benchmark is extracted to obtain a dynamic reference sequence;

[0015] The data sequence is matched and compared with the dynamic reference sequence to obtain the deviation feature value corresponding to the multi-source sensor;

[0016] Based on the fluctuation characteristic value and the deviation characteristic value, the multi-source sensor is comprehensively evaluated to obtain the data quality weight of the multi-source sensor within the corresponding time window.

[0017] In a preferred embodiment, the step of weighted fusion processing of the standardized time-series data from the multi-source sensors based on the data quality weights to obtain the environmental state fusion data of the clock includes:

[0018] Based on the data quality weights, the standardized time series data is subjected to positional weighting to obtain the weighted time series data corresponding to the multi-source sensors;

[0019] Based on the sampling time points of the multi-source sensors, the weighted time series data is summed in pairs to obtain the fused time series data within the corresponding time window;

[0020] Environmental features are extracted from the fused time-series data to obtain the fused environmental state data of the clock within the corresponding time window.

[0021] In a preferred embodiment, the step of performing environmental interactive collaborative calculation on the feature vector of the environmental state fusion data based on a preset environmental factor coupling relationship to obtain the comprehensive compensation parameters for the clock's timekeeping error includes:

[0022] Multidimensional feature decomposition is performed on the environmental state fusion data to obtain the multidimensional factor feature components corresponding to the environmental state fusion data;

[0023] The multidimensional factor feature components include single-factor feature components and coupled feature components, and the preset environmental coupling relationship includes single-factor influence rules and cross-influence rules.

[0024] Based on the single-factor influence rule, independent compensation analysis is performed on the single-factor characteristic components to obtain the single-factor compensation components corresponding to each environmental factor.

[0025] Based on the cross-influence rule, the coordinated compensation amount of the coupled feature components is calculated to obtain the comprehensive timekeeping error compensation parameters of the clock.

[0026] In a preferred embodiment, the calculation formula for the timekeeping comprehensive error compensation parameter is as follows:

[0027] ;

[0028] In the formula, This represents the timekeeping comprehensive error compensation parameter. This represents the single-factor index set of the clock. Represented as the first of the clocks Each single-factor characteristic component , Represented as the first Single-factor influence coefficient, Represented as the set of coupled indexes of the clocks. and This is represented by the two single-factor feature components that constitute the coupling pair. , For the coupling pair The corresponding coupling effect coefficient, The coupling nonlinearity order parameter of the clock is represented as follows. This is represented as the fused environmental state data. The degree of coupling activation is represented by the statistical fluctuation characteristics of the fused environmental state data. The time constant represents the coupling activation degree of the clock.

[0029] In a preferred embodiment, the coupling activation degree is determined by the statistical fluctuation characteristics of the environmental state fusion data, including:

[0030] The information entropy value of the environmental state fusion data is obtained by normalizing the data sequence of the environmental state fusion data within the corresponding time window.

[0031] The temporal gradient of the data sequence is analyzed to determine the gradient intensity value of the environmental state fusion data.

[0032] By performing a weighted analysis on the information entropy value and the gradient intensity value, a comprehensive activity index of the environmental state fusion data is obtained;

[0033] The comprehensive activity index is mapped to a preset activation value range to obtain the coupled activation degree of the clock within the corresponding time window.

[0034] In a preferred embodiment, the step of performing confidence mapping on the covariance matrix of the fused environmental state data to determine the confidence index of the comprehensive compensation parameters for travel time errors includes:

[0035] Construct the covariance matrix of the environmental state fusion data within the corresponding time window;

[0036] The diagonal elements of the covariance matrix represent the data autovariance of each environmental factor, and the off-diagonal elements represent the data covariance between different environmental factors.

[0037] Based on the eigenvalue set of the covariance matrix, the discreteness analysis of the environmental state fusion data is performed to obtain the uncertainty measure of the environmental state fusion data.

[0038] The confidence index of the time error comprehensive compensation parameter is obtained by performing a confidence mapping evaluation on the uncertainty metric value.

[0039] In a preferred embodiment, the step of performing a confidence mapping evaluation on the uncertainty metric to obtain the confidence index of the timekeeping error comprehensive compensation parameter includes:

[0040] The trace of the covariance matrix is ​​used as the total variance of the environmental state fusion data;

[0041] The largest eigenvalue in the set of eigenvalues ​​is used as the main direction discrete measure of the environmental state fusion data;

[0042] The ratio of the total variance to the main direction discreteness metric is processed to obtain the structural stability ratio of the environmental state fusion data.

[0043] Based on the structural stability ratio, the uncertainty metric is weighted and corrected, and the correction result is mapped to a preset confidence interval to obtain the confidence index of the comprehensive compensation parameter for the time travel error.

[0044] In a preferred embodiment, when the confidence index meets the timekeeping requirements of the clock, adjusting the output frequency of the timing pulse generation unit in the clock based on the comprehensive compensation parameters for timekeeping errors includes:

[0045] The confidence index is compared with the confidence requirement threshold corresponding to the preset watch timing requirement. When the confidence index reaches the confidence requirement threshold, it is determined that the timekeeping error comprehensive compensation parameter meets the watch timing requirement.

[0046] When it is determined that the timekeeping requirement of the clock is met, the frequency control dimension analysis is performed on the comprehensive compensation parameter for the timekeeping error to obtain the frequency adjustment amount that matches the timing pulse generation unit.

[0047] The frequency adjustment amount is written into the frequency control register of the timing pulse generation unit to fine-tune the oscillation frequency of the timing pulse generation unit, thereby obtaining the corrected timing pulse signal of the clock.

[0048] To address the aforementioned problems, this invention also provides a clock timekeeping error compensation system based on multi-source sensor data fusion. The system includes a data quality assessment module, a weighted fusion processing module, an environmental collaborative solution module, a confidence mapping module, and a compensation execution module, wherein:

[0049] The data quality assessment module is used to perform reliability analysis on the standardized time series data of the clock, and determine the data quality weights of the multi-source sensors in the clock within the corresponding time window based on the analysis results.

[0050] The weighted fusion processing module is used to perform weighted fusion processing on the standardized time-series data of the multi-source sensors based on the data quality weights to obtain the environmental state fusion data of the clock.

[0051] The environmental collaborative calculation module is used to perform environmental interactive collaborative calculation on the feature vector of the environmental state fusion data based on the preset environmental factor coupling relationship, so as to obtain the comprehensive compensation parameters for the timekeeping error of the clock.

[0052] The confidence mapping module is used to perform confidence mapping on the covariance matrix of the environmental state fusion data to determine the confidence index of the comprehensive compensation parameters for travel time error.

[0053] The compensation execution module is used to adjust the output frequency of the timing pulse generation unit in the clock based on the comprehensive compensation parameters for the timekeeping error when the confidence index meets the clock's timing requirements.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] 1. This invention dynamically evaluates standardized time-series data from both consistency and reliability dimensions and generates data quality weights accordingly. Then, based on these data quality weights, it performs weighted fusion processing on standardized time-series data from multiple sources of sensors. This enables the fusion process to suppress the influence of poor-quality data caused by transient impacts or local interference from individual sensors in real time, ensuring that the fused data has higher stability and reliability in the environmental state before subsequent compensation calculations. This improves the robustness of the clock timekeeping error compensation method in complex and variable environments.

[0056] 2. This invention, based on the pre-defined single-factor influence rules and cross-influence rules contained in the coupling relationship of environmental factors, performs collaborative calculation on the feature vector of the environmental state fusion data, including single-factor feature components and coupled feature components, to generate comprehensive compensation parameters for timekeeping errors. This enables the compensation process to quantify the impact of the interactive coupling effect between multiple environmental factors such as temperature, vibration, and magnetic field on the timekeeping frequency, avoiding compensation deviations caused by ignoring the mutual modulation between factors and improving compensation accuracy. Furthermore, by performing confidence mapping on the covariance matrix of the environmental state fusion data, the confidence index of the comprehensive compensation parameters for timekeeping errors is determined. The comprehensive compensation parameters for timekeeping errors are only used to adjust the output frequency of the timing pulse generation unit when the confidence index meets the watch's timing requirements. This prevents unreliable compensation parameters from intervening due to insufficient current data quality, which could cause secondary degradation of timekeeping accuracy and further ensure the reliability and safety of the compensation operation. Attached Figure Description

[0057] Figure 1 This is a flowchart illustrating a clock timekeeping error compensation method based on multi-source sensor data fusion provided in an embodiment of the present invention.

[0058] Figure 2 This is a functional block diagram of a clock timekeeping error compensation system based on multi-source sensor data fusion provided in an embodiment of the present invention;

[0059] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0060] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0061] This application provides a method for compensating for watch timekeeping errors based on multi-source sensor data fusion. The executing entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method for compensating for watch timekeeping errors based on multi-source sensor data fusion can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0062] Reference Figure 1 The diagram shown is a flowchart illustrating a clock timekeeping error compensation method based on multi-source sensor data fusion according to an embodiment of the present invention. In this embodiment, the clock timekeeping error compensation method based on multi-source sensor data fusion includes:

[0063] B1: Perform reliability analysis on the standardized time series data of the clock, and determine the data quality weights of the multi-source sensors in the clock within the corresponding time window based on the analysis results;

[0064] In this embodiment of the invention, the step of performing reliability analysis on the standardized time-series data of the clock, and determining the data quality weights of the multi-source sensors in the clock within the corresponding time window based on the analysis results, includes:

[0065] Based on the standardized time-series data of the multi-source sensors within the corresponding time window, a data sequence corresponding to the multi-source sensors is constructed.

[0066] Statistical feature extraction is performed on the data sequence to obtain the fluctuation feature values ​​corresponding to the multi-source sensors;

[0067] Based on the normalized time-series data corresponding to the multi-source sensors, a group consensus benchmark is extracted to obtain a dynamic reference sequence;

[0068] The data sequence is matched and compared with the dynamic reference sequence to obtain the deviation feature value corresponding to the multi-source sensor;

[0069] Based on the fluctuation characteristic value and the deviation characteristic value, the multi-source sensor is comprehensively evaluated to obtain the data quality weight of the multi-source sensor within the corresponding time window.

[0070] It needs to be specifically explained that the standardized time-series data of the multi-source sensors within the corresponding time window are serialized and organized. According to the timestamp alignment relationship of each sensor, the data values ​​at the same sampling time point are arranged into rows, and the data values ​​at different sampling time points are arranged into columns along the time axis, thus constructing the data sequence corresponding to each sensor. Each data sequence corresponds to the complete standardized time-series data of one sensor within the time window.

[0071] Statistical features are extracted from the data sequences corresponding to each sensor. The data dispersion of each data sequence within its time window is calculated one by one. That is, the deviations of the data values ​​at each sampling time point in each data sequence from the mean of the data in the sequence are squared and accumulated, and the average is taken within the time window. The fluctuation feature value corresponding to each sensor is obtained, which reflects the degree of data fluctuation. The larger the fluctuation feature value, the higher the instability of the sensor data within the time window.

[0072] After aligning the normalized time-series data corresponding to each sensor by timestamp, a consensus benchmark is extracted. The normalized data value set of all sensors at each sampling time point is obtained. The normalized data value set is sorted and the value in the middle position is taken as the consensus benchmark value of the sampling time point. The consensus benchmark values ​​of all sampling time points are combined sequentially along the time axis to obtain a dynamic reference sequence. The dynamic reference sequence represents the common consistency representation of the multi-source sensors at each sampling time point within the current time window.

[0073] The data sequences corresponding to each sensor are matched and compared with the dynamic reference sequence. The difference between the data value in the data sequence and the consensus benchmark value at the corresponding time point in the dynamic reference sequence is calculated at each sampling time point. The absolute value of the difference is taken and accumulated along each sampling time point within the time window and the average value is taken within the length of the time window to obtain the deviation feature value corresponding to each sensor, which reflects the degree of data deviation. The larger the deviation feature value, the higher the degree of deviation between the sensor data and the common consistency representation of the multi-source sensors.

[0074] A comprehensive evaluation is performed based on the fluctuation and deviation characteristic values ​​corresponding to each sensor. The fluctuation and deviation characteristic values ​​of the same sensor are weighted and combined by multiplying the fluctuation characteristic value with a preset fluctuation weight value to obtain a fluctuation evaluation component, and multiplying the deviation characteristic value with a preset deviation weight value to obtain a deviation evaluation component. The fluctuation evaluation component and the deviation evaluation component are then added to obtain a comprehensive evaluation value for each sensor. The comprehensive evaluation value is then subjected to reciprocal compression transformation by adding the comprehensive evaluation value with a preset benchmark constant to obtain an intermediate value. The reciprocal value of the intermediate value is taken as the data quality weight of the sensor in the corresponding time window. The larger the data quality weight, the higher the reliability of the sensor data in the current time window.

[0075] By extracting fluctuation features and evaluating deviation features from standardized time-series data from various multi-source sensors within corresponding time windows, and by weighting and comprehensively evaluating the fluctuation feature values ​​reflecting the stability of the data itself and the deviation feature values ​​reflecting the consistency of the data with the group consensus benchmark, a data quality weight value corresponding to each sensor is generated. This allows the subsequent weighted fusion processing to dynamically suppress the contribution of poor-quality data caused by transient impacts, local saturation interference, or sudden noise from individual sensors to the fusion result based on the data quality weight value. This ensures that the environmental state fusion data participating in the subsequent compensation calculation has higher reliability and anti-interference ability, thereby improving the stability of clock timekeeping error compensation from the data source.

[0076] B2: Based on the data quality weights, the standardized time-series data from the multi-source sensors are weighted and fused to obtain the environmental state fusion data of the clock.

[0077] In this embodiment of the invention, the step of performing weighted fusion processing on the standardized time-series data of the multi-source sensors based on the data quality weights to obtain the environmental state fusion data of the clock includes:

[0078] Based on the data quality weights, the standardized time series data is subjected to positional weighting to obtain the weighted time series data corresponding to the multi-source sensors;

[0079] Based on the sampling time points of the multi-source sensors, the weighted time series data is summed in pairs to obtain the fused time series data within the corresponding time window;

[0080] Environmental features are extracted from the fused time-series data to obtain the fused environmental state data of the clock within the corresponding time window.

[0081] It should be specifically explained that, based on the data quality weights of each sensor within the corresponding time window, the standardized time-series data corresponding to the same sensor is subjected to positional weighting processing. At each sampling time point, the data value of each time point in the standardized time-series data is multiplied by the data quality weight, so that the value of each sampling time point in the data sequence is scaled proportionally according to the overall reliability of the sensor, generating weighted time-series data corresponding to each sensor. The higher the quality weight of the sensor, the greater the proportion of the values ​​of each time point in the weighted time-series data that are retained, and the lower the quality weight of the sensor, the stronger the suppression of the values ​​of each time point in the weighted time-series data.

[0082] Using the shared sampling time points of all sensors as a reference, the weighted time series data is summed bitwise. At each sampling time point, the weighted data values ​​corresponding to the same time point from the weighted time series data from different sensors are accumulated, so that the effective information of the multi-source sensors at each sampling time point is converged into a single sequence channel. After the accumulation operation is completed at all sampling time points, the data is arranged in chronological order to generate the fused time series data within the corresponding time window. The fused time series data integrates the data contributions of all sensors after quality weight adjustment at each sampling time point.

[0083] Environmental features are extracted from the fused time-series data. The fused time-series data is divided into statistical segments of a preset length along the time axis. Within each statistical segment, the central tendency measure and dispersion measure of the fused time-series data in that segment are extracted. The central tendency measure values ​​of each statistical segment are arranged sequentially along the time axis to generate environmental state trend features. The dispersion measure values ​​of each statistical segment are arranged sequentially along the time axis to generate environmental state fluctuation features. The environmental state trend features and the environmental state fluctuation features are combined to obtain the environmental state fused data of the clock within the corresponding time window. The environmental state fused data contains complete feature information reflecting the overall change trend and fluctuation intensity of multiple environmental factors within the current time window.

[0084] Furthermore, the environmental state fusion data is composed of environmental state trend characteristics and environmental state fluctuation characteristics. Environmental state trend characteristics are obtained by extracting central tendency measures from the fused time-series data along the time axis and arranging them chronologically, reflecting the overall trend of changes in multiple environmental factors such as temperature, vibration, and magnetic fields within the current time window. Environmental state fluctuation characteristics are obtained by extracting dispersion measures from the fused time-series data along the time axis and arranging them chronologically, reflecting the dynamic fluctuation intensity of the fused sequence of data from each sensor within the current time window after quality weight adjustment. The environmental state fusion data carries complete environmental description information from multiple sensors within the corresponding time window after reliability screening and weighted fusion, including both the trend of coordinated changes in various environmental factors and the overall fluctuation degree of the fused data.

[0085] Based on the data quality weights of each sensor within its corresponding time window, the standardized time-series data from multiple sensors are subjected to positional weighting and summation at each sampling time point. This ensures that sensor data with high quality weights dominates the contribution of the fused time-series data, while sensor data with low quality weights is effectively suppressed. This proactively suppresses the contamination of the fusion results by poor-quality data caused by transient impacts or local saturation interference from individual sensors at the data fusion level. It ensures that the generated environmental state fusion data reliably reflects the current state characteristics of the clock's real environment at each sampling time point, providing a stable and reliable data foundation for subsequent environmental interaction and collaborative calculation.

[0086] B3: Based on the preset environmental factor coupling relationship, the feature vector of the environmental state fusion data is subjected to environmental interaction and collaborative solution to obtain the comprehensive compensation parameters for the timekeeping error of the clock.

[0087] In this embodiment of the invention, the step of performing environmental interactive collaborative calculation on the feature vector of the environmental state fusion data based on a preset environmental factor coupling relationship to obtain the comprehensive compensation parameters for the timekeeping error of the clock includes:

[0088] Multidimensional feature decomposition is performed on the environmental state fusion data to obtain the multidimensional factor feature components corresponding to the environmental state fusion data;

[0089] The multidimensional factor feature components include single-factor feature components and coupled feature components, and the preset environmental coupling relationship includes single-factor influence rules and cross-influence rules.

[0090] Based on the single-factor influence rule, independent compensation analysis is performed on the single-factor characteristic components to obtain the single-factor compensation components corresponding to each environmental factor.

[0091] Based on the cross-influence rule, the coordinated compensation amount of the coupled feature components is calculated to obtain the comprehensive timekeeping error compensation parameters of the clock.

[0092] The calculation formula for the timekeeping comprehensive error compensation parameter is as follows:

[0093] ;

[0094] In the formula, This represents the timekeeping comprehensive error compensation parameter. This represents the single-factor index set of the clock. Represented as the first of the clocks Each single-factor characteristic component , Represented as the first Single-factor influence coefficient, Represented as the set of coupled pair indices of the clock. and This is represented by the two single-factor feature components that constitute the coupling pair. , For the coupling pair The corresponding coupling effect coefficient, The coupling nonlinearity order parameter of the clock is represented as follows. This is represented as the fused environmental state data. The degree of coupling activation is represented by the statistical fluctuation characteristics of the fused environmental state data. The time constant represents the coupling activation degree of the clock.

[0095] The coupling activation degree is determined by the statistical fluctuation characteristics of the environmental state fusion data, including:

[0096] The information entropy value of the environmental state fusion data is obtained by normalizing the data sequence of the environmental state fusion data within the corresponding time window.

[0097] The temporal gradient of the data sequence is analyzed to determine the gradient intensity value of the environmental state fusion data.

[0098] By performing a weighted analysis on the information entropy value and the gradient intensity value, a comprehensive activity index of the environmental state fusion data is obtained;

[0099] The comprehensive activity index is mapped to a preset activation value range to obtain the coupled activation degree of the clock within the corresponding time window.

[0100] It is necessary to explain in detail that the environmental state fusion data is subjected to multidimensional feature decomposition. The environmental state fusion data is separated according to the different categories of environmental factors it covers, and the single-factor feature components corresponding to each category are extracted. At the same time, the data of different categories of environmental factors are paired up in a preset coupling pair combination method, and the coupling feature components reflecting the interaction strength of the two factors are extracted for each coupling pair. All single-factor feature components are combined with all coupling feature components to generate multidimensional factor feature components.

[0101] Furthermore, the coupling pair combination method specifically involves pairing any two different factors from temperature, vibration, and magnetic fields into a bidirectional pair, forming a temperature-vibration coupling pair (temperature and vibration combined), a temperature-magnetic coupling pair (temperature and magnetic field combined), and a vibration-magnetic coupling pair (vibration and magnetic field combined). Each coupling pair corresponds to a pair of single-factor feature components: the temperature-vibration coupling pair corresponds to the temperature single-factor feature component and the vibration single-factor feature component; the temperature-magnetic coupling pair corresponds to the temperature single-factor feature component and the magnetic field single-factor feature component; and the vibration-magnetic coupling pair corresponds to the vibration single-factor feature component and the magnetic field single-factor feature component. This coupling pair combination method provides a clear basis for extracting coupling feature components from the environmental state fusion data in the aforementioned steps. Following the coupling pair combination method, the interaction intensity of the corresponding two single-factor feature components is quantified pair by pair to generate the coupling feature components corresponding to each coupling pair. The preset coupling influence coefficient and coupling nonlinearity order parameter in the cross-influence rule are configured one by one according to each coupling pair in the coupling pair combination method, enabling independent quantification of the interaction effect of each coupling pair during the calculation of the collaborative compensation amount based on the coupling influence coefficient and coupling nonlinearity order parameter corresponding to each coupling pair.

[0102] The single-factor influence rule and cross-influence rule are read from the preset environmental factor coupling relationship. The single-factor influence rule includes the independent influence coefficient of each environmental factor acting alone on the clock running frequency. The cross-influence rule includes the coupling influence coefficient, coupling nonlinear order parameter and coupling activation time constant corresponding to each coupling pair.

[0103] Furthermore, the environmental factor coupling relationship is a pre-defined set of rules used to describe the independent influence of individual factors among temperature, vibration, and magnetic field factors on the clock's running frequency, as well as the cross-influence of the interactions between multiple factors on the clock's running frequency. This set of rules consists of two components: single-factor influence rules and cross-influence rules.

[0104] The single-factor influence rule specifies the independent influence coefficient of each environmental factor acting alone on the clock's running frequency. The temperature factor corresponds to the temperature influence coefficient, the vibration factor corresponds to the vibration influence coefficient, and the magnetic field factor corresponds to the magnetic field influence coefficient. Each independent influence coefficient characterizes the change in the clock's running frequency caused by a unit change in the environmental factor. This independent influence coefficient is obtained by conducting single-factor variable environmental tests on the clock and performing linear regression fitting on the test data.

[0105] The cross-influence rule specifies the coupling influence coefficient, coupling nonlinear order parameter, and coupling activation time constant for each coupling pair. The temperature-vibration coupling pair corresponds to the coupling influence coefficient under the interaction of temperature and vibration, the temperature-magnetic coupling pair corresponds to the coupling influence coefficient under the interaction of temperature and magnetic field, and the vibration-magnetic coupling pair corresponds to the coupling influence coefficient under the interaction of vibration and magnetic field. Each coupling influence coefficient characterizes the intensity of the additional influence on the clock's running frequency when the two environmental factors in the coupling pair change synergistically. The coupling nonlinear order parameter characterizes the sensitivity of the coupling effect to the relative imbalance of the two factors. The coupling activation time constant characterizes the response rate of the coupling compensation term to changes in the environmental state. The coupling influence coefficient, coupling nonlinear order parameter, and coupling activation time constant are obtained by conducting multi-factor composite environmental tests on the clock and performing multivariate nonlinear regression fitting on the test data.

[0106] Based on the independent influence coefficients in the single-factor influence rule, an independent compensation analysis is performed on each single-factor characteristic component in the multidimensional factor characteristic components. Each single-factor characteristic component is multiplied by its corresponding independent influence coefficient to obtain the single-factor compensation component corresponding to each environmental factor. Each single-factor compensation component represents the frequency compensation amount required when the environmental factor acts alone.

[0107] Based on the cross-influence rule, the coordinated compensation amount of the coupled feature components in the multi-dimensional factor feature components is converted to obtain the statistical fluctuation characteristics of the environmental state fusion data within the corresponding time window. Based on the statistical fluctuation characteristics, the coupling activation degree is determined. The ratio of the coupling activation degree to the coupling activation degree time constant is processed to obtain the activation ratio value. For each coupling pair, the absolute value of the ratio of the two single-factor feature components contained in the coupled feature components of the coupling pair is taken and then exponentially operated with the coupling nonlinear order parameter as the exponent to obtain the nonlinear coupling ratio of the coupling pair. The nonlinear coupling ratios of each coupling pair are accumulated and added to the preset benchmark value to obtain the coupling base. The logarithm of the coupling base is taken to obtain the number of coupling pairs. The activation ratio value and the number of coupling pairs are multiplied to obtain the coupling index. The coupling scaling factor is obtained by exponential operation with the natural constant term as the base and the coupling index as the exponent. All single-factor compensation components are accumulated to obtain the total linear compensation. The total linear compensation is multiplied with the coupling scaling factor to generate the time-travel comprehensive error compensation parameter.

[0108] Furthermore, the overall timekeeping error compensation parameter, as the final output, directly determines the amount of frequency correction that needs to be applied to the clock timing pulse generation unit. The sign and absolute value of the overall timekeeping error compensation parameter correspond to the direction and magnitude of the frequency correction, respectively.

[0109] The single-factor index set reads all single environmental factor categories participating in the compensation calculation from the preset environmental factor coupling relationship, and combines the identifiers of these environmental factor categories to form the single-factor index set. The single-factor index set includes temperature factor identifiers, vibration factor identifiers, and magnetic field factor identifiers. The single-factor index set limits the range of factors participating in the single-factor linear compensation calculation, and the number of elements in the single-factor index set determines the number of independent compensation terms included in the total linear compensation.

[0110] Indicates the first Each single-factor feature component is extracted by performing multidimensional feature decomposition on the fused environmental state data, and then extracting the single-factor feature components corresponding to each environmental factor for each category. For example... For temperature as a single-factor characteristic component, For vibration single-factor characteristic components, This is a single-factor characteristic component of the magnetic field. The absolute value of the factor directly determines its contribution to single-factor compensation. A positive value indicates that this factor has a positive offset effect on the travel frequency. A negative value indicates that a negative offset effect is applied.

[0111] Indicates the first The single-factor influence coefficient is read from the single-factor influence rule in the pre-set coupling relationship of environmental factors. This coefficient is pre-determined by conducting single-factor variable environmental experiments on the clock and performing linear regression fitting on the experimental data. The decision was made The influence weight of each single-factor feature component on the total linear compensation. The symbol indicates the direction of compensation. The absolute value represents the frequency compensation amount corresponding to a unit change in characteristic component.

[0112] in Representing the total linear compensation, by traversing each single-factor identifier i in the single-factor index set, each single-factor feature component is... Its corresponding single-factor influence coefficient Multiplication is performed to obtain the individual factor compensation components. All individual factor compensation components are then summed to obtain the total linear compensation. The total linear compensation constitutes the base part of the timekeeping comprehensive error compensation parameter, which represents the normal compensation requirement when each environmental factor acts independently. When the total linear compensation is zero, the calculation result of the entire formula is zero, and no compensation is applied at this time.

[0113] The coupling pair index set is read from the cross-influence rules in the preset environmental factor coupling relationship. It contains the identification pairs of each pair of environmental factors participating in the coupling solution. The coupling pair index set contains temperature and vibration combination identification pairs, temperature and magnetic field combination identification pairs, and vibration and magnetic field combination identification pairs. The coupling pair index set limits the range of factor combinations participating in the coupling compensation calculation. The number of elements in the coupling pair index set determines the number of coupling pair terms included in the nonlinear coupling ratio accumulation term.

[0114] and For each single-factor feature component constituting a coupling pair, in the coupling pair index set, the identifier is... Extract the single-factor feature components corresponding to the two environmental factors in the coupling pair from all extracted single-factor feature components. and , and The ratio reflects the relative strength of the two environmental factors in the coupling pair. This ratio determines the contribution of the coupling pair to the nonlinear coupling ratio. The greater the deviation of the ratio from the equilibrium, the more significant the coupling effect of the coupling pair.

[0115] The coupling nonlinear order parameter is read from the cross-influence rules in the preset environmental factor coupling relationship. It is pre-determined by conducting multi-factor composite environmental tests on the clock and performing multivariate nonlinear regression fitting on the test data. The coupling nonlinear order parameter controls the nonlinear rate of change of the coupling effect with the degree of imbalance between factors. When the coupling nonlinear order parameter is greater than 1, the coupling effect shows an accelerating growth trend with the increase of the degree of imbalance of factors. When the coupling nonlinear order parameter is less than 1, it shows a decelerating growth trend. When the coupling nonlinear order parameter is equal to 1, it shows a linear growth trend. The value of the coupling nonlinear order parameter directly affects the compensation strength of the coupling compensation term in the high imbalance scenario.

[0116] The absolute value of the ratio of the factors in the coupling pair is exponentially raised to the nonlinear order parameter of the coupling pair. This exponentiation maps the linear ratio relationship between the factors to a nonlinear ratio relationship, allowing the quantification of the coupling effect to more flexibly match the nonlinear law of the interaction of factors with the degree of imbalance in the actual physical process. The higher the nonlinear order parameter of the coupling pair, the faster the marginal contribution of the coupling pair to the overall coupling effect increases with the degree of imbalance.

[0117] The coupling effect coefficient iterates through each coupling pair in the coupling pair index set, multiplying the coupling effect coefficient of each coupling pair by the nonlinear coupling ratio of that pair to obtain the coupling effect contribution value of each coupling pair. Then, the coupling effect contribution values ​​of all coupling pairs are summed. The cumulative coupling effect integrates the total coupling effect strength of all preset coupling pairs under the current environmental state. The larger this value, the more significant the additional impact of multi-factor interactions on the travel frequency under the current environmental state, and the greater the amount of additional coupling compensation required.

[0118] The environmental state fusion data indirectly affects the coupling activation degree through its statistical fluctuation characteristics. The more drastic the data fluctuation, the higher the coupling activation degree, and the more stable the data, the lower the coupling activation degree, so that the degree of coupling compensation intervention matches the actual complexity of the environment.

[0119] The coupling activation degree is determined based on the statistical fluctuation characteristics of the fused environmental state data within the corresponding time window. Specifically, after normalizing the data sequence, the information entropy value is calculated; after analyzing the degree of change of the temporal gradient of the data sequence, the gradient strength value is calculated; and a weighted analysis of the information entropy value and the gradient strength value is performed to obtain a comprehensive activity index. This comprehensive activity index is then mapped to a preset activation value range. The coupling activation degree falls between a preset lower limit and an upper limit. A higher coupling activation degree indicates a higher degree of complexity and volatility in the current environmental state, and a stronger activation of the coupling compensation term. When the coupling activation degree approaches the lower limit, the coupling compensation term is basically suppressed, and the time-travel comprehensive error compensation parameter is approximately composed only of the total linear compensation amount.

[0120] The coupling activation time constant is read from the cross-influence rules in the preset environmental factor coupling relationship, and is predetermined by conducting multi-factor composite environmental tests on the clock and performing multivariate nonlinear regression fitting on the test data. Controlling the response rate of coupling activation degree to the intervention depth of coupling compensation term, The larger the value, the smaller the coupling activation ratio for the same coupling activation degree, the smoother the coupling compensation intervention, and the smoother the transition of the formula from pure linear compensation to coupling compensation when the environment fluctuates. The smaller the value, the more sensitive the coupling compensation is to environmental changes, and the faster the formula switches when the environmental state changes.

[0121] The environmental state fusion data sequence within the corresponding time window is normalized to compress each data value into the same range, resulting in a normalized data sequence. The frequency of each numerical interval in the normalized data sequence is counted, and the ratio of the frequency of each numerical interval to the total number of data points within the corresponding time window is taken as the probability of occurrence of each numerical interval. The product of the probability of occurrence of each numerical interval and its corresponding value is accumulated and the negative value is taken to obtain the information entropy value of the environmental state fusion data. The larger the information entropy value, the higher the degree of disorder of the environmental state fusion data within the time window.

[0122] The temporal gradient of the data sequence is analyzed. The data values ​​in the data sequence are differentially processed along the time axis for each adjacent sampling time point to obtain the change amplitude value between each adjacent sampling time point. The absolute values ​​of all change amplitude values ​​are accumulated and divided by the total number of intervals between adjacent points within the time window to obtain the gradient intensity value of the environmental state fusion data. The larger the gradient intensity value, the more drastic the change of the environmental state fusion data within the time window.

[0123] The information entropy value and gradient strength value are weighted and analyzed. The information entropy value is multiplied by the preset information entropy weight coefficient to obtain the information entropy contribution component. The gradient strength value is multiplied by the preset gradient weight coefficient to obtain the gradient contribution component. The information entropy contribution component and the gradient contribution component are added together to obtain the comprehensive activity index of the environmental state fusion data.

[0124] The overall activity index is mapped to a preset activation value range, which has a lower and upper limit. When the overall activity index is below the preset lower threshold, the coupled activation value is set to the lower limit of the activation value range. When the overall activity index is above the preset upper threshold, the coupled activation value is set to the upper limit of the activation value range. When the overall activity index is between the preset lower and upper thresholds, the difference between the overall activity index and the lower threshold is compared to the difference between the upper and lower thresholds. The result of this ratio is multiplied by the range of the activation value range and then added to the lower limit of the activation value range to obtain the coupled activation value of the clock within the corresponding time window.

[0125] By simultaneously extracting single-factor feature components and coupled feature components from environmental state fusion data, and performing independent compensation analysis and collaborative compensation calculation based on the preset single-factor influence rules and cross-influence rules in the coupling relationship of environmental factors, the final generated comprehensive timekeeping error compensation parameters not only include the conventional compensation quantities when each environmental factor acts independently, but also quantify the additional influence of the interactive coupling effects between multiple factors such as temperature and vibration, temperature and magnetic field, and vibration and magnetic field on the timekeeping frequency. This overcomes the shortcomings of traditional schemes that simply superimpose the compensation quantities of each factor while ignoring the mutual modulation between factors, and significantly improves the accuracy of timekeeping error compensation of clocks in multi-factor composite environments. By adjusting the scaling ratio of the coupling compensation term to the total linear compensation by the coupling activation degree determined by the statistical fluctuation characteristics of the environmental state fusion data and the preset coupling activation degree time constant, the coupling compensation term automatically degenerates into linear compensation when the environment is stable, and the coupling compensation fully intervenes when the environment changes drastically. This achieves adaptive matching for environments of different complexity, avoids excessive compensation due to over-introduction of coupling terms when the environment is stable, and avoids insufficient compensation due to insufficient intervention of coupling terms when the environment is complex. This further improves the adaptability and stability of the time travel error compensation method under variable environmental conditions.

[0126] B4: Perform confidence mapping on the covariance matrix of the environmental state fusion data to determine the confidence index of the comprehensive compensation parameter for travel time error;

[0127] In this embodiment of the invention, the step of performing confidence mapping on the covariance matrix of the environmental state fusion data to determine the confidence index of the comprehensive compensation parameters for travel time errors includes:

[0128] Construct the covariance matrix of the environmental state fusion data within the corresponding time window;

[0129] The diagonal elements of the covariance matrix represent the data autovariance of each environmental factor, and the off-diagonal elements represent the data covariance between different environmental factors.

[0130] Based on the eigenvalue set of the covariance matrix, the discreteness analysis of the environmental state fusion data is performed to obtain the uncertainty measure of the environmental state fusion data.

[0131] The confidence index of the time error comprehensive compensation parameter is obtained by performing a confidence mapping evaluation on the uncertainty metric value.

[0132] The confidence index for obtaining the comprehensive compensation parameter of the time travel error by performing confidence mapping evaluation on the uncertainty metric value includes:

[0133] The trace of the covariance matrix is ​​used as the total variance of the environmental state fusion data;

[0134] The largest eigenvalue in the set of eigenvalues ​​is used as the main direction discrete measure of the environmental state fusion data;

[0135] The ratio of the total variance to the main direction discreteness metric is processed to obtain the structural stability ratio of the environmental state fusion data.

[0136] Based on the structural stability ratio, the uncertainty metric is weighted and corrected, and the correction result is mapped to a preset confidence interval to obtain the confidence index of the comprehensive compensation parameter for the time travel error.

[0137] Specifically, the covariance matrix of the environmental state fusion data within the corresponding time window is constructed by treating the data sequences corresponding to each environmental factor in the environmental state fusion data as observation sequences of multidimensional random variables. For each pair of environmental factor data sequences, the average of the product of the deviations of their data values ​​from their respective means within the time window is calculated. The average of the squares of the deviations of the data values ​​of each environmental factor's own data sequence from its own mean within the time window is used as the diagonal element. The average of the products of the deviations of the data values ​​of two different environmental factor data sequences from their respective means within the time window is used as the off-diagonal element. All diagonal and off-diagonal elements are arranged into a square matrix according to the order of environmental factors to obtain the covariance matrix of the environmental state fusion data.

[0138] Eigenvalue decomposition is performed on the covariance matrix, which is decomposed according to the direction of its eigenvectors and the magnitude of its eigenvalues. All eigenvalues ​​and their corresponding eigenvectors of the covariance matrix are obtained. All eigenvalues ​​are arranged in descending order of value and combined to form an eigenvalue set. Each eigenvalue represents the variance of the environmental state fusion data in the direction of the corresponding eigenvector. The larger the eigenvalue, the higher the dispersion of the data in that direction.

[0139] Discreteness analysis is performed on environmental state fusion data based on the feature value set. All feature values ​​in the feature value set are obtained, and all feature values ​​are summed to obtain the feature value sum. The feature value mean is obtained by dividing the feature value sum by the number of feature values ​​in the feature value set. The feature value mean is used as the uncertainty measure of environmental state fusion data. The larger the uncertainty measure value, the higher the overall data dispersion of environmental state fusion data within the current time window and the stronger the noise component contained in the data.

[0140] Obtain all diagonal elements of the covariance matrix, sum the diagonal elements to obtain the trace of the covariance matrix, and use the trace of the covariance matrix as the total variance of the environmental state fusion data. The total variance represents the sum of the data variances of all environmental factors in the environmental state fusion data. The larger the total variance, the greater the sum of the independent fluctuations of each environmental factor.

[0141] The largest eigenvalue is taken from the set of eigenvalues ​​and used as the main direction dispersion measure of the environmental state fusion data. The main direction dispersion measure characterizes the variance strength of the environmental state fusion data in the direction of the greatest change. The larger the main direction dispersion measure, the more prominent the concentration dispersion trend of the data in a certain dominant direction.

[0142] The ratio of the total variance to the dispersion measure of the main direction is processed. The ratio obtained by dividing the total variance by the dispersion measure of the main direction is used as the structural stability ratio of the environmental state fusion data. The closer the structural stability ratio is to the total number of environmental factors, the more uniform the data dispersion structure is distributed in all directions. The farther the structural stability ratio deviates from the total number of environmental factors, the more concentrated the data dispersion is in a certain dominant direction.

[0143] The uncertainty metric is weighted and corrected based on the structural stability ratio. The absolute value of the difference between the structural stability ratio and the total number of environmental factors is taken to obtain the stability deviation. The stability deviation is multiplied by a preset deviation penalty coefficient to obtain the correction weight. The uncertainty metric is multiplied by the correction weight to obtain the corrected uncertainty metric. The corrected uncertainty metric integrates information on both the overall dispersion of the data and the structural stability of the data distribution.

[0144] The corrected uncertainty metric is mapped to a preset confidence interval, which has an upper and lower limit. When the corrected uncertainty metric is greater than the upper limit of the preset confidence interval, the confidence index is taken as the lower limit of the confidence interval. When the corrected uncertainty metric is less than the lower limit of the preset confidence interval, the confidence index is taken as the upper limit of the confidence interval. When the corrected uncertainty metric is between the lower and upper limits of the preset confidence interval, the upper limit of the confidence interval and the result of the corrected uncertainty metric normalized by the upper limit threshold are multiplied to obtain the confidence index of the time travel error comprehensive compensation parameter. The higher the value of the confidence index, the higher the reliability of the time travel error comprehensive compensation parameter within the current time window.

[0145] By constructing a covariance matrix of the fused environmental state data and performing eigenvalue decomposition on the covariance matrix, both the overall dispersion of the data and the stability of the data distribution structure are incorporated into the confidence assessment system. This allows the generated confidence index to not only reflect the noise level of the fused data but also identify the risk of multi-factor imbalance caused by excessive concentration of data dispersion in a certain dominant direction. This provides a reliable basis for judging whether the comprehensive compensation parameters for timekeeping errors are suitable for the current time window. By setting the confidence index as a prerequisite for the effectiveness of the compensation parameters, when the quality of the fused environmental state data deteriorates due to short-term sensor failure or sudden strong interference, the confidence index automatically decreases, preventing unreliable compensation parameters from being applied to the timing pulse generation unit. This avoids introducing additional timekeeping deviations due to poor compensation, ensuring the safety and stability of the watch's timekeeping error compensation operation in complex and changing environments.

[0146] B5: When the confidence index meets the timekeeping requirements of the clock, the output frequency of the timekeeping pulse generation unit in the clock is adjusted based on the comprehensive compensation parameters for the timekeeping error.

[0147] In this embodiment of the invention, when the confidence index meets the timekeeping requirements of the clock, adjusting the output frequency of the timing pulse generation unit in the clock based on the timekeeping error comprehensive compensation parameter includes:

[0148] The confidence index is compared with the confidence requirement threshold corresponding to the preset watch timing requirement. When the confidence index reaches the confidence requirement threshold, it is determined that the timekeeping error comprehensive compensation parameter meets the watch timing requirement.

[0149] When it is determined that the timekeeping requirement of the clock is met, the frequency control dimension analysis is performed on the comprehensive compensation parameter for the timekeeping error to obtain the frequency adjustment amount that matches the timing pulse generation unit.

[0150] The frequency adjustment amount is written into the frequency control register of the timing pulse generation unit to fine-tune the oscillation frequency of the timing pulse generation unit, thereby obtaining the corrected timing pulse signal of the clock.

[0151] It should be specifically explained that the confidence index is compared with the confidence requirement threshold corresponding to the preset timekeeping requirements of the clock. The confidence requirement threshold is pre-stored in the non-volatile memory of the clock and is read each time a compensation judgment is performed. When the value of the confidence index is greater than or equal to the confidence requirement threshold, it is determined that the comprehensive compensation parameter for the timekeeping error meets the timekeeping requirements of the clock within the current time window, and a compensation enable signal is generated. When the value of the confidence index is less than the confidence requirement threshold, it is determined that the comprehensive compensation parameter for the timekeeping error does not meet the timekeeping requirements of the clock within the current time window, a compensation prohibition signal is generated, and the compensation operation is not performed.

[0152] When the compensation enable signal is generated, the frequency control dimension is analyzed on the comprehensive compensation parameters for timekeeping error. The relative frequency change value corresponding to the frequency correction is extracted from the comprehensive compensation parameters for timekeeping error. The nominal oscillation frequency value of the timing pulse generation unit is read. The relative frequency change value is multiplied by the nominal oscillation frequency value to obtain the absolute frequency offset. Then, the bit width and frequency adjustment resolution of the frequency control register of the timing pulse generation unit are read. The absolute frequency offset is divided by the frequency adjustment resolution to obtain the frequency control adjustment step number. The frequency control adjustment step number is quantized into a binary frequency adjustment amount that matches the bit width of the frequency control register.

[0153] The binary frequency adjustment is written into the frequency control register of the timing pulse generation unit through the internal bus interface of the clock. After receiving the binary frequency adjustment, the frequency control register updates the frequency control word stored inside. The numerically controlled oscillator inside the timing pulse generation unit adjusts the connection state of its internal capacitor array or current source according to the updated frequency control word, changes the resonance parameters of the numerically controlled oscillator, and fine-tunes the oscillation frequency generated by the numerically controlled oscillator so that the frequency of the timing pulse signal output by the timing pulse generation unit approaches the target frequency value after compensation and correction, thus obtaining the timing pulse signal corrected by the clock.

[0154] By setting a comparison mechanism between the confidence index and the confidence requirement threshold as a precondition for compensation execution, the frequency adjustment operation is triggered only when the confidence level of the comprehensive compensation parameters for timekeeping error reaches the preset timing requirement standard. This prevents unreliable compensation parameters from being written into the timing pulse generation unit due to deterioration in the quality of environmental condition fusion data, thus ensuring the safety of the compensation operation. By performing frequency control dimension analysis on the comprehensive compensation parameters for timekeeping error and writing the obtained frequency adjustment amount into the frequency control register through the bus interface, the internal resonance parameters of the numerically controlled oscillator are directly adjusted. This allows the compensation correction to be accurately transmitted digitally to the source of the timing pulse signal, eliminating the transmission error between the compensation parameters and the frequency correction, and ensuring high-precision tracking consistency between the actual output frequency of the timing pulse signal and the target compensation frequency calculated through collaborative solution.

[0155] like Figure 2 The diagram shown is a functional block diagram of a clock timekeeping error compensation system based on multi-source sensor data fusion provided in an embodiment of the present invention.

[0156] The clock timekeeping error compensation system 100 based on multi-source sensor data fusion described in this invention can be installed in an electronic device. Depending on the functions implemented, the clock timekeeping error compensation system 100 may include a data quality assessment module 101, a weighted fusion processing module 102, an environmental collaborative solution module 103, a confidence mapping module 104, and a compensation execution module 105. The module described in this invention can also be referred to as a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0157] In this embodiment, the functions of each module / unit are as follows:

[0158] The data quality assessment module 101 is used to perform reliability analysis on the standardized time series data of the clock and determine the data quality weights of the multi-source sensors in the clock within the corresponding time window based on the analysis results.

[0159] The weighted fusion processing module 102 is used to perform weighted fusion processing on the standardized time series data of the multi-source sensors based on the data quality weights to obtain the environmental state fusion data of the clock.

[0160] The environmental collaborative calculation module 103 is used to perform environmental interactive collaborative calculation on the feature vector of the environmental state fusion data based on the preset environmental factor coupling relationship, so as to obtain the comprehensive compensation parameters for the timekeeping error of the clock.

[0161] The confidence mapping module 104 is used to perform confidence mapping on the covariance matrix of the environmental state fusion data to determine the confidence index of the travel time error comprehensive compensation parameter.

[0162] The compensation execution module 105 is used to adjust the output frequency of the timing pulse generation unit in the clock based on the comprehensive compensation parameters for the timekeeping error when the confidence index meets the clock's timing requirements.

[0163] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0164] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0165] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0166] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0167] The embodiments of this application can acquire and process relevant data based on an artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0168] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A clock timekeeping error compensation method based on multi-source sensor data fusion, characterized in that, The method includes: B1: Perform reliability analysis on the standardized time series data of the clock, and determine the data quality weights of the multi-source sensors in the clock within the corresponding time window based on the analysis results; B2: Based on the data quality weights, the standardized time-series data from the multi-source sensors are weighted and fused to obtain the environmental state fusion data of the clock. B3: Based on the preset environmental factor coupling relationship, the feature vector of the environmental state fusion data is subjected to environmental interactive collaborative solution to obtain the comprehensive compensation parameters for the timekeeping error of the clock, including: Multidimensional feature decomposition is performed on the environmental state fusion data to obtain the multidimensional factor feature components corresponding to the environmental state fusion data; The multidimensional factor feature components include single-factor feature components and coupled feature components, and the preset environmental coupling relationship includes single-factor influence rules and cross-influence rules. Based on the single-factor influence rule, independent compensation analysis is performed on the single-factor characteristic components to obtain the single-factor compensation components corresponding to each environmental factor. Based on the cross-influence rule, the coordinated compensation amount of the coupled feature components is calculated to obtain the comprehensive timekeeping error compensation parameters of the clock. The calculation formula for the timekeeping comprehensive error compensation parameter is as follows: ； In the formula, This represents the timekeeping comprehensive error compensation parameter. This represents the single-factor index set of the clock. Represented as the first of the clocks Each single-factor characteristic component , Represented as the first Single-factor influence coefficient, Represented as the set of coupled pair indices of the clock. and This is represented by the two single-factor feature components that constitute the coupling pair. , For the coupling pair The corresponding coupling effect coefficient, The coupling nonlinearity order parameter of the clock is represented as follows. This is represented as the fused environmental state data. The degree of coupling activation is represented by the statistical fluctuation characteristics of the fused environmental state data. This represents the coupling activation time constant of the clock; The information entropy value of the environmental state fusion data is obtained by normalizing the data sequence of the environmental state fusion data within the corresponding time window. The temporal gradient of the data sequence is analyzed to determine the gradient intensity value of the environmental state fusion data. By performing a weighted analysis on the information entropy value and the gradient intensity value, a comprehensive activity index of the environmental state fusion data is obtained; The comprehensive activity index is mapped to a preset activation value range to obtain the coupled activation degree of the clock within the corresponding time window; B4: Perform confidence mapping on the covariance matrix of the environmental state fusion data to determine the confidence index of the comprehensive compensation parameter for travel time error; B5: When the confidence index meets the timekeeping requirements of the clock, the output frequency of the timekeeping pulse generation unit in the clock is adjusted based on the comprehensive compensation parameters for the timekeeping error.

2. The clock timekeeping error compensation method based on multi-source sensor data fusion as described in claim 1, characterized in that, The reliability analysis of the standardized time-series data of the clock, and the determination of the data quality weights of the multi-source sensors in the clock within the corresponding time window based on the analysis results, includes: Based on the standardized time-series data of the multi-source sensors within the corresponding time window, a data sequence corresponding to the multi-source sensors is constructed. Statistical feature extraction is performed on the data sequence to obtain the fluctuation feature values ​​corresponding to the multi-source sensors; Based on the normalized time-series data corresponding to the multi-source sensors, a group consensus benchmark is extracted to obtain a dynamic reference sequence; The data sequence is matched and compared with the dynamic reference sequence to obtain the deviation feature value corresponding to the multi-source sensor; Based on the fluctuation characteristic value and the deviation characteristic value, the multi-source sensor is comprehensively evaluated to obtain the data quality weight of the multi-source sensor within the corresponding time window.

3. The clock timekeeping error compensation method based on multi-source sensor data fusion as described in claim 1, characterized in that, The step of performing weighted fusion processing on the standardized time-series data of the multi-source sensors based on the data quality weights to obtain the environmental state fusion data of the clock includes: Based on the data quality weights, the standardized time series data is subjected to positional weighting to obtain the weighted time series data corresponding to the multi-source sensors; Based on the sampling time points of the multi-source sensors, the weighted time series data is summed in pairs to obtain the fused time series data within the corresponding time window; Environmental features are extracted from the fused time-series data to obtain the fused environmental state data of the clock within the corresponding time window.

4. The clock timekeeping error compensation method based on multi-source sensor data fusion as described in claim 1, characterized in that, The step of performing confidence mapping on the covariance matrix of the fused environmental state data to determine the confidence index of the comprehensive compensation parameters for travel time errors includes: Construct the covariance matrix of the environmental state fusion data within the corresponding time window; The diagonal elements of the covariance matrix represent the data autovariance of each environmental factor, and the off-diagonal elements represent the data covariance between different environmental factors. Based on the eigenvalue set of the covariance matrix, the discreteness analysis of the environmental state fusion data is performed to obtain the uncertainty measure of the environmental state fusion data. The confidence index of the time error comprehensive compensation parameter is obtained by performing a confidence mapping evaluation on the uncertainty metric value.

5. The clock timekeeping error compensation method based on multi-source sensor data fusion as described in claim 4, characterized in that, The confidence index for obtaining the comprehensive compensation parameter of the time travel error by performing confidence mapping evaluation on the uncertainty metric value includes: The trace of the covariance matrix is ​​used as the total variance of the environmental state fusion data; The largest eigenvalue in the set of eigenvalues ​​is used as the main direction discrete measure of the environmental state fusion data; The ratio of the total variance to the main direction discreteness metric is processed to obtain the structural stability ratio of the environmental state fusion data. Based on the structural stability ratio, the uncertainty metric is weighted and corrected, and the correction result is mapped to a preset confidence interval to obtain the confidence index of the comprehensive compensation parameter for the time travel error.

6. The clock timekeeping error compensation method based on multi-source sensor data fusion as described in claim 1, characterized in that, When the confidence index meets the timekeeping requirements of the clock, the output frequency of the timekeeping pulse generation unit in the clock is adjusted based on the comprehensive compensation parameters for the timekeeping error, including: The confidence index is compared with the confidence requirement threshold corresponding to the preset watch timing requirement. When the confidence index reaches the confidence requirement threshold, it is determined that the timekeeping error comprehensive compensation parameter meets the watch timing requirement. When it is determined that the timekeeping requirement of the clock is met, the frequency control dimension analysis is performed on the comprehensive compensation parameter for the timekeeping error to obtain the frequency adjustment amount that matches the timing pulse generation unit. The frequency adjustment amount is written into the frequency control register of the timing pulse generation unit to fine-tune the oscillation frequency of the timing pulse generation unit, thereby obtaining the corrected timing pulse signal of the clock.

7. A clock timekeeping error compensation system based on multi-source sensor data fusion, characterized in that, To implement the clock timekeeping error compensation method based on multi-source sensor data fusion as described in claim 1, the system includes a data quality assessment module, a weighted fusion processing module, an environmental collaborative solution module, a confidence mapping module, and a compensation execution module, wherein: The data quality assessment module is used to perform reliability analysis on the standardized time series data of the clock, and determine the data quality weights of the multi-source sensors in the clock within the corresponding time window based on the analysis results. The weighted fusion processing module is used to perform weighted fusion processing on the standardized time-series data of the multi-source sensors based on the data quality weights to obtain the environmental state fusion data of the clock. The environmental collaborative calculation module is used to perform environmental interactive collaborative calculation on the feature vector of the environmental state fusion data based on the preset environmental factor coupling relationship, so as to obtain the comprehensive compensation parameters for the timekeeping error of the clock. The confidence mapping module is used to perform confidence mapping on the covariance matrix of the environmental state fusion data to determine the confidence index of the time travel error comprehensive compensation parameter. The compensation execution module is used to adjust the output frequency of the timing pulse generation unit in the clock based on the comprehensive compensation parameters for the timekeeping error when the confidence index meets the clock's timing requirements.

Images