Calibration data correction method and device, electronic equipment and storage medium
By identifying operating conditions and extracting multi-dimensional features, combined with data integrity and noise level assessment, and dynamically adjusting algorithm weights, the problem of poor vehicle calibration data correction effect is solved, achieving higher accuracy and robustness in calibration data correction.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHONGQING JINKANG NEW ENERGY VEHICLE CO LTD
- Filing Date
- 2026-01-05
- Publication Date
- 2026-04-14
AI Technical Summary
In the existing technology, the calibration data correction method of vehicle control system cannot adapt to the changing driving environment, resulting in limited correction effect, poor robustness, and inability to meet the complex requirements of various working conditions.
By integrating comprehensive working condition identification, multi-dimensional feature extraction and adaptive algorithm fusion, the calibration data is dynamically corrected. This includes determining the working condition type, extracting time-domain and frequency-domain features, dynamically evaluating algorithm weights based on data integrity and noise levels, and constructing an adaptive algorithm combination for correction.
It significantly improves the accuracy and robustness of calibration data correction, enabling it to better adapt to dynamic operating conditions and enhance overall performance and environmental adaptability.
Smart Images
Figure CN121858910A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of vehicle technology, and specifically to a method, apparatus, electronic device, and storage medium for correcting calibration data. Background Technology
[0002] Vehicle control systems (such as engine, transmission, and battery management systems) rely on a large amount of pre-set calibration data (such as MAP diagrams and lookup table values) to ensure their performance, efficiency, and emissions under different operating conditions. Related technologies use static filtering parameters and fixed algorithm strategies to correct the calibration data. However, during vehicle operation, the variable driving environment leads to poor adaptability of the corrected calibration data, making it difficult to meet the combined needs of various operating conditions, resulting in limited correction effectiveness and poor robustness. Summary of the Invention
[0003] In view of the above problems, this application provides a method, apparatus, electronic device and storage medium for correcting calibration data. By integrating comprehensive working condition identification, multi-dimensional feature extraction and adaptive algorithm fusion, dynamic, accurate and robust online correction of the original calibration data is achieved.
[0004] According to one aspect of this application, a method for correcting calibration data is provided. The method includes: determining the operating condition type corresponding to the current operating condition data of a vehicle, and extracting the time-domain features and frequency-domain features of the current operating condition data; determining the current adaptation algorithm according to the operating condition type, and determining the weights of each current adaptation algorithm according to the time-domain features, the frequency-domain features, data integrity, and data noise, so as to obtain a combination of current adaptation algorithms; and using the combination of current adaptation algorithms to correct the original calibration data to obtain corrected calibration data.
[0005] In one optional approach, determining the operating condition type corresponding to the vehicle's current operating condition data includes: calculating the similarity between the vehicle's current operating condition data and preset operating condition data to obtain the similarity corresponding to each preset operating condition data; and taking the preset operating condition type corresponding to the preset operating condition data with a similarity greater than the preset similarity as the operating condition type corresponding to the current operating condition data.
[0006] In one optional approach, the weights of each current adaptation algorithm are determined based on the time-domain features, the frequency-domain features, data integrity, and data noise levels. This includes: determining the data integrity of the current operating data based on the missing rate of the current operating data, and determining the data noise level of the current operating data based on the signal-to-noise ratio of the current operating data; performing quality evaluations on each current adaptation algorithm based on the time-domain features, the frequency-domain features, the data integrity, and the data noise levels to obtain a quality score for each current adaptation algorithm; and determining the corresponding weights for each current adaptation algorithm based on its quality score.
[0007] In one optional approach, determining the data integrity of the current operating condition data based on the missing rate of the current operating condition data includes: calculating the matching rate of each operating condition data in the current operating condition data with the corresponding preset operating condition data to obtain the matching rate corresponding to each operating condition data; identifying operating condition data with a matching rate less than the preset matching rate as abnormal data; and determining the data integrity of the current operating condition data based on the number of operating condition data in the current operating condition data and the number of abnormal data.
[0008] In one optional approach, the correction method further includes: if the quality score of the target adaptation algorithm is lower than a preset score, adjusting the weights corresponding to the target adaptation algorithm to obtain an adjusted current adaptation algorithm combination; wherein the target adaptation algorithm is any algorithm among the current adaptation algorithms.
[0009] In one alternative approach, the weights of each current adaptation algorithm are determined based on their respective quality scores, including: summing the quality scores of each current adaptation algorithm to obtain a total quality score; and using the ratio between the quality score of each current adaptation algorithm and the total quality score as their respective weights.
[0010] In one alternative approach, the original calibration data is corrected using the current adaptation algorithm combination to obtain corrected calibration data. This includes: inputting the original calibration data into each current adaptation algorithm in the current adaptation algorithm combination so that each current adaptation algorithm outputs its corresponding corrected data; weighting and fusing the corrected data; and using the fused corrected data as the corrected calibration data.
[0011] According to another aspect of this application, a calibration data correction device is provided, the correction device comprising: a working condition identification module, configured to determine the working condition type corresponding to the current working condition data of the vehicle, and extract the time domain features and frequency domain features of the current working condition data; an algorithm strategy determination module, configured to determine the current adaptation algorithm according to the working condition type, and determine the weight of each current adaptation algorithm according to the time domain features, the frequency domain features, data integrity and data noise, so as to obtain a current adaptation algorithm combination; and a correction module, configured to correct the original calibration data using the current adaptation algorithm combination to obtain corrected calibration data.
[0012] According to one aspect of this application, an electronic device is provided, comprising: a controller; and a memory for storing one or more programs, which, when executed by the controller, perform the modified method described above.
[0013] According to one aspect of this application, a computer-readable storage medium is also provided, on which computer-readable instructions are stored, which, when executed by a computer's processor, cause the computer to perform the above-described correction method.
[0014] According to one aspect of this application, a computer program product or computer program is also provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the modified method described above.
[0015] This application extracts time-domain and frequency-domain features from the vehicle's current operating condition data and dynamically evaluates the applicability of each algorithm based on data integrity and noise level. It realizes the construction of an adaptive algorithm combination based on operating condition type, which solves the problems that static filtering strategies cannot adapt to dynamic operating conditions and that module isolation leads to overall performance degradation. This significantly improves the correction accuracy, robustness and environmental adaptability of calibration data.
[0016] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description
[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.
[0018] Figure 1 This is a flowchart illustrating an exemplary embodiment of a calibration data correction method.
[0019] Figure 2 Based on Figure 1 The exemplary embodiment shown illustrates a flowchart of another method for correcting calibration data.
[0020] Figure 3 This is a schematic diagram of the structure of a calibration data correction device shown in an exemplary embodiment of this application.
[0021] Figure 4 This is a schematic diagram of the structure of a computer system for an electronic device illustrated in an exemplary embodiment of this application. Detailed Implementation
[0022] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.
[0023] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.
[0024] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.
[0025] In this application, "multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.
[0026] Existing technologies for correcting vehicle calibration data typically employ static filtering parameters and fixed algorithm strategies, failing to adequately consider the dynamic changes in vehicle operating conditions during operation. Furthermore, the modules for operating condition identification, filtering, and quality assessment in these technologies often operate independently, lacking a collaborative mechanism. This prevents adjustments to the filtering strategy based on real-time data quality feedback, resulting in limited overall performance. Especially when data is missing or subject to high noise interference, forcibly using unsuitable algorithms further reduces the reliability of the output data.
[0027] To address this, one aspect of this application provides a method for correcting calibration data. This method integrates comprehensive operating condition identification, multi-dimensional feature extraction, and adaptive algorithms to achieve dynamic, accurate, and robust online correction of the original calibration data. Please refer to the details below. Figure 1 , Figure 1 This is a flowchart illustrating an exemplary embodiment of a calibration data correction method. The correction method includes at least steps S110 to S130, which are described in detail below: S110: Determine the operating condition type corresponding to the current operating condition data of the vehicle, and extract the time domain features and frequency domain features of the current operating condition data.
[0028] This embodiment analyzes real-time multi-source sensor data collected by the vehicle (such as vehicle speed, acceleration, engine speed, throttle opening, braking status, etc.) to determine its typical driving mode, such as idling, low-speed crawling, frequent acceleration and deceleration on urban roads, constant speed driving on highways, and driving on mountain slopes. This process can be a data matching process, where the current operating condition data is matched with preset operating condition data, and the preset operating condition type corresponding to the successfully matched preset operating condition data is taken as the operating condition type corresponding to the current operating condition data. Alternatively, this process can be a similarity calculation process: the similarity between the vehicle's current operating condition data and the preset operating condition data is calculated to obtain the similarity score for each preset operating condition data; the preset operating condition type corresponding to the preset operating condition data with a similarity score greater than the preset similarity score is taken as the operating condition type corresponding to the current operating condition data. Here, the preset similarity score is a threshold parameter used to determine whether to accept a certain operating condition type hypothesis, and its specific value can be adjusted according to the application scenario. If the similarity of multiple preset working condition data exceeds the threshold, the working condition type with the highest similarity is selected as the final judgment result; if no working condition has a similarity that reaches the threshold, it is judged as an "unknown working condition", triggering the default processing flow or entering the anomaly detection mode.
[0029] Similarity calculation measures the degree of feature similarity between current operating condition data and preset operating condition data. It can be implemented using various mathematical methods, such as Euclidean distance, Pearson correlation coefficient, Dynamic Time Warping (DTW), cosine similarity, or Mahalanobis distance. When using Euclidean distance, feature vectors of the same dimension (such as mean, variance, dominant frequency, spectral energy concentration, etc.) are first extracted from both the current and preset operating condition data. Then, the normalized distance between them is calculated and converted into a similarity value (e.g., similarity = 1 / (1+d); where similarity represents the similarity and d represents the normalized distance). When using the DTW algorithm, it is suitable for processing time-series signals that are not of equal length or have time offsets, and can more accurately reflect the consistency of dynamic change trends.
[0030] Time-domain and frequency-domain features refer to the extraction of representative statistical quantities and spectral information from raw time-series data (i.e., current operating condition data). Time-domain features include, but are not limited to, mean, variance, skewness, kurtosis, zero-crossing rate, and crest factor, used to characterize the central tendency, dispersion, and transient behavior of a signal. Frequency-domain features include dominant frequency, spectral energy distribution, power spectral density (PSD), and band energy ratio, reflecting the energy concentration of the signal in different frequency bands, which helps identify periodic vibrations or specific mechanical resonance phenomena. To ensure the effectiveness of time-domain and frequency-domain features, the raw data (i.e., current operating condition data) can be preprocessed before extraction, such as detrending, filtering and noise reduction, and normalization, to eliminate the influence of dimensional differences and abnormal disturbances.
[0031] In some embodiments, high-dimensional feature vectors are constructed based on time-domain and frequency-domain features, serving as crucial parameters for operating condition identification. This means that operating condition type identification is not limited to the current operating condition data itself but also requires combining time-domain and frequency-domain features extracted from the current operating condition data to ensure more accurate identification of the operating condition type. For example, the current operating condition data is first cleaned and standardized to ensure the stability of subsequent feature extraction; then, complementary features are extracted in the time and frequency domains respectively to enhance the representation of complex operating conditions; finally, these features are fused and input into a pre-constructed operating condition classification model, which uses machine learning algorithms to determine the category of the current data stream, achieving accurate determination of the operating condition type. This multi-dimensional feature extraction method more comprehensively reflects the changing patterns of vehicle operating states compared to single-domain analysis.
[0032] S120: Determine the current adaptation algorithm based on the working condition type, and determine the weight of each current adaptation algorithm based on time domain characteristics, frequency domain characteristics, data integrity and data noise, so as to obtain the current adaptation algorithm combination.
[0033] Data integrity can be measured by calculating the missing data rate, which is the ratio of valid sampling points to the theoretically required number of sampling points. Data noise can be estimated using the signal-to-noise ratio (SNR), and the residual energy after high-pass filtering can be used as a proxy for noise power. Combining the aforementioned time-domain and frequency-domain characteristics, a multi-dimensional quality evaluation system is constructed to score the applicability of each candidate algorithm under the current operating conditions.
[0034] The current matching algorithm combination is a set of filtering or correction algorithms dynamically determined based on the vehicle's current operating condition, time-domain characteristics, frequency-domain characteristics, data integrity, and noise levels. Its composition adaptively adjusts with the operating environment. This combination includes two or more current matching algorithms, each with advantages for specific types of noise or signal distortion. For example, moving average filtering is suitable for suppressing random noise, Savitzky-Golay filtering excels at preserving signal peak shapes, Kalman filtering is suitable for dynamic process modeling in state estimation, FFT filtering can be used to remove periodic interference, and median filtering has good suppression capabilities for impulse noise.
[0035] The current adaptation algorithm and its weights together constitute the current adaptation algorithm combination. This combination is not fixed, but dynamically adjusted as the working conditions change and the data quality fluctuates, reflecting the system's adaptive capability.
[0036] S130: The original calibration data is corrected using the current combination of adaptation algorithms to obtain the corrected calibration data.
[0037] Raw calibration data refers to unprocessed or only pre-processed calibration-related data collected from vehicle sensors, controllers, or bus systems, such as engine speed, vehicle speed, throttle opening, and temperature parameters. This type of data contains quality issues such as noise, drift, and missing values, and requires subsequent correction to improve its accuracy and stability.
[0038] The original calibration data is input in parallel into each filtering algorithm in the current adaptive algorithm combination. Each algorithm independently performs filtering operations, generating its own intermediate correction results. Then, according to the weights determined in S120 above, all intermediate results are weighted and fused to output the final corrected calibration data. This process supports a multi-threaded parallel execution architecture, with each filtering algorithm running in an independent thread, improving processing efficiency. The fusion method can use linear weighted averaging or introduce non-linear fusion rules (such as a confidence-based voting mechanism). The fused data not only retains the advantages of each algorithm but also effectively suppresses the biases generated by individual algorithms under specific conditions. In addition, the operating condition type, algorithm combination, weight distribution, and original / corrected data comparison information used for each correction can be recorded to form a metadata log for subsequent quality traceability, model iteration, and closed-loop optimization.
[0039] For example, the original calibration data is input into each current adaptation algorithm in the current adaptation algorithm combination, so that each current adaptation algorithm outputs its corresponding correction data; the correction data are weighted and fused, and the fused correction data is used as the corrected calibration data.
[0040] Each current adaptation algorithm in the current algorithm combination independently receives the same raw calibration data and performs corresponding mathematical operations or signal processing logic according to preset parameters, outputting a set of intermediate results, namely "their respective corrected data". This corrected data reflects optimized versions of the same input signal under different processing strategies. For example, in a transient acceleration condition, Kalman filtering more accurately tracks signal trends, while Savitzky-Golay filtering better preserves the detailed characteristics of torque changes. Multiple current adaptation algorithms are executed synchronously through a multi-threaded parallel architecture, with each algorithm running in an independent thread to avoid mutual blocking and improve overall processing efficiency. The thread pool dynamically allocates resources based on the number of algorithms, ensuring millisecond-level response even in scenarios with high real-time requirements. Furthermore, during algorithm execution, runtime, memory usage, and other performance metrics are recorded for subsequent quality assessment and weight updates.
[0041] Weighted fusion refers to the linear combination of the corrected data outputs of each currently adapted algorithm based on predetermined weight coefficients to form the final output result. The weight coefficients can be derived from a corresponding quality scoring mechanism, reflecting the relative reliability of each algorithm under the current operating conditions and data quality conditions. A higher weight indicates better filtering performance of the algorithm at that moment, and a larger proportion of its output in the fusion result. The fusion method is not limited to linear weighted averaging; nonlinear fusion strategies can also be used, such as adaptive fusion models based on neural networks, fuzzy logic fusion devices, or maximum likelihood estimation methods, to cope with complex non-stationary signal environments. The weight update mechanism introduces a time decay factor, giving higher influence to algorithms with better recent performance and enhancing the system's dynamic adaptability.
[0042] This example demonstrates how to effectively overcome the technical bottleneck of a single filtering algorithm in balancing noise reduction and signal detail preservation when dealing with complex and ever-changing vehicle operating conditions. By inputting multiple adaptation algorithms into the original calibration data in parallel and performing fusion processing based on real-time evaluation weights, it improves the overall quality and reliability of the corrected calibration data. This approach is particularly suitable for data optimization tasks in real-world vehicle environments with high dynamics and strong interference.
[0043] This embodiment extracts time-domain and frequency-domain features from the vehicle's current operating condition data and dynamically evaluates the applicability of each algorithm based on data integrity and noise level. It realizes the construction of an adaptive algorithm combination based on operating condition type, which solves the problems that static filtering strategies cannot adapt to dynamic operating conditions and that module isolation leads to overall performance degradation. This significantly improves the correction accuracy, robustness and environmental adaptability of calibration data.
[0044] In another exemplary embodiment of this application, it is described in detail how the weights of each current adaptation algorithm are determined based on time-domain characteristics, frequency-domain characteristics, data integrity, and data noise. Please refer to [link to relevant documentation] for details. Figure 2 , Figure 2 Based on Figure 1 The exemplary embodiment shown illustrates a flowchart of another calibration data correction method. This correction method, in... Figure 1 The S120 shown includes S210 to S230, which are described in detail below: S210: Determine the data integrity of the current operating condition data based on the missing rate of the current operating condition data, and determine the data noise level of the current operating condition data based on the signal-to-noise ratio of the current operating condition data.
[0045] This embodiment quantifies the continuity and availability of data by analyzing the proportion of valid sampling points in the operating condition data, thereby determining the data integrity.
[0046] For example, the matching rate of each working condition data in the current working condition data is calculated by comparing it with the corresponding preset working condition data to obtain the matching rate for each working condition data. Working condition data with a matching rate lower than the preset matching rate are considered abnormal data, and the data integrity of the current working condition data is determined based on the number of working condition data and the number of abnormal data in the current working condition data. Treating working condition data with a matching rate lower than the preset matching rate as abnormal data involves the joint evaluation of multiple variables within each sampling time or data window. A weighted matching mechanism can be used, assigning different weights based on the importance of different signals to working condition identification (e.g., vehicle speed has a higher weight than light signals), thereby improving the sensitivity of key parameter anomaly detection. Furthermore, to avoid misjudgment due to instantaneous fluctuations, a sliding window mechanism can be introduced, marking a persistent anomaly only when multiple consecutive sampling points are below the matching threshold. Determining the data integrity of the current working condition data based on the number of working condition data and the number of abnormal data is a quantitative evaluation achieved by statistically analyzing the proportion of normal data. Specifically, the integrity index I=N valid / N total ; where N valid The number of valid data points that meet the matching rate standard, i.e., the difference between the number of data points in the current operating condition data and the number of abnormal data points; N total This represents the total number of data points collected, i.e., the number of data points related to the current operating condition. This indicator is output as a percentage to facilitate normalization processing by the subsequent algorithm quality scoring module. For example, when the integrity rate is below 85%, the system can trigger an early warning and reduce the initial weights of filtering algorithms that rely on high-precision input (such as Kalman filtering).
[0047] This embodiment measures signal purity by comparing the ratio of useful signal power to background noise power to determine the level of data noise. The signal-to-noise ratio (SNR) is defined as: SNR = 10 × log 10(P_signal / P_noise), where P_signal is the average power of the effective signal component and P_noise is the estimated power of the noise component. In practical implementation, high-frequency noise components can be separated by high-pass filtering, or the signal frequency band can be decomposed using wavelet transform to estimate the noise energy. Alternatively, residual analysis can be used, where the difference between the pre-filtered signal and the original signal is calculated, and the energy of the residual sequence is extracted as an approximation of the noise power. A high signal-to-noise ratio (SNR) indicates good signal quality, making it suitable for fine-grained filtering algorithms sensitive to details (such as Savitzky-Golay filtering); while in low SNR environments, robust algorithms with strong anti-interference capabilities (such as moving average or exponentially weighted filtering) should be preferred. As an optional implementation, a dynamic threshold mechanism can be introduced to set differentiated SNR judgment criteria according to different operating conditions, enhancing adaptability.
[0048] S220: Based on time domain characteristics, frequency domain characteristics, data integrity, and data noise, perform quality evaluation on each current adaptation algorithm to obtain a quality score for each current adaptation algorithm.
[0049] This embodiment comprehensively considers the characteristics of the input data (i.e., time domain features, frequency domain features, data integrity, and data noise) and the expected performance of each candidate algorithm under the current conditions, outputting a quantifiable applicability score. The quality assessment process can be implemented based on a rule engine (such as a fuzzy logic system) or a machine learning model (such as a support vector machine, random forest, or lightweight neural network). For example, if a signal is found to have obvious periodicity and low noise levels, Savitzky-Golay filtering may obtain a higher score; if the signal has frequent abrupt changes and contains many glitches, median filtering is better. Each currently adapted algorithm corresponds to an independent quality scoring module, and the score result is a value in the range [0,1], representing the recommendation strength of the algorithm under the current environment.
[0050] In another exemplary embodiment, if the quality score of the target adaptation algorithm is lower than a preset score, the weights corresponding to the target adaptation algorithm are adjusted to obtain the adjusted current adaptation algorithm combination; wherein, the target adaptation algorithm is any algorithm in the current adaptation algorithm.
[0051] In the process of multi-algorithm fusion, although the weights of each adapted algorithm have been initially determined based on operating conditions, time-domain characteristics, frequency-domain characteristics, data integrity, and noise levels, some algorithms may still experience performance degradation due to sudden environmental changes or signal anomalies during actual operation. To prevent such inefficient algorithms from affecting the overall output quality, a weight correction logic based on quality scores is introduced to achieve real-time optimization of the algorithm combination. Specifically, after the initial weight allocation is completed, a quality assessment and verification operation is further performed. For each adapted algorithm currently participating in the fusion, its corresponding quality score is obtained. This score is derived from multiple dimensions, including but not limited to signal integrity score, noise level score, frequency domain feature clarity score, and time-domain stability score. The quality score of each algorithm is compared with a preset score. If the quality score of a target adapted algorithm is found to be lower than the preset threshold, it is determined to be an unreliable algorithm, and its weight needs to be adjusted. The weight adjustment methods may include, but are not limited to, the following: (1) setting the weight of the target adaptation algorithm directly to zero so that it no longer participates in the weighted fusion process; (2) reducing its weight proportionally, for example, by multiplying it by a decay factor α (0<α<1) to retain some contribution capability; (3) introducing a sliding window mechanism, completely eliminating the algorithm if it is below the threshold for N consecutive times, otherwise only temporarily reducing its weight to enhance the robustness of the system; (4) redistributing the weight originally allocated to the algorithm to other high-scoring algorithms to maintain the normalization characteristic of the total weight. This exemplary embodiment effectively eliminates the interference of unreliable algorithms on the final calibration data by adjusting the weight of the target algorithm that is below the preset quality score, improves the stability and fault tolerance of the system in a dynamically changing environment, and ensures the high-quality output of the corrected calibration data.
[0052] S230: Determine the corresponding weights for each of the current adaptation algorithms based on their quality scores.
[0053] The quality scores of each currently adapted algorithm are converted into weighting coefficients used in the fusion process to ensure that the final output is dominated by the optimal combination. A typical implementation is to normalize all quality scores: sum the quality scores of each currently adapted algorithm to obtain the total quality score; and use the ratio of the quality score of each currently adapted algorithm to the total quality score as their respective weights.
[0054] First, the total quality score is calculated, which is the sum of the scores of all algorithms. Then, the quality score of each algorithm is divided by the total quality score, and the resulting ratio is its weight in the weighted fusion. The mathematical expression is: w_i = Q_i / ΣQ_j, where w_i is the weight of the i-th algorithm, Q_i is its quality score, and ΣQ_j is the sum of the scores of all algorithms. This method ensures that the weights are non-negative and the sum is 1, which meets the probability distribution requirements and is suitable for linear weighted fusion strategies. As an optional implementation, a temperature factor τ can be introduced for softmax normalization: w_i = exp(Q_i / τ) / Σexp(Q_j / τ), where exp represents the exponential function. Exp maps real numbers to positive numbers, and adjusting τ controls the smoothness of the weight distribution: a smaller τ allows the highest-scoring algorithm to dominate, while a larger τ tends to result in a more even distribution. In addition, a minimum score threshold can be set before normalization to force the weights of algorithms with scores below the preset value to be reset to zero, preventing inferior algorithms from interfering with the fusion results.
[0055] This application implements a technical approach to dynamically generate filtering algorithm weights based on real-time data quality and features in complex and ever-changing vehicle operating environments. By introducing a data quality quantification mechanism based on missing rate and signal-to-noise ratio, it solves the problem that traditional static weight allocation cannot respond to changes in operating conditions. By combining time-domain and frequency-domain features to independently score each algorithm, the weight decision-making becomes more scientific and refined. By using a normalization method to map the scores to the weights, the mathematical rationality and engineering feasibility of the fusion process are ensured. Therefore, the intelligence and robustness of the calibration data correction process are improved, achieving the technical effect of improving the accuracy and stability of the corrected data.
[0056] In another exemplary embodiment of this application, the application scenarios of the above-mentioned multiple correction methods are illustrated. The application scenarios include vehicles and servers, and the two ends can be connected wirelessly. This application does not limit the connection method between them.
[0057] The server executes any of the above correction methods, as illustrated below: The operating condition type corresponding to the current operating condition data of the vehicle is determined, and the time domain features and frequency domain features of the current operating condition data are extracted. The current adaptation algorithm is determined according to the operating condition type, and the weight of each current adaptation algorithm is determined according to the time domain features, frequency domain features, data integrity and data noise, so as to obtain the current adaptation algorithm combination. The original calibration data is corrected by using the current adaptation algorithm combination to obtain the corrected calibration data.
[0058] Servers can be located inside vehicles, or they can be physical servers independent of vehicles. They can also be server clusters or distributed systems composed of multiple physical servers, where multiple servers can form a blockchain, and the server is a node on the blockchain. Servers can also be cloud servers that provide 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, CDN (Content Delivery Network), and big data and artificial intelligence platforms. This document does not impose any restrictions on this.
[0059] Another aspect of this application provides a calibration data correction device, such as... Figure 3 As shown, Figure 3 This is a schematic diagram illustrating the structure of a calibration data correction device according to an exemplary embodiment of this application. The correction device 300 includes: The operating condition identification module 310 is used to determine the operating condition type corresponding to the current operating condition data of the vehicle, and to extract the time domain features and frequency domain features of the current operating condition data.
[0060] The algorithm strategy determination module 330 is used to determine the current suitable algorithm according to the working condition type, and to determine the weight of each current suitable algorithm according to time domain characteristics, frequency domain characteristics, data integrity and data noise, so as to obtain the current suitable algorithm combination.
[0061] The correction module 350 is used to correct the original calibration data using the current adaptation algorithm combination to obtain the corrected calibration data.
[0062] In another exemplary embodiment, the working condition identification module 310 includes: The similarity calculation unit is used to calculate the similarity between the vehicle's current operating condition data and preset operating condition data, and obtain the similarity corresponding to each preset operating condition data.
[0063] The working condition identification unit is used to identify the preset working condition type corresponding to the preset working condition data with a similarity greater than the preset similarity as the working condition type corresponding to the current working condition data.
[0064] In another exemplary embodiment, the algorithm strategy determination module 330 includes: The determination unit is used to determine the data integrity of the current operating condition data based on the missing rate of the current operating condition data, and to determine the data noise of the current operating condition data based on the signal-to-noise ratio of the current operating condition data.
[0065] The quality assessment unit is used to evaluate the quality of each current adaptation algorithm based on time domain characteristics, frequency domain characteristics, data integrity, and data noise, and obtain a quality score for each current adaptation algorithm.
[0066] The weight determination unit is used to determine the corresponding weights of each current adaptation algorithm based on their quality scores.
[0067] In another exemplary embodiment, the determining unit includes: The matching rate calculation module is used to calculate the matching rate between each working condition data in the current working condition data and the corresponding preset working condition data, so as to obtain the matching rate corresponding to each working condition data.
[0068] The system identifies sections that treat operating condition data with a matching rate lower than a preset matching rate as abnormal data, and determines the data integrity of the current operating condition data based on the number of operating condition data and the number of abnormal data in the current operating condition data.
[0069] In another exemplary embodiment, the correction device 300 further includes: The adjustment module is used to adjust the weights of the target adaptation algorithm if the quality score of the target adaptation algorithm is lower than the preset score, so as to obtain the adjusted current adaptation algorithm combination; wherein, the target adaptation algorithm is any algorithm in the current adaptation algorithm.
[0070] In another exemplary embodiment, the weight determination unit includes: The overall quality score determination section is used to sum the quality scores of each currently adapted algorithm to obtain the overall quality score.
[0071] The weight determination section is used to determine the weight of each currently adapted algorithm by comparing its quality score with the total quality score.
[0072] In another exemplary embodiment, the correction module 350 includes: The correction data unit is used to input the original calibration data into each current adaptation algorithm in the current adaptation algorithm combination, so that each current adaptation algorithm outputs its corresponding correction data.
[0073] The calibration data correction unit is used to perform weighted fusion of various correction data and use the fused correction data as the corrected calibration data.
[0074] The correction device of this application extracts time-domain and frequency-domain features from the current operating condition data of the vehicle, and dynamically evaluates the applicability of each algorithm in combination with data integrity and noise level. It realizes the construction of adaptive algorithm combination based on operating condition type, solves the problems that static filtering strategy cannot adapt to dynamic operating conditions and module isolation leads to overall performance degradation, thereby significantly improving the correction accuracy, robustness and environmental adaptability of calibration data.
[0075] It should be noted that the correction device provided in the above embodiments and the correction method provided in the foregoing embodiments belong to the same concept. The specific way in which each module and unit performs operations has been described in detail in the method embodiments, and will not be repeated here.
[0076] Another aspect of this application provides an electronic device, including: a controller; and a memory for storing one or more programs, which, when executed by the controller, perform the modified method described above.
[0077] Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of a computer system for an electronic device according to an exemplary embodiment of this application, illustrating a schematic diagram of the structure of a computer system suitable for implementing the embodiments of this application.
[0078] It should be noted that, Figure 4 The computer system 400 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.
[0079] like Figure 4 As shown, the computer system 400 includes a Central Processing Unit (CPU) 401, which can perform various appropriate actions and processes, such as executing the methods described in the above embodiments, based on a program stored in Read-Only Memory (ROM) 402 or a program loaded from storage portion 408 into Random Access Memory (RAM) 403. The RAM 403 also stores various programs and data required for system operation. The CPU 401, ROM 402, and RAM 403 are interconnected via a bus 404. An Input / Output (I / O) interface 405 is also connected to the bus 404.
[0080] The following components are connected to I / O interface 405: an input section 406 including a keyboard, mouse, etc.; an output section 407 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to I / O interface 405 as needed. A removable medium 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 410 as needed so that computer programs read from it can be installed into storage section 408 as needed.
[0081] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit (CPU) 401, it performs various functions defined in the system of this application.
[0082] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. The transmitted data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.
[0083] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.
[0084] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.
[0085] Another aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned correction method. This computer-readable storage medium may be included in the electronic device described in the above embodiments, or it may exist independently and not incorporated into the electronic device.
[0086] Another aspect of this application provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the modified methods provided in the various embodiments described above.
[0087] According to one aspect of the embodiments of this application, a computer system is also provided, including a Central Processing Unit (CPU), which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) or a program loaded from storage into random access memory (RAM), such as performing the methods described above. Various programs and data required for system operation are also stored in the RAM. The CPU, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.
[0088] The following components are connected to the I / O interface: input components including keyboards, mice, etc.; output components including cathode ray tubes (CRTs), liquid crystal displays (LCDs), and speakers; storage components including hard drives; and communication components including network interface cards such as LAN (Local Area Network) cards and modems. The communication components perform communication processing via networks such as the Internet. Drives are also connected to the I / O interface as needed. Removable media, such as disks, optical discs, magneto-optical discs, semiconductor memories, etc., are installed on the drive as needed so that computer programs read from them can be installed into the storage components as required.
[0089] The above description is merely a preferred exemplary embodiment of this application and is not intended to limit the implementation of this application. Those skilled in the art can easily make corresponding modifications or alterations based on the main concept and spirit of this application. Therefore, the scope of protection of this application should be determined by the scope of protection claimed in the claims.
Claims
1. A method for correcting calibration data, characterized in that, The correction method includes: Determine the operating condition type corresponding to the current operating condition data of the vehicle, and extract the time domain features and frequency domain features of the current operating condition data; The current adaptation algorithm is determined based on the working condition type, and the weights of each current adaptation algorithm are determined based on the time domain characteristics, the frequency domain characteristics, data integrity, and data noise, so as to obtain the current adaptation algorithm combination. The original calibration data is corrected using the current combination of adaptive algorithms to obtain the corrected calibration data.
2. The correction method according to claim 1, characterized in that, Determine the operating condition type corresponding to the vehicle's current operating condition data, including: The similarity between the vehicle's current operating condition data and the preset operating condition data is calculated to obtain the similarity corresponding to each preset operating condition data. The preset working condition type corresponding to the preset working condition data with a similarity greater than the preset similarity is taken as the working condition type corresponding to the current working condition data.
3. The correction method according to claim 1, characterized in that, The weights of each current adaptation algorithm are determined based on the time-domain features, the frequency-domain features, data integrity, and data noise levels, including: The data integrity of the current operating condition data is determined based on the missing rate of the current operating condition data, and the data noise level of the current operating condition data is determined based on the signal-to-noise ratio of the current operating condition data. Based on the time-domain features, frequency-domain features, data integrity, and data noise, the quality of each current adaptation algorithm is evaluated to obtain a quality score for each current adaptation algorithm. The weights of each algorithm are determined based on their quality scores.
4. The correction method according to claim 3, characterized in that, The data integrity of the current operating condition data is determined based on the missing rate of the current operating condition data, including: The matching rate of each working condition data in the current working condition data is calculated with the corresponding preset working condition data to obtain the matching rate of each working condition data. Operating condition data with a matching rate lower than a preset matching rate are considered abnormal data, and the data integrity of the current operating condition data is determined based on the number of operating condition data and the number of abnormal data in the current operating condition data.
5. The correction method according to claim 3, characterized in that, The correction method further includes: If the quality score of the target adaptation algorithm is lower than the preset score, the weights corresponding to the target adaptation algorithm are adjusted to obtain the adjusted current adaptation algorithm combination; wherein, the target adaptation algorithm is any algorithm in the current adaptation algorithm.
6. The correction method according to claim 3, characterized in that, The weights corresponding to each current adaptation algorithm are determined based on their quality scores, including: The quality scores of each currently adapted algorithm are summed to obtain the total quality score. The ratio between the quality score of each current adaptation algorithm and the total quality score is used as the weight of each algorithm.
7. The correction method according to claim 1, characterized in that, The original calibration data is corrected using the current combination of adaptation algorithms to obtain corrected calibration data, including: The original calibration data is input into each current adaptation algorithm in the current adaptation algorithm combination, so that each current adaptation algorithm outputs its corresponding correction data. The various corrected data are weighted and merged, and the merged corrected data is used as the corrected calibration data.
8. A calibration data correction device, characterized in that, The correction device includes: The operating condition identification module is used to determine the operating condition type corresponding to the current operating condition data of the vehicle, and to extract the time domain features and frequency domain features of the current operating condition data. The algorithm strategy determination module is used to determine the current adaptive algorithm according to the working condition type, and to determine the weight of each current adaptive algorithm according to the time domain characteristics, the frequency domain characteristics, data integrity and data noise, so as to obtain the current adaptive algorithm combination. The correction module is used to correct the original calibration data using the current combination of adaptive algorithms to obtain the corrected calibration data.
9. An electronic device, characterized in that, include: Controller; A memory for storing one or more programs, which, when executed by a controller, cause the controller to implement the correction method as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, It stores computer-readable instructions that, when executed by the computer's processor, cause the computer to perform the correction method as described in any one of claims 1 to 7.