32 results about "Process noise" patented technology

The invention discloses a gas sensor multi-mode adaptive preprocessing method based on a Mama state space model, and relates to the field of gas detection, and the method comprises the following steps: synchronously collecting original signals, environmental parameters and historical time sequence data of a sensor; performing feature extraction on the three types of data to generate a sensor feature vector, an environment feature vector and a historical feature vector; the method comprises the following steps of: obtaining a final fusion feature through processing, mapping the final fusion feature to a 128-dimensional embedding space through a contrast learning encoder, inputting a three-layer residual error connection Mamba time sequence prediction network to output a zero offset, dynamically predicting process noise and observation noise through a neural network, executing adaptive Kalman filtering to output a smooth signal, and outputting a final fusion signal. Adopting a learnable piecewise linear network to carry out nonlinear correction on the filtering signal; the gas concentration value is calculated by integrating the correction signal and the zero offset, and finally a confidence score and a data quality mark are provided; according to the method, high-precision, real-time and robust pretreatment of various gas sensors in a complex environment is realized.

The invention relates to a high-dynamic time division navigation signal tracking method assisted by adaptive Kalman filtering, and belongs to the technical field of satellite navigation. Aiming at the problems of tracking delay, precision reduction and easy lock loss caused by discontinuous loop update in a high dynamic environment in the existing time division navigation signal tracking method, the method comprises the steps of carrier and code stripping, time-slot error identification, adaptive Kalman filter AKF preprocessing and loop filtering and tracking of three-order FLL-assisted four-order PLL. And full-time-slot continuous tracking is realized. The core of the method is that process noise covariance is adaptively adjusted through AKF, optimal estimation and smoothing are carried out on errors, and then processing is carried out through a post-stage loop filter. According to the method, the tracking continuity, robustness and precision in a high dynamic environment are effectively improved.

The invention relates to the technical field of GNSS high-precision positioning calculation, and discloses a GNSS troposphere delay error real-time correction method and device and a medium, and the method comprises the steps: presetting a first process noise model and a second process noise model corresponding to different constraint values; acquiring motion state parameters of a GNSS receiver, atmospheric disturbance characteristic measurement of bypass information and a model mismatch index of a Kalman filter; when a preset dynamic motion condition is met, selecting a first process noise model; when the preset dynamic motion condition is not met, the first process noise model or the second process noise model is selected according to the atmospheric disturbance characteristic measurement and the model mismatch index, multi-source information is used for cooperative judgment, an algorithm closed loop capable of achieving autonomous recognition and avoiding carrier motion pollution and bypass information failure is constructed, and the algorithm closed loop is optimized. Mathematical coupling of elevation and zenith delay is avoided, and high-reliability rapid convergence can be achieved for independent users without network assistance.

The application discloses an object grabbing method based on a denoising diffusion model, and relates to the technical field of robots, and comprises the following steps: S100, scene point cloud data processing; S200, forward process noise adding network training; S300, point cloud data completion; S400, object grabbing simulation; S500, feature aggregation; S600, aggregated feature noise adding; S700, feature heat map grabbing; and S800, final grabbing posture matrix calculation. The application realizes generation of a six-degree-of-freedom diversity grabbing posture for parallel clamping jaws, and improves the success rate, accuracy and diversity of grabbing.

This invention discloses a joint estimation method and system for state and characteristic parameters based on maneuver intensity perception. It primarily addresses the problems of filter divergence and low parameter estimation accuracy caused by sudden changes in target maneuvering when estimating parameters for hypersonic targets in existing technologies. The implementation scheme is as follows: receiving observation data and calculating innovation; introducing a forgetting factor to adaptively update the process noise covariance, and using the innovation norm to construct a gain factor to adjust the observation weights, completing the posterior state update; calculating the sliding window length based on the adaptive process noise covariance, and constructing the arrival cost using the output of the posterior state update; transforming physical constraints into a penalty function, and combining the arrival cost to establish a rolling time-domain estimation objective function and solve for the optimal estimate. This invention can perceive maneuver intensity in real time, dynamically adjust the estimation window and observation weights, significantly improving the accuracy and robustness of the joint estimation, and can be applied to guidance information extraction and trajectory prediction for near-space hypersonic gliders.

The embodiment of the invention discloses a GNSS receiver clock modeling method based on power-law noise, and relates to the technical field of satellite navigation positioning, and the method comprises the steps: obtaining phase difference data between a local clock and a reference clock, and obtaining a diffusion coefficient of clock noise through extraction based on the phase difference data; establishing a clock state space equation containing a clock offset parameter and a frequency offset parameter, and calculating a process noise covariance matrix of the clock state space equation based on the diffusion coefficient; and embedding the two-dimensional clock random state model and the process noise covariance matrix into a process noise embedding filtering estimation algorithm, and calculating to obtain a clock offset parameter, a frequency offset parameter and a three-dimensional coordinate parameter of the receiver. According to the method, estimation noise of receiver frequency deviation can be remarkably suppressed, and the short-term stability of the elevation direction in GNSS dynamic positioning is effectively improved. The method and the device are suitable for the GNSS terminal adopting a low-cost clock so as to realize high-precision time synchronization and positioning.

The invention relates to the technical field of loading control, and discloses an LQG control method and system for an integral configuration link, and the method comprises the steps: obtaining an actual measurement output vector, an expected output vector and a system state observation vector of a loading control system in a current control period, and constructing an integral type performance index functional; obtaining an augmented state space model and an augmented state estimation vector according to an original state space equation, a process noise covariance matrix, a measurement noise covariance matrix and a system state observation vector of the loading control system; based on the augmented state space model and the integral type performance index functional, establishing an extended Riccati differential equation to obtain an optimal feedback gain matrix; and performing linear feedback operation on the augmented state estimation vector based on the optimal feedback gain matrix, generating a control instruction of a loading control system, and realizing zero-static-error optimal tracking, independent and accurate setting of multi-channel integral intensity, adaptive optimization of controller parameters and robust stable operation under uncertainty.

This invention relates to the field of rail transit technology and discloses a rail corrugation detection method based on anomaly detection and dynamic adaptation. The method includes: acquiring the axle box acceleration vibration signal of the rail; performing state prediction using a Kalman filter model to obtain a priori state estimate and calculating the measurement residual; dynamically adjusting the process noise covariance matrix of the Kalman filter model based on the measurement residual; determining a dynamic threshold based on the measurement residual; comparing the measurement residual with the dynamic threshold to determine whether the measurement residual is an outlier; if so, robustly correcting the priori state estimate to obtain a posterior state estimate. This invention effectively solves the problem that traditional fixed-parameter models cannot accommodate different track conditions, maintaining optimal filtering performance on various track grades and avoiding common problems in traditional methods such as filter divergence or tracking hysteresis.

The application discloses a kind of self-adapting noise reduction method and noise reduction device, including following process: the noise source of specific environment is collected, and is converted into noise data, the record keeping of noise data is carried out;Analysis processing noise data, determine the A rate coefficient of noise source under environment;Record keeping and send A rate coefficient to adaptive controller, carry out adaptive filtering noise reduction.The application realizes the analysis to specific environmental noise, can evaluate the use effect of protective earmuff, provides direction for protective earmuff improvement, and can grasp the relative relationship of personnel hearing and noise, strong universality, better practicality, easy to popularize and apply, with greater practical value.

The application provides a satellite autonomous continuous maneuvering orbit determination method with adaptive process noise, and aims at the problem of orbit determination accuracy reduction of a satellite in a maneuvering process.The method is based on innovation, and adaptively adjusts process noise in a non-maneuvering stage and a maneuvering stage, so that the estimation accuracy reduction problem caused by fixed noise parameters is solved, and the filter can more accurately represent the dynamic state of the system.In a Gaussian noise, compared with a traditional unscented particle filter, the APNUPF exhibits better orbit determination accuracy during the maneuvering, and even if the state mutates, the APNUPF can still maintain high-precision orbit determination.In addition, in order to meet the actual engineering requirements, the robustness and reliability of the method for satellite autonomous continuous maneuvering orbit determination are comprehensively verified under non-Gaussian noise conditions.

The application discloses a power quality analysis method and system for a data acquisition terminal, and relates to the technical field of electric power engineering.The method comprises the following steps: acquiring data of various electrical parameters and preprocessing the data to obtain standardized multidimensional sequences; quantifying the correlation between each dimension based on the multidimensional sequences, and calculating the overall correlation level of each sampling time in the sequence; generating a forgetting factor for each sampling time in the sequence based on the overall correlation level; constructing adaptive measurement noise matrices and process noise matrices based on the forgetting factor; performing Kalman filtering based on the two adaptive noise matrices to obtain adaptively filtered multidimensional sequences; and performing inverse standardization transformation and FFT analysis on the adaptively filtered multidimensional sequences to analyze power quality parameters.The application solves the defect that traditional Kalman filtering algorithms ignore the correlation between different electrical parameters, can improve the denoising effect, and enhances the accuracy and reliability of power quality analysis.

The application discloses an optimal output feedback control method suitable for a direct current microgrid, belongs to the technical field of direct current microgrid control, and comprises the following steps: based on a projection theorem, an optimal state filter is designed to eliminate the time correlation of process noise and measurement noise; by means of dynamic programming and a mathematical induction method, the performance index of the system is decomposed into two parts; by means of a state augmentation method, a non-public information set is used to transform the interconnected system; by means of orthogonal decomposition technology, the control input, the state variable, the process noise and the measurement noise of the transformed system are all decomposed into two independent parts; and based on the time correlation noise filter and the system variable decomposition, an optimal output feedback control method suitable for the characteristics of the direct current microgrid is designed. The application effectively improves the control precision and stability of the direct current microgrid in a complex noise environment, solves the information asymmetry problem caused by communication delay, and provides a complete solution convenient for engineering implementation.

The invention provides a method and device for estimating a noise covariance matrix in a Kalman filter. Time series measurements of the state relate to a state-space model, process noise, and measurement noise. And estimating a noise covariance matrix through the time sequence measurement value, wherein the noise covariance matrix is used for estimating the state of the future time. According to the scheme, it is assumed that a Kalman filter is realized by using an initial noise covariance matrix, a target function obtained based on an innovation sequence is minimized by changing a Kalman filtering gain, and the noise covariance matrix is updated according to the updated Kalman filtering gain. The process is an updating iteration process. In the process of updating the noise covariance matrix, the updated Kalman filtering gain and the covariance matrix corresponding to the key parameters obtained in the Kalman filtering process are used for iterative updating, so that a converged noise covariance matrix is obtained when the corresponding value of the target function is minimum. In the process, actual time sequence measurement values are used instead of prior information, and the estimation method is more stable and reliable.

The invention provides a robust Kalman filtering method based on STIW distribution, and relates to the technical field of state estimation and filtering. The method comprises the following steps: S1, constructing a state space model, and initializing parameters; step S2, on the basis of the initialized parameters, modeling the process noise as Gaussian-inverse Wishart distribution; s3, modeling the measurement noise as STIW distribution based on the initialized parameters, and generating a three-layer layering model; s4, based on the state space model, executing time updating, predicting state and covariance, and updating noise parameters; step S5, based on the output of the step S4, using a VB method to update posterior distribution through fixed-point iteration; and step S6, repeating the step S5 until a convergence condition or a maximum number of iterations is reached, and outputting a final estimation result. According to the robust Kalman filtering method based on STIW distribution, the precision and robustness of an estimation result in a complex noise environment are improved.

The invention discloses a noise-adaptive gyroscope null drift and attitude estimation method, and belongs to the technical field of satellite attitude measurement. The method is used for obtaining the gyroscope null drift and the attitude quaternion of the micro-nano satellite, the gyroscope measurement angular velocity is obtained through the method, the noise of the gyroscope measurement angular velocity is stripped through a total variation denoising method, and a smooth component is obtained; on the basis of stripping the noise, a preset process noise covariance matrix is adopted, so that the dependence on real noise is avoided; according to the method, estimation of attitude quaternion and gyroscope null drift is realized through the MEKF method. According to the method, uncertain process noise is bypassed, and the precision is improved.

The application discloses a kind of based on timing supervision and whole link occlusion perception's self-adapting target tracking method and system, the method constructs double-flow parallel prediction architecture, utilizes the Mamba network with linear calculation complexity to carry out nonlinear motion supervision to Kalman filter, when monitoring that prediction deviation exceeds threshold value, the process noise of filter or reset speed state and covariance matrix is adaptively adjusted, to eliminate cumulative inertia error;In detection stage, the intelligent non-maximum suppression strategy based on minimum area intersection-over-union (IoS) and depth difference is used, and the occluded effective target is retained;In association and update stage, appearance feature fusing is triggered by detecting trajectory collision risk, and feature update cooling lock is started for occluded or re-found target, and feature library pollution is blocked.The application can effectively improve the tracking accuracy of nonlinear maneuvering target, and significantly reduce the identity switching rate in dense crowded scene.

The invention provides a method and a device for estimating a noise covariance matrix of a Kalman filter. In a state space model used by a Kalman filter, measurement noise and process noise can influence a measurement value and an estimation value of a state, and an error existing in a true value of the state is mainly caused by a measurement noise covariance and a process noise covariance. According to the scheme, a method for accurately estimating the measurement noise covariance matrix and the process noise covariance matrix of the Kalman filter is provided through the time sequence measurement value of the reference object measured by the target sensor so as to represent the noise. Therefore, the target value of the reference object in the next time state can be accurately estimated by using the measurement noise covariance matrix and the process noise covariance matrix. Namely, the optimal state value of the reference object in the future time is estimated.

The invention relates to the technical field of precision displacement measurement, and discloses a dynamic accumulative error compensation method based on error state Kalman filtering and two-parameter adaptive adjustment. Aiming at the accumulative error problem of a grating displacement sensor caused by mechanical vibration, temperature drift and the like and the limitation that classical Kalman filtering cannot effectively process, the method is realized by the following steps of: establishing a mathematical model of the accumulative error and defining a system state vector; performing prediction estimation on the error state and the covariance; calculating an observation residual error and a Kalman gain; based on residual statistical characteristics, performing adaptive adjustment on covariance of the process noise and the measurement noise; and finally, the error state is updated by using Kalman gain, and the displacement measurement value is compensated in real time. The core of the method is that error state separation estimation is combined with a dual-noise parameter adaptive mechanism, time-varying accumulative errors are effectively tracked and compensated, and the long-term precision and stability of displacement measurement are remarkably improved.

The invention discloses a wide-area time transfer method based on adaptive random noise constraint, and the method comprises the steps: obtaining the historical observation data of a receiver, and carrying out the calculation based on a precise single-point positioning model, and obtaining a receiver clock error sequence; performing variance analysis on the receiver clock error sequence, identifying dominant noise types under different smoothing times, and determining a time interval of segmented modeling; selecting a corresponding variance estimator or constructing a mixed variance model according to the dominant noise type, and calculating an adaptive random noise constraint value; and introducing the adaptive random noise constraint value into a process noise matrix of the precise point positioning model, and carrying out dynamic constraint estimation on a receiver clock error to realize time transfer. The method provided by the embodiment of the invention can overcome systematic deviation introduced by a traditional fixed constraint strategy, adapts to physical property differences of different atomic clocks, and remarkably improves receiver clock error estimation precision and frequency stability of wide-area time transfer.

The present application relates to the technical field of computer processing. Provided are an audio processing method and apparatus, a model training method and apparatus, and an electronic device. The audio processing method comprises: acquiring target audio recorded by means of a rotating device, a device rotation speed and environmental background audio; separately performing spectrum extraction processing on the target audio and the environmental background audio, so as to obtain a target spectrogram and a background audio spectrogram; inputting the target spectrogram, the device rotation speed and the background audio spectrogram into a noise filtering model, so as to obtain a noise-reduced spectrogram, wherein the noise filtering model is used for performing noise-filtering processing on the target spectrogram on the basis of the device rotation speed and the background audio spectrogram and outputting the processed noise-reduced spectrogram; and performing audio conversion processing on the noise-reduced spectrogram, so as to obtain noise-reduced audio corresponding to the target audio. By means of the present application, more accurate noise reduction can be implemented, thereby improving the effect of noise cancellation for audio recorded by means of a rotating device.

The application discloses a kind of adaptive turntable space attitude estimation method and device based on multiple code disc, belong to the field of attitude estimation.Method includes: for each time's turntable space attitude estimation, all execute: the turntable azimuth angle measurement value of current time absolute code disc feedback and the turntable pitch angular velocity measurement value of incremental code disc feedback are acquired, to construct the space attitude measurement matrix of current time;By Kalman filtering algorithm, the measurement value of absolute code disc and incremental code disc feedback is fused, and the turntable space attitude estimation value of current time is obtained;Residual-based adaptive method is constructed, and process noise covariance matrix and observation noise covariance matrix are dynamically adjusted to cope with time-varying noise and model error.The scheme combines the advantages of two kinds of sensors of absolute code disc and incremental code disc, and adjusts covariance weight based on residual in real time to balance the uncertainty of system noise, with high accuracy and high safety.

The invention relates to the technical field of noise control of a power equipment control room, in particular to an active noise reduction method for an indoor substation, which is characterized in that under a steady-state working condition, a pre-trained deep learning model is adopted, and relevant parameters of an inversion signal are generated by an original noise signal, so that the problem that a traditional fixed filter cannot be matched with a phase is effectively solved; under the unsteady-state working condition, the type of signal disturbance is further judged, and frequency tracking and amplitude estimation are performed in the original noise signal decomposition process, so that the method can quickly respond and effectively process noise under the unsteady-state conditions such as frequency micro-deviation, amplitude sudden change and new harmonic waves, and the problem of noise reduction failure caused by unsteady-state disturbance is relieved. Through the geometric inversion phase compensation technology, the door slot reflection sound path is fully considered, the problem of phase mismatch caused by neglecting the factor in a traditional scheme is effectively solved, the phase matching error is controlled within a small range, and the stability and reliability of the noise reduction effect are improved.

The present disclosure relates to process noise adjustment in target trackers. A computer-implemented method for adjusting a process noise of a target tracker in a camera (100) configured to monitor a scene (200) comprising a moving target (102a-c), the target tracker (107) using a motion model (134) associated with a process noise (132), the method comprising: using the target tracker, determining (S100) a target trajectory of the moving target travelling along a path in the scene over a time period using a process noise level; evaluating (S105) at least one spatial parameter of the determined target trajectory relative to a predetermined criterion (140); repeating the steps of determining (S100) the target trajectory and evaluating (S105) the at least one spatial parameter of the target trajectory during a subsequent time period using an increased process noise for each iteration until the predetermined criterion is met or a maximum process noise limit is reached; and storing (S110) data (144) indicative of a final process noise.

A computer-implemented method to adjust a process noise of an object tracker in a camera is configured to monitor a scene including moving objects, the object tracker uses a motion model associated with process noise, the method comprising: determining, using the object tracker, object tracks of moving objects travelling along a path in the scene for a time period using a level of process noise, evaluating at least one spatial parameter of the determined object tracks in relation to a predetermined criterion, repeating the steps of determining object tracks and evaluating at least one spatial parameter for object tracks during subsequent time periods using increased process noise for each iteration until the predetermined criterion is fulfilled or a maximum process noise limit is reached, and storing data indicating the final process noise.

The invention discloses a transformer substation noise intelligent separation method and device, equipment and a medium, and relates to the field of audio processing, and the method comprises the steps: collecting equipment noise generated when a plurality of target equipment works, carrying out the preprocessing of the equipment noise, and constructing a target mixed noise data set based on the obtained preprocessed noise data; mixing the original audio signals, decomposing the obtained mixed audio signals to obtain decomposed audio signals, and constructing a signal distortion ratio and a scale-invariant signal distortion ratio through the original audio signals, the decomposed audio signals and the mixed audio signals; and optimizing a preset multi-channel speech enhancement model through the distortion ratio and the target mixed noise data set, and performing noise separation on the to-be-processed noise data based on the obtained target multi-channel speech enhancement model to obtain a separated noise signal and a separated audio signal. Therefore, the noise signal can be processed through the multi-channel speech enhancement model, and the noise separation effect is further improved.

The invention discloses an unknown related multi-sensor state and noise covariance joint estimation method. The method comprises the following steps: S1, initializing a multi-sensor system model and parameters of a local Kalman filter of each sensor; s2, on the basis of the initialized system model and filter parameters, each sensor carries out target state prediction and updating, and an innovation sequence is generated and accumulated; s3, based on the accumulated information sequence and the sliding time window with the preset length, each sensor carries out estimation to obtain a local process noise covariance and a local measurement noise covariance; s4, based on the estimated local noise covariance, distributing a fusion weight for each sensor; and S5, performing weighted fusion on the local process noise covariance and the local measurement noise covariance based on the fusion weight distributed in the step S4 to obtain global process noise covariance estimation and global measurement noise covariance estimation.

The invention discloses a big data identification method for periodic weighting of a coal face, and belongs to the technical field of coal mine pressure analysis. The method comprises the steps that S1, support resistance data are cleaned based on pressure gradient characteristics, and process noise is eliminated; s2, constructing a nonlinear propulsion model, and realizing accurate mapping from time domain data to a space propulsion degree; s3, the working face is divided into regions and boxes, the group energy of each region is calculated, and a multi-dimensional energy sequence is constructed; s4, extracting a static load baseline by adopting an anti-noise trend separation method based on quantiles, separating load disturbance characteristics, and enhancing by using a morphological method; and S5, constructing an adaptive threshold based on the robust statistics, and performing consistency verification in combination with the geological prior periodic step pitch to realize automatic and accurate identification of the periodic weighting. According to the method, the problems of incomplete data cleaning, distortion of space-time mapping, weak anti-noise capability, dependence on empirical thresholds and the like of a traditional method are solved, and the accuracy, the robustness and the automation level of weighting identification are remarkably improved.

The application discloses a kind of periodic weighting big data identification methods of coal mining face, belong to coal mine strata pressure analysis technical field.The method includes: S1, based on pressure gradient feature cleaning support resistance data, eliminate process noise;S2, construct nonlinear propulsion model, realize the accurate mapping of time domain data to space propulsion degree;S3, working face is divided and is boxed, calculates each regional group energy, constructs multidimensional energy sequence;S4, using the anti-noise trend separation method based on quantile extracts static load baseline, separates load disturbance characteristics, and is enhanced using morphological method;S5, based on robust statistics, construct adaptive threshold, combine geological priori period step distance and carry out consistency check, realize the automatic, accurate identification of periodic weighting.The application solves the problems of traditional method, such as incomplete data cleaning, space-time mapping distortion, weak noise resistance and identification relying on empirical threshold, significantly improves the accuracy, robustness and automation level of weighting identification.

This application relates to a pipeline inspection method, apparatus, computer equipment, storage medium, and computer program product. The method includes: extracting the power spectrum of an original time-series signal; determining the number of modes required for signal decomposition based on peak distribution; extracting multiple spectral shape features from the power spectrum and determining a penalty factor for signal decomposition based on a comprehensive evaluation of these features; performing variational mode decomposition on the original time-series signal using the number of modes and the penalty factor to obtain decomposed sub-signals; extracting time-frequency domain features of the decomposed sub-signals and dividing them into an effective signal subset and a noise signal subset based on these features; performing noise suppression processing on the decomposed sub-signals in the noise signal subset and merging the processed noise signal subset with the effective signal subset to reconstruct a denoised time-series signal. This method can more accurately extract pipeline structural state features, thereby improving the accuracy and reliability of pipeline defect identification.