23 results about "Robust filtering" patented technology

The invention provides a physical driving measurement method for monocular three-dimensional dynamic displacement of a rotary machine, and belongs to the technical field of crossing of computer vision and industrial state monitoring. According to the method, a high-speed dynamic visual acquisition system is constructed, and a time sequence video stream is obtained by using a cooperative marker; constructing a'Gaussian + motion blur 'composite gradient model taking the motion blur width as an endogenous variable, and jointly resolving sub-pixel edge coordinates and ambiguity by adopting a nonlinear optimization algorithm; decoupling the pixel displacement of the radial X axis, the radial Y axis and the axial Z axis by using the width change and the edge displacement of the marker in combination with the geometric principles of sequential robust filtering and monocular imaging; and finally, outputting a three-way physical vibration waveform by using the calibration conversion factor. According to the method, the fuzzy influence is adaptively eliminated from a physical imaging mechanism, micron-level precision three-direction vibration synchronous measurement is realized only by a single camera, and the hardware cost and the deployment difficulty are reduced.

The invention provides a gross error detection method, gross error detection equipment, a readable storage medium and a product, and relates to the technical field of positioning. The method comprises the following steps: inputting a satellite observation value into an RTK filter, and obtaining a residual error output by the RTK filter at the current moment and a prediction variance covariance matrix of the residual error; determining an error in a unit weight according to the residual error and the predicted variance covariance matrix, and calculating a dynamic threshold value according to the error in the unit weight; according to a preset experience threshold value and a dynamic threshold value, performing first gross error elimination on the residual error to obtain a first residual error sample; according to the condition number of the predicted variance covariance matrix, performing second gross error elimination on the first residual error sample to obtain a second residual error sample; and performing de-correlation calculation on the second residual error sample construction, and performing adaptive robust filtering on a de-correlation residual error subset. According to the method, gross error elimination is carried out from a data layer and a model layer, and more accurate and efficient RTK gross error detection capability is achieved.

The application provides a non-Gaussian noise suppression method and device with adaptive kernel width, and belongs to the field of inertial base combined navigation algorithm and state estimation. The method solves the problems that the robust filtering method based on fixed kernel width is difficult to adapt to time-varying noise characteristics, and the scheme depending on an optimization algorithm has the problem of insufficient real-time performance. The method comprises the following steps: initializing filter parameters according to a SINS / DVL combined navigation system; performing time updating to obtain a predicted state vector and a predicted state covariance matrix at the current moment; updating a measurement noise covariance matrix through a variational Bayesian method, wherein the measurement noise covariance matrix is modeled as an inverse Wishart distribution; adaptively updating a kernel width parameter according to a filter innovation at the current moment and the measurement noise covariance matrix; and updating a state quantity estimation value and a state covariance matrix at the current moment through a fixed-point iteration algorithm by using the updated kernel width. The method is used in the field of underwater resource exploration.

The application provides a physical driving measurement method for monocular three-dimensional dynamic displacement of a rotating machine, and belongs to the technical field of computer vision and industrial state monitoring; a high-speed dynamic vision acquisition system is constructed, and a time sequence video stream is acquired by using a cooperative marker; a "Gaussian + motion blur" composite gradient model taking a motion blur width as an endogenous variable is constructed, and a nonlinear optimization algorithm is used to jointly solve sub-pixel edge coordinates and a blur degree; in combination with time sequence robust filtering and monocular imaging geometric principles, pixel displacement of radial X and Y axes and axial Z axis is decoupled by using marker width variation and edge displacement; finally, three-way physical vibration waveforms are output by using a calibration conversion factor. The application starts from a physical imaging mechanism, adaptively eliminates blur influence, and realizes three-way vibration synchronous measurement with micron-level precision by using only a single camera, so that hardware cost and deployment difficulty are reduced.

The application discloses a kind of unmanned aerial vehicle flight control method and system, for solving the protection level failure caused by navigation error non-gaussian distribution and the problem of frequent false trigger of flight control of unmanned aerial vehicle in complex electromagnetic environment of city low altitude. Method includes obtaining bottom sensor original observation value, using robust filtering algorithm to suppress multipath interference and output residual error;Non-gaussian error boundary is constructed based on set theory geometric envelope algorithm, and the rigidity quantitative index of navigation reliability is deduced by protection level;Dynamic mapping operation risk alarm limit, and introduce time anti-shake tolerance window to filter transient signal burr;When reliability is continuously broken down, trigger state machine to execute stepwise autonomous degradation. The application breaks the limitation of traditional gaussian hypothesis, and considers the absolute safety and task continuity of flight.

The invention particularly relates to a sea clutter pulse interference suppression method for cascaded Hammerstein robust filtering, and the method comprises the steps: receiving an echo signal, and carrying out the preprocessing of the echo signal, so as to obtain a direct signal and a sea clutter signal, which are separated from each other; a cascaded Hammerstein filter is used to carry out multi-stage filtering processing on the separation signal, and a non-linear mapping output signal is obtained; wherein the cascaded Hammerstein filter comprises a first nonlinear filtering module, a second linear filtering module and a third nonlinear filtering module which are arranged in a cascaded manner; and carrying out filtering processing on the nonlinear mapping output signal through a robust filter to obtain an output signal after interference suppression. According to the method, nonlinear distortion and non-Gaussian pulse interference can be effectively suppressed in a complex marine environment, and the anti-interference capability, the signal fidelity and the adaptivity of a radar and a communication system are improved.

The application belongs to the technical field of navigation, and proposes an adaptive robust filtering navigation method based on multivariate t distribution and Bayesian shrinkage. First, multi-source heterogeneous sensor observation data is obtained. Second, the multi-source heterogeneous sensor data is modeled based on multivariate t distribution, and the innovation of the multi-source heterogeneous sensor observation data is extracted. Then, a continuous confidence factor is constructed, the fault detection is converted into a Bayesian inference problem, and the confidence diagnosis of the observation innovation is performed based on the confidence factor. Then, the innovation discount factor is constructed based on the confidence factor, and the weighted fusion target function is constructed based on the observation fitting term and the virtual robust term, so as to obtain the equivalent innovation observation, the equivalent observation noise covariance matrix, the equivalent innovation covariance and the robust Kalman gain. Finally, the observation information weight is dynamically adjusted, the equivalent parameters are used to complete the robust filtering iteration, and the optimal navigation solution is obtained. The application can improve the navigation positioning precision and robustness in the multi-source fusion navigation process.

This invention discloses an attention-integrated multimodal federated learning method and system for risk monitoring. On the client side, an attention-integrated hybrid early fusion module dynamically injects multimodal features into globally shared latent variables through attention mechanisms and entropy-based gating mechanisms. On the server side, a plug-and-play dual-track federated aggregation strategy decouples state aggregation from model aggregation. The state track employs robust filtering based on the geometric median, while the model track supports algorithm-independent parameter updates. Between communication rounds, a neighborhood-weighted personalized reference mechanism constructs a gradient-based semantic topology to balance global consistency and local task preferences. This invention simultaneously addresses the technical challenges of modality incompleteness, data statistical heterogeneity, and Byzantine adversarial robustness in multimodal federated learning, thus providing an efficient and robust collaborative learning solution for disaster risk perception under privacy-preserving conditions.

The application discloses a noise-induced score function-based Kalman filtering method for target tracking, and belongs to the technical field of intelligent perception and state estimation. In view of the problem that sensors such as radars and millimeter wave radars are easily interfered by multipath reflection and other pulse interference in a complex environment, leading to abnormal measurement, the application constructs a class of noise-induced M score function, and rewrites the observation update step of Kalman filtering into a generalized M estimation form; meanwhile, an online adaptive update algorithm of a noise scale parameter is designed according to a measurement residual, and collaborative adaptive estimation of system state and noise parameters is realized. The application does not need prior noise distribution assumption, can automatically soft-restrain abnormal measurement under pulse interference, and significantly improves the robustness and precision of state estimation. Compared with traditional Kalman filtering and existing robust filtering methods, the application has higher estimation precision and stronger anti-interference ability in applications such as radar target tracking, automatic driving front vehicle tracking and unmanned aerial vehicle tracking.

A flight calibration method and system for the structural parameters of a star sensor and an inertial unit (IMU) includes: setting the parameters of the star sensor and IMU, and environmental parameters, and setting the initial state of the structural parameter deviation estimator; after the IMU and star sensor output data, performing time propagation on the estimated values ​​and estimated covariance of constant zero bias, zero bias drift, and structural parameter deviation; calculating the attitude increment of the IMU and star sensor using angular velocity and star observation information, and constructing the measurement equation for the structural parameter deviation; solving the corresponding regularized least squares problem using a regularized robust filtering framework to update the estimated values ​​and estimated covariance of constant zero bias, zero bias drift, and structural parameter deviation; and compensating for the structural parameters after obtaining a stable estimated value of the structural parameter deviation. This invention enables high-precision calibration of the structural parameters of the star sensor and IMU without the need for complex astronomical-ground coordinate transformations using timing and positioning information.

This invention relates to the field of traffic control systems, specifically to a checkpoint travel time estimation method based on mixed-integer optimization and spatial KNN, comprising the following steps: cleaning mapped checkpoint data, calculating initial spatiotemporal parameters, and removing anomalies based on absolute median; decomposing travel time and removing anomaly dwellers; dividing the road network according to traffic flow thresholds, estimating high-traffic road segments using mixed-integer optimization verification and Bayesian dynamic fusion, and performing parameter extrapolation based on spatial KNN topological constraints and congestion characteristics inverse solution for low-traffic road segments; fusing the results from dual-source road segments to construct a global parameter vector, and iteratively optimizing through an iterative convergence mechanism to output a high-precision global travel time field. This method utilizes absolute median to construct a robust filtering mechanism to remove extreme anomalies; it performs parameter extrapolation for low-traffic road segments, achieving high-precision estimation while ensuring physical interpretability, accurately reconstructing the dynamic operating state of the road network.

The invention relates to the field of traffic control systems, in particular to a bayonet travel time estimation method based on mixed integer optimization and space KNN, which comprises the following steps: cleaning mapping bayonet data, measuring and calculating initial space-time parameters, and eliminating anomalies based on absolute median difference; decomposing the travel time and stripping abnormal residence; road networks are divided according to flow threshold values, high-flow road sections are estimated through mixed integer optimization check and Bayesian dynamic fusion, and low-flow road sections are subjected to parameter deduction based on space KNN topological constraint and congestion feature reverse inverse solution; and fusing double-source road section results to construct a global parameter vector, and outputting a high-precision global travel time field through iterative convergence mechanism loop optimization. According to the method, an absolute median difference is utilized to construct a robust filtering mechanism to eliminate extreme anomalies; parameter deduction is implemented for a low-flow road section, high-precision estimation is realized on the basis of ensuring physical interpretability, and the dynamic operation state of a road network is precisely restored.

The invention discloses a thick-tail filtering method for reducing inaccuracy noise covariance sensitivity, and belongs to the technical field of control engineering and signal processing, and the method comprises the steps: S1, initializing a state, a hyper-parameter, covariance priori and GEG thick-tail noise parameters; s2, executing a time updating operation, and obtaining a prior state and related distribution parameters; s3, carrying out iterative measurement updating, and completing posterior distribution estimation of each variable through fixed-point variational Bayesian VB iteration; and S4, outputting a filtering result, and keeping update parameters for subsequent use. According to the thick-tail filtering method for reducing the sensitivity of the covariance of the inaccurate noise, the problems of the thick-tail noise and the inaccurate covariance can be effectively solved, the sensitivity of the algorithm to the noise covariance error is remarkably reduced, high estimation precision and high robustness are achieved, the calculation complexity is moderate, and the method is suitable for large-scale popularization and application. And the performance in scenes such as target tracking is superior to that of a traditional robust filtering algorithm.

The application provides a remote sensing image building change detection method based on matching optimization, comprising the following steps: step 1: high-precision building recognition is performed on two images before and after, a graph model is constructed according to image change intensity information and edge intensity characteristics, and a building change candidate area of two time phases before and after is obtained; step 2: a method based on motion statistical feature matching is used to estimate the roof relationship, and a fast robust filtering strategy is used for judgment and correction, so that the stability of the matching process is improved; step 3: a fully connected conditional random field model is established, building change information of the candidate area of two time phases before and after and roof matching results are integrated together for cooperative optimization, and the influence of position difference on building change detection is eliminated. The application can accurately capture potential building change areas, and overcomes the problem that building change detection of high-resolution remote sensing images is difficult to find and complete under complex and diverse scenes.

The present application relates to the field of satellite positioning technology, and particularly relates to a robust filtering method based on satellite quality evaluation, comprising the following steps: step 1: a receiver collects navigation messages in real time; step 2: single point positioning is used to obtain the receiver position and signal quality evaluation parameters; step 3: satellite signal quality evaluation is performed to determine whether the observation matrix is full rank; step 4: a robustness factor is calculated, and a Kalman gain is calculated; step 5: multi-epoch joint observation is performed to establish full-rank complete geometric constraints to solve new predicted values; and step 6: the current epoch parameters are retained, and the next epoch is solved. The scheme of the present application has the advantages that it can be used to reduce the influence of weak satellite signal strength, multipath effect and the like on positioning accuracy, and improve the continuous positioning calculation capability.

In view of the problem of robust HRRP sequence target recognition under the condition of intermittent sampling and retransmission interference, the application provides a robust target recognition method, system and device for intermittent sampling and retransmission interference, a target recognition network model is constructed, and the model is trained by using high-resolution one-dimensional range profile data of different targets under the condition of intermittent sampling and retransmission interference, sparse features with structured characteristics are extracted from the disturbed signal first, then the target and interference components are decoupled and fused in the feature space through a mask estimator, a soft threshold mask that suppresses interference and enhances target features is generated, and finally the pure target signal is recovered through a reconstruction module and the recognition is completed by a Bayesian recognition network. The application realizes the robust filtering of intermittent sampling and retransmission interference and the enhancement and extraction of target features, thereby significantly improving the accuracy and reliability of radar high-resolution one-dimensional range profile target recognition in a complex interference environment.

The invention discloses a two-stage compression denoising imaging method for a large-area array Geiger mode APD (avalanche photo diode). The method comprises the following steps: intercepting photon events in a pre-sampling time window at the front part of a data block to carry out global histogram statistics, fitting and estimating time domain characteristics of background noise based on Poisson distribution, screening and compressing full-frame data, and eliminating most noise; carrying out local joint histogram statistics and peak value extraction on the compressed data based on a pixel space neighborhood, and reconstructing a preliminary distance image and a preliminary intensity image; if the effective data in the neighborhood of the data meets the flat area condition, standard median filtering is executed; otherwise, starting a robust filtering algorithm based on absolute median difference MAD for processing. According to the method, the cooperation of front-end data efficient compression and rear-end accurate denoising and edge preserving is realized, and the method has the advantages of high compression ratio, high adaptability and excellent imaging quality, and is particularly suitable for real-time three-dimensional imaging of a large-area-array GM-APD laser radar system.

The application discloses a kind of based on multiple adaptive factors interactive robust state estimator design method, for improving the estimation accuracy of estimator and the curve smoothness of filtering under the condition that observation noise occurs unpredictable surge, to meet the filtering estimation demand of higher accuracy.Standard robust state estimation method requires known observation noise covariance value, therefore cannot be normally used under the condition that observation noise surges;Robust filtering method combined with single adaptive factor can compensate the estimation error caused by noise surge to a certain extent, but its performance cannot achieve universal optimization under different noise surge degrees.The application can effectively improve the estimation effect of state estimator under complex observation noise surge condition, break through the limitation of traditional robust state estimation method, realize fast estimation of actual observation noise under complex observation noise surge condition, effectively improve the estimation accuracy and estimation curve smoothness of state estimator, optimize the estimation effect of estimator.

The invention relates to a time-frequency dual-robust filtering method for demagnetization diagnosis of a permanent magnet motor, which belongs to the field of fault diagnosis of the permanent magnet motor, and comprises the following steps: loading an original signal of the permanent magnet motor, and carrying out time-domain filtering through a self-adaptive Hampel filter; framing windowing processing: converting the signal after time domain filtering into a frequency domain signal; training a Gaussian process regression model by using a normal sample; calculating a reference noise spectrum; performing frequency domain depth noise reduction by adopting a weighted spectral subtraction method; and reconstructing the filtered signal. According to the invention, sudden abnormal values and pulse interference can be effectively eliminated; and frequency domain deep noise reduction is carried out through weight spectrum subtraction, so that deep suppression can be carried out in a strong noise frequency band, and weak suppression or reservation can be carried out in a weak noise frequency band and a frequency band possibly containing fault characteristics. According to the method, weak characteristic components related to demagnetization faults can be reserved and even enhanced while background noise is restrained to the maximum extent, loss of diagnosis information is fundamentally avoided, and the diagnosis reliability is improved.

This invention discloses a bridge multi-source fusion positioning system and method based on digital twins, belonging to the field of navigation and positioning technology. Specifically, it involves constructing a bridge digital twin model containing environmental semantic annotation information; collecting BeiDou positioning data, IMU inertial data, and UWB ranging data and performing spatiotemporal synchronization; calculating real-time positioning quality indicators based on BeiDou positioning data; identifying the current environmental state in real time based on multi-source positioning data and predicting future motion trajectories using the digital twin model; adaptively switching positioning states based on the predicted trajectory and real-time positioning quality indicators; fusing multi-source positioning data using a robust improved capacitive Kalman filter algorithm; dynamically adjusting the observation noise covariance based on the Mahalanobis distance of the innovation vector; and outputting the fused positioning result. This invention provides prior environmental information through the digital twin model, and combined with adaptive state switching and robust filtering fusion, enables continuous and reliable centimeter-level positioning in complex environments.

The invention belongs to the field of multi-source data time unification and fusion filtering systems, and relates to a multi-source heterogeneous sensor space-time alignment method based on kinematics bidirectional compensation. According to the method, traditional interpolation limitation is broken through a kinematics bidirectional compensation mechanism, and dynamic weight distribution and robust filtering time deviation closed-loop correction are combined. When satellite data arrives, forward and backward bidirectional calculation is carried out through inertial navigation, so that lag of pure interpolation is avoided, inertial navigation drift can be inhibited through filtering, the influence of abnormal values is weakened, and a high-precision space-time reference can be provided for the fields of automatic driving, unmanned aerial vehicle navigation and the like.

The invention relates to the technical field of high-precision positioning, in particular to a high-precision positioning method, system and device and a medium, and the method comprises the steps: obtaining the original observation data of a Beidou receiver; performing multipath error suppression processing on the original observation data to obtain target observation data; based on the target observation data, searching and fixing the integer ambiguity by using an intelligent algorithm to obtain an integer ambiguity fixed solution; and taking the integer ambiguity fixed solution as a known quantity, resolving the dynamic positioning model by using an adaptive robust filtering algorithm, and outputting a positioning result. The method has the beneficial effects that the positioning precision, the refresh rate and the reliability are integrally improved in a complex mountain area environment, tiny deformation of a dam body and a side slope can be captured in time, and the actual requirements of small hydropower station safety monitoring for high-precision and high-timeliness early warning are met.