69 results about "Bayesian filtering" patented technology

The invention relates to the field of mechanical equipment health management, in particular to a large sliding bearing fault detection and evaluation method, device and system. Comprising the following steps: collecting multi-source sensing data, and constructing a comprehensive data set; constructing a state space model based on a sliding bearing physical mechanism; the multi-source sensing data and the state space model are fused through Bayesian filtering, and hidden state parameter posterior distribution is dynamically estimated; generating a virtual fault sample by using a generative adversarial network in combination with a physical rule base; designing a Bayesian space-time sequence diagnosis model based on an attention mechanism, and generating fusion health state features; processing and fusing the health state features by using a degradation process model, and predicting the remaining service life of the bearing; and based on the health state, the fault probability and the remaining service life, setting multi-stage early warning threshold values, and triggering intelligent early warning. According to the method, the defect that a single model is insufficient in adaptability and generalization ability under complex working conditions is overcome, and the accuracy and reliability of fault detection are remarkably improved.

The invention discloses a multi-source equipment health collaborative operation and maintenance method in a complex environment based on edge federation, and relates to the technical field of equipment collaborative operation and maintenance. Vibration acoustic emission current waveforms are mapped to a unified time-frequency grid at an edge gateway, and an encrypted sparse index is generated and uploaded; training a global model by combining a graph regular base network with a gradient direction and a distribution distance, and injecting fault information by using a new working condition protection door; after being issued by a sparse adaptation layer of double-temperature-zone distillation and random projection compression, fine adjustment is carried out on site under few samples through temperature gradual fusion and reversible orthogonal mapping, and only unit gradient direction and health labels are uploaded; the center adopts entropy constraint Bayesian filtering to fuse information to generate a health index, a maintenance schedule and a spare part plan are formed by integer programming according to confidence intensity mapping risk popularity, a result is differentially pushed and audited, and a closed loop of collection, learning, evaluation and decision is realized.

The invention provides a multi-level semantic map construction method based on multi-sensor fusion, and the method carries out the construction of an environment grid layer, and comprises the steps: constructing an environment grid map in real time through fusing perception data; scene semantic layer construction: extracting image scene semantic probability distribution by using a deep convolutional network, fusing time sequence observation through Bayesian filtering, and mapping a scene category to a grid unit by using an occupation probability model; constructing an object semantic layer, namely identifying an object by adopting a target detection network in which an information aggregation-distribution mechanism is introduced, extracting an object point cloud, and dynamically updating object semantic attributes of grid units through multi-source observation fusion; and scene atlas generation: constructing a hierarchical scene atlas which takes the marker object as a reference core and comprises a spatial topological relation. According to the method, the dynamic environment adaptability and the multi-modal data fusion precision of semantic mapping are improved, a more visual environment understanding mode is provided for the robot, and the practicability of the semantic map in robot positioning and navigation is improved.

The invention discloses a public building environment monitoring system based on the Internet of Things, and belongs to the technical field of safety monitoring management, and the public building environment monitoring system specifically comprises the steps that a feature extraction module obtains Wi-Fi RTT time delay features, Bluetooth RSSI features and UWB TDoA features in parallel through a multi-mode feature extractor, and marks environment quality marks for each type of features; the signal attenuation correction module performs joint correction on a Wi-Fi RTT refraction error, a Bluetooth signal absorption error and a UWB multi-path reflection error based on a physical attenuation model and an offline training environment response curve, and outputs a corrected feature vector; the filtering fusion module is used for firstly applying Kalman filtering in a time domain and then applying particle filtering in a space domain through a multilayer Bayesian filtering fusion engine, fusing the corrected feature vectors and generating a target dynamic grid cell point probability heat map; and selecting a unit with the highest posterior probability from the target dynamic grid unit grid point probability heat map as a final alarm position.

This invention relates to the field of hydraulic engineering, and more particularly to a method for safety assessment and optimization design of large-flow aqueduct structures. The method includes: collecting displacement, strain, and temperature monitoring sequences based on a triple probe, completing grouping, coordinate and time registration, and establishing a monitoring baseline; constructing a spatial model including support nonlinearity and fluid-structure boundary conditions, applying loading conditions, and obtaining response predictions; aligning monitoring and prediction in time and location, performing hierarchical Bayesian filtering inversion, and obtaining parameter posteriors and structural health; under the integration of in-service monitoring and accelerated degradation characteristics, using multi-task learning to generate material evolution and lifespan intervals, and automatically generating and simulating optimization schemes under reliability and maintainability constraints. This application forms a closed loop from assessment to design, effectively enhancing data comparability, traceability of uncertain information, and adaptability of the scheme to engineering projects.

The invention relates to the technical field of market trend prediction, and discloses a market trend prediction system driven by macroeconomic indicators. An index acquisition module of the system dynamically acquires core economic indexes such as GDP growth rate, CPI, PMI and currency supply; the data preprocessing module is used for carrying out layered noise reduction processing on the multi-source heterogeneous data; the feature engineering module constructs a market sensitive feature set through spatio-temporal feature fusion; the prediction model building module is used for building a multi-layer nonlinear prediction model based on a deep belief network; the dynamic adjustment module adopts reinforcement learning to optimize a decision threshold value and combines a Markov chain to carry out state transition planning; and the feedback iteration module analyzes and predicts deviation through Bayesian filtering and realizes strategy updating. According to the method, deep learning and reinforcement learning technologies are creatively fused, the prediction precision is remarkably improved through a dynamic calibration mechanism, and the method can be widely applied to the macroeconomic analysis fields of financial investment, industrial planning and the like.

Computer-implemented method (30) for calculating an estimate of a parameter characterising concrete (e.g. slump) in a drum (101) of a concrete mixing vehicle (100), comprising the steps of: - receiving (31) a drum speed value indicative of a rotational speed of the drum; - receiving (31) a torque value indicative of a torque required to turn the drum; - providing (32) the drum speed value and torque value as input measurements to a relationship estimator Bayesian filter; - calculating (33) at least one vehicle parameter characterising a relationship between drum speed and torque using the relationship estimator Bayesian filter, wherein the relationship estimator Bayesian filter is configured such that the at least one vehicle parameter comprises a hidden state variable of the relationship estimator Bayesian filter; and - calculating (34) an estimated parameter value by providing the at least one vehicle parameter as an input to a parameter calculation model.

The invention relates to the technical field of obstacle avoidance path planning, and discloses an intelligent sports vehicle dynamic obstacle avoidance path continuous planning method comprising the following steps: S1, obtaining a racing track pre-stored map, and constructing a racing track reference coordinate system; relates to the technical field of obstacle avoidance path planning, and the method comprises the steps: predicting the probability distribution of the future position of an obstacle through employing a recursive Bayesian filtering algorithm based on the historical observation data of the obstacle, carrying out the modeling through multivariate normal distribution, converting the modeling into a geometric confidence region, and forming a confidence occupancy set covering the uncertainty of the obstacle; a space-time safety channel is generated through set operation by combining the intelligent vehicle appearance envelope sequence and the racing track boundary; the method does not need to depend on a single prediction track, and by depicting the uncertainty of obstacle movement, even if the actual position of the obstacle deviates from prediction, the intelligent vehicle can still run in the safety channel, so that the problems of planning failure and conflict between the vehicle and the obstacle caused by deterministic prediction of the actual position of the obstacle in the prior art are effectively relieved.

The invention relates to the technical field of satellite communication, and discloses a low-orbit satellite communication Doppler frequency offset processing method, system and device and a medium, and the method comprises the steps: separating a pilot signal from a received signal; taking the Doppler frequency as a first system state, taking the received signal as first observation data, and performing Doppler frequency offset estimation on the pilot signal by adopting a frequency offset estimation algorithm to obtain a frequency offset estimation value; taking the Doppler frequency as a second system state, taking the frequency offset estimation value as second observation data, and adopting a frequency offset tracking algorithm to perform Doppler frequency offset tracking on the pilot signal to obtain a frequency offset tracking value; and performing frequency offset compensation on the received signal according to the frequency offset tracking value to obtain the received signal after frequency offset compensation. According to the method, the Doppler frequency offset is subjected to streaming processing through Bayesian filtering in the estimation and tracking stages, the complexity of data storage and calculation is effectively reduced, the data updating rate is increased, and therefore the accuracy of low-orbit satellite data is effectively improved.

The invention provides a chip data detection method and system based on artificial intelligence, and relates to the technical field of chip detection, and the method comprises the steps: obtaining the dynamic power data of a chip; constructing a chip data detection model based on the random forest; judging whether the structural complexity of the chip is smaller than preset structural complexity or not; if yes, carrying out denoising processing on the dynamic power data by adopting a Bayesian filtering algorithm to obtain first denoised power data, and carrying out anomaly detection through a chip data detection model; and otherwise, de-noising the dynamic power data by adopting a bidirectional long-short time memory network to obtain second de-noised power data, acquiring logic activity data of the chip, and performing anomaly detection through a chip data detection model by jointly using the second de-noised power data and the logic activity data. According to the invention, the structural complexity is de-noised by using the Bayesian filtering algorithm and is processed by using the bidirectional long and short time memory network, so that the detection efficiency and accuracy are effectively improved.

The invention discloses an environment situation scanning three-dimensional reconstruction method based on infrared and SLAM, and belongs to the technical field of computer vision and three-dimensional reconstruction. According to the method, thermal radiation information and environment image information of a target environment are obtained through infrared sensing and visual imaging respectively, after feature extraction and matching processing are conducted on the two kinds of information, synchronous positioning and map construction are completed, and environment pose data and environment map data are generated; a sensing uncertainty probability model is established, after multi-source data denoising is completed through Bayesian filtering, feature level fusion is carried out to obtain fusion environment data, and finally a three-dimensional reconstruction model of a target environment is constructed based on the fusion environment data. According to the method, the full-process closed loop of environment situation scanning and three-dimensional reconstruction is realized, the application limitation of single sensing information is made up, the comprehensiveness of environment sensing data and the scene adaptability of three-dimensional reconstruction are improved, and the digital reconstruction requirements in various complex environments can be met.

The invention discloses a digital twin co-evolution method, and belongs to the field of structural health monitoring and prediction. Comprising the following steps: constructing life prediction digital twins; carrying out online load-damage hybrid monitoring and fusion diagnosis; the digital twin co-evolution based on the Bayesian filtering model is driven; collaborative updating of the crack propagation simulation finite element model and the stress intensity factor agent model is achieved; and performing crack propagation and life prediction based on the digital twin updated by co-evolution. Through a multi-model closed-loop co-evolution mechanism, the problem of model misalignment under the influence of a strong uncertain environment is effectively solved, the accuracy and reliability of structural damage diagnosis and life prediction are remarkably improved, and a core technical support is provided for implementing accurate condition-based maintenance.

A wireless communications system includes a network of acoustic modems for communicating messages between downhole equipment and a surface control and telemetry system. The acoustic modems include multi-channel acoustic receivers that are uniquely deployed along an acoustic transmission medium to provide spatial diversity. Acoustic signals received on the multiple receiver channels are combined and filtered using a Bayesian-type filter to reduce noise.

The invention discloses an extended target tracking method based on a long-short-term memory network. The method comprises the following steps: learning motion characteristics of a target by using the long-short-term memory network, capturing non-Markov characteristics of long-term motion of the target by using long-term memory, and capturing speed change characteristics of short-term motion of the target by using short-term memory; priori knowledge and a long and short term memory network are combined, and under a reasonable assumption condition, recursion forms of a prediction formula and an updating formula of the extended target tracking method are deduced based on a Bayesian filtering framework; a memory link, a prediction link and an updating link of an algorithm combining the neural network and the prior knowledge are constructed, and the three links are executed in a recursive mode to achieve tracking of the extended target. According to the method, the tracking precision of the extended target can be effectively improved.

This invention relates to the field of environmental perception fusion positioning technology, and provides a positioning method, system, and electronic device based on a Bayesian filtering neural network. The method includes: acquiring an observation waveform matrix generated by a base station after receiving a waveform signal transmitted by a target object at each sampling time; inputting the observation waveform matrix into a preset positioning model to obtain the positioning information of the target object; fitting the Bayesian filtering result to the positioning model using a recurrent neural network, which is trained using a cost function including motion likelihood and observation likelihood. The motion likelihood is generated based on the motion model of the target object sample, and the observation likelihood is generated based on the position information of the target object sample determined by a parameter estimation algorithm. This invention addresses the shortcomings of existing technologies where the application of Bayesian filtering algorithms in environmental perception fusion positioning requires approximation algorithms and complex derivations to obtain positioning results, thus affecting positioning speed and accuracy.

This invention relates to the field of satellite communication technology and discloses a method, system, device, and medium for processing Doppler frequency offset in low-Earth orbit (LEO) satellite communication. The method includes: separating a pilot signal from a received signal; using the Doppler frequency as a first system state and the received signal as first observation data, performing Doppler frequency offset estimation on the pilot signal using a frequency offset estimation algorithm to obtain an estimated frequency offset value; using the Doppler frequency as a second system state and the estimated frequency offset value as second observation data, performing Doppler frequency offset tracking on the pilot signal using a frequency offset tracking algorithm to obtain a tracked frequency offset value; and performing frequency offset compensation on the received signal based on the tracked frequency offset value to obtain a frequency offset-compensated received signal. This invention uses Bayesian filtering for streaming processing of the Doppler frequency offset during the estimation and tracking stages, effectively reducing the complexity of data storage and computation, increasing the data update rate, and thus effectively improving the accuracy of LEO satellite data.

The present disclosure provides a system (104) and a method (800) for vehicle localization in a tunnel, utilizing sensor data and tunnel-specific features. The system (100) is configured to determine the lateral and longitudinal positions of the vehicle (101) in real time. The lateral position is identified by updating a light probability distribution and applying Bayesian filtering to establish the vehicle's ego lanes. Furthermore, a real-time landmark identifier (ID) for lights is generated based on normalized tunnel features, including distances, light intensity, and tunnel geometry. The longitudinal position is determined by matching the generated landmark ID with predefined map data, enabling robust localization for navigation and thus allowing the vehicle (101) to be located in tunnels.

This invention discloses an intelligent tracking method for maneuvering swarm targets based on deep neural networks. This method incorporates deep learning into a Bayesian filtering framework, extracting the motion characteristics and process noise statistical characteristics of the swarm targets from measurement data using multiple deep neural networks. It then estimates the motion state transition matrix and process noise variance matrix of the swarm targets online, and uses Bayesian filtering to accurately estimate the motion state of the swarm targets' centroid. By modeling the swarm target contour as an ellipse centered at the centroid, the high-precision centroid motion state estimation results improve the contour estimation accuracy, ultimately achieving swarm target tracking. Compared to existing swarm target tracking methods, the proposed method does not require prior information to establish a target motion model, enabling high-precision tracking of non-cooperative maneuvering swarm targets in complex environments where prior information is lacking.

A computer includes a processor and a memory, and the memory stores instructions executable by the processor to associate detected objects from sensor data with tracked objects in a tracking list stored in the memory by executing an association model, determine kinematic states of the tracked objects by executing a Bayesian filter, and update which of the tracked objects are stored in the tracking list by executing a track-management model. The association model includes association weights and outputs probabilities that the detected objects are the same as the tracked objects. An association-model gradient of the association model is computable. The association-model gradient is of the probabilities with respect to the association weights. The Bayesian filter includes Bayesian parameters. A Bayesian-filter gradient of the Bayesian filter is computable. The Bayesian-filter gradient is of the kinematic states with respect to the Bayesian parameters. A track-management gradient of the track-management model is computable.

The invention relates to the technical field of hard capsule polishing, and discloses an automatic polishing system based on hard capsules and a device thereof. A data acquisition module of the system acquires surface characteristic data such as hard capsule surface roughness, stain distribution, capsule size and the like; the data analysis module receives the data and then obtains a polishing demand analysis result through random forest regression analysis; the optimization module adopts a genetic algorithm to optimize parameters and rules of the polishing controller according to the result; the control module carries out fuzzy logic calculation on the polishing amount through particle swarm optimization according to the optimized parameters and rules and converts the polishing amount into polishing control signals; after the execution module receives the signal, the polishing amount constraint condition is adjusted through a near-end gradient method, and integer programming is conducted on the polishing process in combination with a branch and bound method so as to automatically adjust the polishing amount; and the detection module detects surface characteristic data after polishing through Bayesian filtering and feeds back the processed data to the control, data analysis and optimization module so as to adjust a polishing strategy.

The application relates to a danger judgment and active guidance method and system for power construction safety supervision, and belongs to the technical field of power construction safety. The method comprises the following steps: collecting multi-modal data of a construction scene through intelligent glasses; analyzing the data at an edge computing end to obtain a scene semantic category and an interactive behavior of a worker; according to the scene semantic category, dynamically scheduling and hot loading a matched risk identification model set from a model library to identify a risk factor and a three-dimensional position thereof; acquiring a three-dimensional voxelized static danger level base map; mapping the risk factor into a dynamic risk observation point cloud in a voxel space; taking the static base map as a priori and taking the dynamic point cloud as observation evidence, performing recursive Bayesian filtering through a dynamic Bayesian network to generate a dynamic danger level three-dimensional distribution; generating and rendering graded AR early warning information; when a violation trend occurs, generating and rendering a three-dimensional safety path of a risk avoidance area. The method realizes adaptive, accurate and active construction safety supervision.

The invention discloses a voiceprint continuous tracking system based on multi-mode space-time fusion and SE (3) manifold optimization and a tracking method thereof. The tracking method comprises the following steps: S1, acquiring an acoustic signal from a microphone array deployed on a user wearing device, inertial data from an inertial measurement unit, and optionally visual data from a camera; s2, estimating the real-time pose of the equipment worn by the user, wherein the pose is expressed on the SE (3) manifold; s3, defining a joint state vector of the sound source, wherein the state vector comprises a three-dimensional position and a voiceprint feature vector of the sound source in a world coordinate system; and S4, constructing a Bayesian filtering framework running on the SE (3) manifold, and utilizing the real-time pose. A wave field synthesis theory and a neural radiation field technology are combined, an extended acoustic NeRF model is constructed, a dynamic holographic sound field and spatially distributed voiceprint features are reconstructed in real time by using an estimated sound source state and a glasses pose, and visualization is carried out in XR equipment.

This invention discloses an intelligent multi-target association tracking method (BAIT) based on a Bayesian inference Transformer architecture, combining the advantages of classical Bayesian filtering (BF) recursive inference with the Transformer's ability to handle long sequence tasks. First, BAIT employs a state predictive encoder during the prediction process to fully extract past target motion information. Then, mimicking the classical BF recursive inference structure, an association decoder is used between state prediction and filter update to achieve optimal matching and association between the target and the measurement. Next, based on the target's past motion information and the association result, a state update decoder is used to estimate the target's motion state in the current frame. Finally, by combining the classical BF iterative inference structure, BAIT can achieve high-precision target association tracking in complex data association scenarios. This method has advantages such as high precision, continuity, and accurate association, and can be applied in many fields, including military and civilian applications.

The invention relates to the technical field of agricultural meteorological early warning, and discloses a greenhouse crop meteorological early warning system based on artificial intelligence. The system comprises a data acquisition module, a data analysis module, a model training module, an early warning generation module, a strategy adjustment module and a feedback verification module. The data acquisition module optimizes an acquisition process and acquires parameters such as air temperature and humidity; the data analysis module obtains an analysis result through sliding window statistics; the model training module predicts model parameters and rules through particle swarm optimization; the early warning generation module combines a genetic algorithm to calculate a meteorological risk and converts the meteorological risk into an early warning control signal; the strategy adjusting module automatically adjusts an early warning strategy through dynamic planning and a branch and bound method; and the feedback verification module verifies the data through Bayesian filtering and feeds back and corrects the data. According to the system, intelligent early warning and strategy optimization of greenhouse meteorological risks are realized, the early warning accuracy and timeliness are improved, the system adapts to dynamic changes of a greenhouse environment, and reliable guarantee is provided for growth of greenhouse crops.

The present application relates to the technical field of blood oxygen detection, and more particularly to a pulse blood oxygen saturation detection method based on image processing, which introduces a transfer learning strategy and a domain adaptation network, can be adjusted according to the individual physiological parameters of the user, thereby improving the applicability to different populations. In particular for dark-skinned population, the present application significantly reduces the measurement error, reduces the error rate by more than 30%. The present application designs a set of real-time calculation mechanism of current signal quality index Q, and combines the Bayesian filtering model for dynamic calibration, so as to ensure that reliable blood oxygen saturation estimation value can be obtained even under low signal-to-noise ratio condition. When the signal quality decreases, the historical data is compensated for the current detection value through the LSTM network prediction module, so as to further ensure the accuracy and continuity of the result.

The application discloses a skipping counting detection method based on laser radar area recognition, and relates to the technical field of motion monitoring, which comprises the following steps: continuously scanning each effective detection ROI area to obtain original point cloud data, and acquiring time sequence point cloud data through preprocessing; training a multiple hidden Markov model based on the time sequence point cloud data, and independently modeling the state of a rope in parallel, and solving the optimal state sequence of each rope through a Viterbi algorithm; based on the optimal state sequence, applying Bayesian filtering to dynamically optimize cycle parameters, and acquiring a complete counting sequence through interpolation counting; performing event counting statistics on the complete counting sequence, aligning through a time window, mapping the effective detection ROI area, and outputting the independent counting value of each person. Through the parallel independent modeling mechanism of the multiple hidden Markov model, the application realizes accurate separation and recognition of a multi-person skipping scene; through the Viterbi algorithm, the optimal state sequence is solved, and the recognition accuracy and stability are maintained.

A method is for tracking an object using an environment sensor. The object is represented by an object status. The method includes detecting a sensor value of the environment sensor, predicting a future object status of the object, and updating the object status using a Bayesian filter. The updating includes using an artificial intelligence module (“AI module”). The AI module is trained such that the detected sensor value is associated with the object and the object status of the object is updated based on the predicted future object status of the object and the detected sensor value.

The invention discloses a rope skipping counting detection method based on laser radar area identification, and relates to the technical field of motion monitoring, and the method comprises the steps: carrying out the continuous scanning of each effective detection ROI area, obtaining original point cloud data, and obtaining time sequence point cloud data through preprocessing; a multi-hidden Markov model is trained based on the time sequence point cloud data, parallel independent modeling is carried out on the states of the ropes, and the optimal state sequence of each rope is solved through a Viterbi algorithm; based on the optimal state sequence, applying Bayesian filtering to dynamically optimize periodic parameters, and obtaining a complete counting sequence through interpolation counting; and performing event counting statistics on the complete counting sequence, and outputting an independent counting value of each person through time window alignment and effective detection ROI mapping. According to the invention, through a parallel independent modeling mechanism of the multi-hidden Markov model, accurate separation and identification of a multi-person rope skipping scene are realized; the optimal state sequence is solved through the Viterbi algorithm, and the recognition accuracy and stability are kept.

This application relates to a method and apparatus for characterizing the quantity-time relationship of pharmacokinetics of monomeric components of traditional Chinese medicine (TCM). Addressing the problem that traditional methods, which directly utilize noisy experimental data to invert pharmacokinetic parameters of TCM monomeric components, suffer from significant errors, making it difficult to reliably reveal metabolic patterns in different experimental individuals, this application proposes a method using modern statistical parameter estimation techniques to dynamically track and analyze the ADME process of TCM monomeric components. Specifically, the metabolic patterns of TCM monomeric components in vivo are modeled as state variables of a dynamic system, and the measured random experimental data are modeled as noisy observation data. Bayesian filtering techniques are then used to dynamically track and analyze the ADME process of TCM monomeric components, achieving accurate, robust, and reliable characterization of the pharmacokinetic behavior of TCM monomeric components.

The invention discloses a multi-target tracking method and system based on tandem joint probability data association and variational Bayesian filtering, electronic equipment and a medium, and the method comprises the following steps: 1, predicting each target, and calculating the interconnection probability between each measurement and the target; step 2, generating equivalent fusion measurement and equivalent covariance for each target; step 3, performing dynamic correction on the fusion measurement noise covariance of each target; 4, updating and iterating each target by using Kalman filtering, and returning to the step 3 if the number of iterations is less than the set number of iterations; otherwise, storing the optimal estimation value and the posterior error covariance of each target. According to the method, the calculation complexity and the resource consumption are remarkably reduced, and meanwhile, the multi-target state estimation precision and the system robustness are synchronously improved.