127 results about "Particle filter" patented technology

The invention relates to the technical field of laser ranging, in particular to an unmanned aerial vehicle dynamic laser ranging method based on multi-source fusion and adaptive compensation, and the method comprises the following steps: obtaining a ranging pulse of a laser ranging module, and carrying out three-axis rotation and coordinate translation in combination with an IMU angle to generate a geographic registration coordinate set; radar and ultrasonic ranging are fused to construct redundant distance compensation data, particle filtering is executed based on IMU acceleration angular velocity to generate an attitude constraint distance, SLAM weight is adjusted according to illumination and meteorological parameters, and an unmanned aerial vehicle dynamic ranging coordinate result set is generated. Spatial registration is executed by combining longitude and latitude height data, by establishing dynamic association between the ranging data and IMU attitude data, smoothness and continuity of the ranging data can be kept when attitude disturbance is severe, the weight proportion of the vision ranging data in multi-source fusion is dynamically adjusted, and the high-precision estimation capacity of the target distance in the dynamic flight state is improved.

The invention relates to the field of gas leakage identification, and discloses a gas leakage identification method and system based on sound localization, a medium and equipment, and the method comprises the steps: collecting a leakage sound wave signal through a microphone array, and extracting acoustic features in a complex background; simulating a diffusion process of gas in a turbulence environment after gas leakage through an established leakage gas convection-diffusion model, and combining gas concentration distribution in the simulated diffusion process with acoustic characteristics to predict sound source localization at a leakage position; modeling a sound source localization search process as POMDP, and iteratively updating a confidence state of a sound source position through a particle filter; spatial distribution features are extracted from the confidence state through DBSCAN clustering to serve as input of the LSTM-DQN network, spatial and temporal features are fused through the constructed LSTM-DQN network, cross-scene generalization is achieved through transfer learning, and sound source coordinates are dynamically optimized and output through the Q network. According to the invention, the positioning accuracy and real-time performance in a complex environment are improved.

The invention discloses a fuel cell vehicle control method integrating fault diagnosis and energy management. According to the method, vehicle key operation parameters are collected in real time, a whole vehicle operation state vector is constructed, and fuel cell health diagnosis and preliminary power action generation are carried out at the same time based on the state vector. The diagnosis module obtains a health factor and a severity index through residual calculation, particle filtering and generalized likelihood ratio test, and the action generation module outputs an intelligent agent power distribution vector. And according to the severity index, the dynamic reconstruction power, the SOC and the power change rate constraint, optimal power distribution is obtained through optimization solution, and meanwhile, a high-severity state is intervened in combination with a hierarchical fault-tolerant strategy. The method also comprises a demand power prediction and online learning module which is used for improving the robustness and the adaptive capability of the optimization strategy. According to the invention, by tightly combining the health state of the fuel cell with energy management optimization, high-energy-efficiency, long-service-life and reliable dynamic control of the fuel cell system is realized.

The invention discloses an intelligent substation virtual loop online transmission verification method based on GOOSE / SV closed loop test, and relates to the technical field of intelligent substation operation and maintenance. Through precise clock synchronization and an improved cross-correlation algorithm, in combination with wavelet noise reduction and spectral clustering analysis, nanosecond synchronization quality evaluation of GOOSE / SV signals is realized, and the hidden transmission risk discovery time is shortened from regular maintenance to real-time monitoring; a hierarchical hidden Markov model is adopted to analyze equipment-level to system-level behavior modes, real-time probabilistic reasoning is realized in combination with a dynamic Bayesian network and particle filtering, a multi-dimensional evaluation system is constructed, the reliability evaluation capability under complex working conditions is remarkably improved, an optimization scheme is generated based on network path characteristic analysis and bottleneck identification, and the reliability of the system is improved. Multi-scene closed-loop verification is carried out by means of a digital twin technology, safety and reliability of parameter optimization are ensured, and full-process intelligent operation and maintenance of the virtual circuit of the intelligent substation from state perception to optimization verification are realized.

The invention discloses an end-to-end automatic driving long tail scene data acquisition and automatic labeling method. The method comprises the following steps: firstly, collecting multi-modal time series data including images, point cloud and odometer information; thirdly, calculating a comprehensive long-tail scene judgment coefficient by constructing an occupation map and fusing perception uncertainty and prediction uncertainty, thereby automatically identifying a high-risk and high-uncertainty long-tail scene; then, only for the data judged as the long-tail scene, starting a particle filtering-based track truth value generation process, fusing the motion speculation of the odometer and the map matching of the laser radar point cloud in the process, and carrying out state prediction, observation updating, resampling and weighted averaging to obtain a track truth value of the long-tail scene; and generating a high-quality self-vehicle track point sequence conforming to the vehicle kinematics constraint as a truth value label. According to the method, accurate screening and true value automatic generation of the long-tail scene are realized, the data acquisition efficiency and the labeling quality are improved, and the safety and generalization ability of an end-to-end automatic driving model in a complex scene are enhanced.

The invention relates to the technical field of batteries, in particular to a battery capacity prediction method and system and a computer program product. The invention provides a battery capacity prediction method, which comprises the following steps of: obtaining a constant voltage stage time characteristic sequence of a battery, and inputting the constant voltage stage time characteristic sequence into a bidirectional long short-term memory network to obtain an initial prediction sequence; generating a capacity degradation sequence including battery capacity corresponding to the battery based on the capacity degradation model; constructing a mapping relation between the initial prediction sequence and the capacity degradation sequence through a dynamic time warping strategy; determining a target prediction residual error through a particle filtering model and based on the initial prediction sequence and the capacity degradation sequence which have the mapping relation; and correcting the initial prediction sequence through the target prediction residual error to obtain a capacity prediction value. The problem of reference correction error caused by time migration in a traditional battery capacity prediction method is solved, and the prediction precision and long-term stability in a capacity jump scene are improved.

A system is provided for localizing an autonomous work vehicle within mapped agricultural environments when GPS may be unavailable. Range sensors capture LiDAR images that are segmented by a trained model into features such as trunks and canopy. Trunk points are clustered to compute centroids, which are used to estimate left and right tree lines and a lane centerline relative to the vehicle. This sensor-derived centerline is compared with georeferenced lane and row geometry stored in a digital map. A particle filter maintains multiple pose hypotheses, propagates them using vehicle odometry, and weights them according to alignment between the lane centerline relative to the vehicle and the mapped lane geometry for the particle. Resampling and averaging of high-weight hypotheses produces an adjusted pose that corrects for lateral and heading errors. The adjusted pose serves as input to a navigation controller that issues steering and speed commands, enabling accurate autonomous operation.

The invention discloses an intelligent heating control method and system based on a liquid heating container, and the method comprises the following steps: obtaining an observation sequence, completing time alignment and scale normalization, and forming a synchronous observation sequence; carrying out duration inference by using a semi-Markov model, and calculating a residence constraint parameter; constructing a discrete-continuous mixed state space model and executing Rao-Blackwell particle filtering to obtain a particle weight, a phase sequence and posterior edge distribution, and calculating a weight degradation index; generating a noise portrait, and giving a quantile vector, a tail ratio and polarity bias; implementing quantile gating correlation entropy re-calibration according to the noise portrait, and refreshing a weight degradation index; a discrete Schrodinger bridge is constructed in a window with the phase state kept unchanged, deterministic carrying is carried out, and a calibration prior transfer kernel and proposal distribution are formed; and in combination with the weight degradation index and the resident constraint parameter, self-adaptive resampling is carried out, and a heating power track and a safety constraint parameter are output. The state judgment stability and the power control precision are improved.

The invention relates to the technical field of positioning and navigation, in particular to a two-dimensional laser positioning method and system in a sparse scene, and the method comprises the steps: endowing a point cloud with a corner semantic label through a semantic segmentation model through fusing two-dimensional laser radar and camera data, and removing a dynamic object; extracting wall corner line and wall corner point features, performing data association with a pre-constructed semantic map, performing coarse positioning to obtain an initial pose, and performing fine optimization by using a semantic PLICP algorithm to obtain a fine positioning pose; and finally, carrying out confidence evaluation and fusion on the fine positioning pose in a particle filtering mode, and outputting a final optimized pose. According to the invention, the wall corner instance and features in the wall corner semantic point cloud are extracted, nearest neighbor search association and coarse positioning are carried out on the semantic map, the initial pose is obtained, semantic PLICP fine optimization is carried out, and particle filtering fusion is supplemented, so that high-precision and robust positioning independent of manual marking in a sparse scene is realized.

The invention provides a hetero-variance migration network and particle filter-based battery life prediction method, and relates to the technical field of battery health management, and the method comprises the steps: collecting a part of charging data fragments from a target battery under an actual operation condition, and extracting an incremental capacity feature; processing the incremental capacity characteristics by using a heterovariance migration network trained on a tagged source domain and an untagged target domain, outputting a mean value and a variance of a battery health state, and describing an evolution process of the battery health state by using a parameterized double-index degradation function; generating an initial parameter particle by taking a degradation function parameter of source domain fitting as a mean value, and updating the parameter particle by taking a state prediction value output by the heterovariance neural network as an observation value and a corresponding uncertainty variance as observation noise; and finally, predicting the remaining life of the battery based on the updated parameter result. According to the method, cross-working-condition and cross-battery-type residual life prediction is realized, and the prediction precision and robustness are improved.

The invention discloses a battery pack health state prediction method and system driven by a data fusion model, and the method comprises the steps: extracting aging features according to battery test data, and obtaining feature data; according to the feature data, constructing a state space model based on Gaussian process regression, and obtaining an initial prediction model; carrying out transfer learning on the initial prediction model according to the pre-training data set to obtain a transfer model adaptive to the target battery pack; constructing a particle filtering framework according to the migration model, inputting real-time characteristic data acquired on line into the particle filtering framework for state updating and parameter correction, and obtaining a closed-loop estimation result; and outputting a health state prediction value of the battery pack according to a closed-loop estimation result. According to the method, high-precision and high-robustness online prediction of the full-life-cycle health state of the power battery pack is realized, and the generalization ability of the model in a data scarcity scene and the long-term estimation stability in a dynamic environment are remarkably improved.

The invention discloses a nonlinear system dynamic boundary constraint particle filtering method, and belongs to the technical field of deep space exploration. The implementation method comprises the following steps: independently sampling all nodes of the flexible lander, and introducing constraints in a particle filtering initial sampling link to enable initial particles to meet the constraints; in the filtering process, as many particles as possible fall within the constraint range in the initial stage of filtering by widening the constraint boundary, and the constraint boundary is continuously shrunk to be gradually close to the real constraint range, so that the situation that a large number of particles fall out of the constraint boundary due to the fact that the real constraint range is too narrow is avoided. According to the method, the constraint boundary is dynamically adjusted, the number of effective particles is increased, and a higher-precision multi-node state estimation result of the flexible lander is obtained; according to the multi-node state of the flexible lander, the particle diversity is ensured through dynamic scaling constraint boundaries, and filtering stop caused by particle degradation is prevented; and realizing real-time trajectory planning and guidance control of the flexible lander by using a node state estimation result.

The invention belongs to the technical field of positioning and navigation, and provides a geomagnetic sequence positioning method based on BiLSTM and multi-head attention. The method comprises the following steps: firstly, acquiring data of geomagnetism, an accelerometer, a gyroscope and other sensors by using a smart phone, and dividing a geomagnetism sequence into a plurality of subsequences to enhance the time sequence modeling capability; then constructing a multi-scale feature extraction model: extracting deep time sequence features of each sub-sequence by adopting a bidirectional long-short term memory network (BiLSTM), and constructing global representation through splicing; a self-adaptive multi-head attention mechanism is introduced, magnetic signal key features are captured from multiple dimensions, feature weights are dynamically adjusted in combination with a gating mechanism, and important information enhancement is achieved. And the fused multi-scale features output initial position estimation through a full connection layer. In order to further improve the positioning precision, data of a magnetometer and a gyroscope are fused through complementary filtering to correct a course angle, a geomagnetic positioning result serves as an observation value, a pedestrian track plotting result serves as a state prediction value, and final fusion positioning is achieved through particle filtering.

The invention discloses an airborne three-dimensional graphic engine display optimization method based on particle filter optimization, and relates to the technical field of airborne avionics graphics. According to the method, optimization strategies such as a particle filtering optimization technology, a three-dimensional pixelating technology, dynamic scene management, resource optimization management, hardware acceleration, an asynchronous shader technology and optimization renderer design are introduced, so that the rendering efficiency, the display quality and the resource utilization rate of the graphics engine are remarkably improved, and the rendering efficiency of the graphics engine can be greatly improved in a resource limited environment. And high-precision and low-delay three-dimensional graphic rendering is realized, the requirement of the aircraft on real-time display is met, and the method has important application value and popularization prospect.

The application discloses a kind of complex scene under factory area fusion positioning method and system, belong to mobile positioning technical field, applied to mobile device, comprising: judge the number of received position data corresponding ultra-wideband UWB positioning base station whether greater than or equal to first preset value;When the number of UWB positioning base station is greater than or equal to first preset value, the first confidence of position data is calculated according to the signal-to-noise ratio of position data and the signal strength of position data;When the first confidence is greater than the number of UWB positioning base station corresponding to the position data of second preset value is greater than or equal to 3, select the first confidence of the first 3 position data;The observation coordinates of the particle group of the particle filter of mobile device and the second confidence corresponding to observation coordinates are calculated;Wherein, particle group includes N particles, and the second confidence is the average of the first confidence of the first 3 position data.

A high-precision positioning method based on multimodal fusion filtering, relating to the field of intelligent positioning technology, includes the following steps: acquiring sensor data collected in real time by a GNSS submodule, an IMU submodule, and a barometer submodule, and performing collaborative preprocessing on the sensor data; performing state estimation through an extended Kalman filter (EKF) and an adaptive particle filter (APF) to obtain the corresponding EKF positioning output and APF positioning output; calculating a multi-dimensional positioning reliability reflecting the geometric accuracy of the GNSS signal, the stability of the IMU data, and the high consistency between the barometer and GNSS; dynamically allocating fusion weights for the EKF positioning output, APF positioning output, and historical positioning data; and performing weighted fusion of the EKF positioning output, APF positioning output, and historical positioning data to output the final positioning result. This invention, through a three-layer architecture of collaborative data preprocessing, intelligent filtering, and dynamic decision-making, systematically improves the accuracy, continuity, and overall robustness of positioning in complex environments.

This invention provides a target detection method, apparatus, electronic device, and storage medium. The method includes: acquiring echo data received by a vehicle-mounted radar; performing a Fast Fourier Transform (FFT) on the echo data to obtain range-Doppler two-dimensional data; and performing detection processing on the range-Doppler two-dimensional data according to a preset pre-detection tracking model to obtain a target detection result. The preset pre-detection tracking model includes at least a particle filter-based pre-detection tracking model. The preset pre-detection tracking model employs a first preset likelihood ratio function, which includes a background noise amplitude term and a sum term of the target's amplitude and the background noise amplitude. Using this invention can reduce the computational resources consumed by particle filter-based pre-detection tracking algorithms for target detection, thereby reducing its computational cost.

The invention discloses a bearing fault diagnosis method and device based on an optimization algorithm, and a medium, and relates to the technical field of bearing fault diagnosis. The method comprises the following steps: collecting a vibration signal sequence of a target bearing in a current monitoring time period; inputting the fault signals into a pre-calibrated fault signal generation type model library; based on a Bayesian inference framework, taking the vibration signal sequence as evidence, and constructing posterior probability estimation taking a fault physical parameter set as an unknown quantity; carrying out iterative solution on posterior probability estimation by adopting particle filtering in a sequence Monte Carlo optimization algorithm, and generating and updating parameter particles carrying weights so as to approach probability distribution of a fault physical parameter set at the current moment; and based on the probability distribution, calculating the marginal probability of the fault type identifier, determining the type corresponding to the maximum marginal probability as the current fault, and performing extrapolation calculation to obtain residual service life probability distribution. According to the method, the probabilization and quantification of the diagnosis result are realized instead of a single label.

The invention discloses a sodium-ion battery pack health state online evaluation method and system, and the method comprises the following steps: S1, constructing a sodium-ion battery aging model, collecting and preprocessing pack operation data, and generating a data sequence; s2, calculating a feature entropy value, adjusting the state of the liquid neural network, and extracting a fusion feature sequence; s3, performing path signature transformation on the fusion feature and the simulation sequence, calculating a structural deviation, and generating a deviation sequence; s4, updating model parameters by using particle filtering, and reasoning a capacity estimation value to form a capacity sequence; s5, performing Loess fitting and smooth filtering on the capacity sequence to generate a capacity estimation curve; and S6, evaluating the Pack health state according to the capacity estimation curve, and outputting an online result. According to the method, digital twin modeling, a liquid neural network and the like are fused, and the method has the advantages of high evaluation precision, high response speed and high adaptability.

The invention provides an SBA-based underwater object detection and imaging method and system. The method comprises the steps of obtaining original sonar data of an underwater array element; constructing an underwater multi-array element model based on the original sonar data, and performing filtering and denoising; constructing an underwater beam concentration signal model through signal minimum variance beam concentration and sound wave beam concentration signal processing; the cluster signal model is optimized, the optimization comprises signal cluster convolutional neural network model optimization and underwater sound wave particle filtering manipulation mask optimization, and the optimization is used for obtaining an optimized underwater sonar signal; constructing a two-dimensional image of the underwater object by using an SBA algorithm based on the optimized underwater sonar signal; and generating a three-dimensional image of the underwater object according to the two-dimensional image. According to the method, the detection capability and the efficiency level of the ocean underwater sonar are improved and enhanced through the innovatively constructed acoustic wave bundling mathematical model and functional expression operation on the acoustic wave bundling signals transmitted, fed back and recycled by the ocean sonar equipment.

The application discloses a method for determining the elevation of a giant mixed-flow type hydro-generator set during overall replacement of a stator, and belongs to the technical field of installation of hydroelectric generation equipment. The method is characterized by the following steps: arranging laser displacement and inclination sensors on the stator lifting lugs and in the circumferential direction, collecting multi-source heterogeneous data and aligning the data based on millisecond-level timestamps; uniformly mapping the data to a global coordinate system with the center line of the fixed guide vane of the ring as a reference to construct a unified spatial posture framework; dynamically inhibiting displacement fluctuations caused by interference such as elongation of the steel wire rope and deformation of the crossbeam by using Kalman filtering; fusing the filtered displacement and inclination data to determine the tilting state of the stator as a whole by weighted average; further combining particle filtering to estimate the influence of the deformation of the crossbeam on the trajectory of the center of mass, and fitting the trajectory of the center of mass by using the least square method to calculate the elevation value in real time; when the elevation deviation is out of limit, the fusion parameters are optimized through a feedback loop to realize closed-loop regulation and control of the elevation and the posture. The application significantly improves the precision and stability of the determination of the elevation of the center of mass during the hoisting of a kiloton-class stator, and guarantees the safety and efficiency of the overall replacement operation.

The invention discloses a gas well production state estimation method and system based on particle filtering. The method specifically comprises the following steps: S1, collecting production data of different types of gas wells; s2, visual curve drawing is carried out according to the production data of different types of gas wells; and S3, outputting an estimated gas well state through the constructed particle filtering prediction model based on the acquired data of gas well production. Dynamic estimation of gas well parameters is achieved based on particle filtering, the particle filtering serves as a suboptimal estimation framework, and when the number of particles is large enough, the particle filtering can approach the optimal estimation result; meanwhile, compared with Kalman filtering, the particle filtering has the advantages that the particle filtering can also approach optimal estimation for a nonlinear and non-Gaussian process; the gas well state estimation accuracy is improved; meanwhile, the production trend and the occurrence of abnormal conditions are judged through the estimated gas well state, and the management level and safety are improved.

The application discloses a mobile robot positioning method and system based on laser radar scanning information and belongs to the technical field of robot positioning, and aims to solve the technical problem of how to improve the shortcomings of robot particle filtering positioning algorithm in global positioning and following positioning in a large-scale map. The mobile robot is applied to a mobile robot provided with a perception sensor, the perception sensor comprises a mileage meter, an IMU and a laser radar, is used for collecting pose information and environment information, comprises a pre-processing module, a global initialization positioning module and a following positioning module, a scanning matching process based on a multi-resolution grid map and a high-confidence area is adopted, in the robot following positioning module, a high-confidence area and particle genetic mutation are introduced, in the resampling process, a particle is updated by applying a likelihood domain observation model and simultaneously adding a high-confidence area reference, so that the particle set converges to the vicinity of an execution track, the particle convergence effect is good, and the robot following precision is ensured.

The application discloses a kind of unknown distribution's laser interferometer measurement result's uncertainty evaluation method, is used to the measurement result of non-Gaussian and distribution type unknown for uncertainty evaluation;The steps of this method include: 1 using beta distribution method to express measurement result;2 the parameter of beta distribution is regarded as to be estimated parameter, and state space model is established using Bayesian statistical method;3 determine the prior distribution of beta distribution parameter, and the beta distribution parameter of measurement result is estimated recursively based on particle filtering method;4 according to distribution parameter estimated value, the distribution type of interferometer measurement result and its uncertainty are obtained.The method of the application can represent the multiple distribution types of laser interferometer measurement result by beta distribution, so as to solve the optimal estimation and uncertainty evaluation problem of the measurement result of distribution type unknown.

The invention provides a self-adaptive particle filtering method for a maritime maneuvering target, and solves the technical problems of mismatching of motion models and untimely response caused by frequent switching of motion states of an existing maritime maneuvering target. The method comprises the following steps: performing prediction updating on an IMM-PF filter and an MHM-PF filter introducing a multi-hypothesis idea according to target measurement data, and respectively obtaining target state estimation of the IMM-PF filter and the MHM-PF filter; calculating scores of three indexes of a target according to target state estimation of an IMM-PF filter or an MHM-PF filter, performing weighting to obtain a fused original target maneuvering scoring function, and performing smoothing to obtain a target maneuvering probability; and setting a maneuverability discrimination threshold, comparing the maneuverability probability of the target with the maneuverability discrimination threshold, and adaptively fusing the target state estimation of the IMM-PF filter and the MHM-PF filter to obtain the final state estimation of the target. The method can be widely applied to the technical field of target tracking.

The invention discloses an unmanned aerial vehicle-satellite image matching visual positioning method based on particle filtering, which is used for realizing accurate positioning of an unmanned aerial vehicle in a satellite denial environment. The method comprises the following steps: firstly, realizing relative visual odometer positioning through an unmanned aerial vehicle image on the basis of a Light image feature matching algorithm; then, carrying out scattering point sampling according to a relative positioning result, extracting a candidate sub-graph from a satellite remote sensing image, completing absolute positioning of an unmanned aerial vehicle image and a satellite image by using a GIM-dkm feature matching algorithm, and improving cross-domain matching robustness; on the basis, relative positioning and absolute positioning results are input into a particle filtering framework, fusion is realized based on a weighting function, and a stable and accurate final positioning result is output. Experimental results show that the method of the invention effectively combines relative and absolute positioning advantages, and significantly improves the positioning performance of the unmanned aerial vehicle in a satellite denial environment.

A multi-frame coherent detection and tracking method for weak target line spectra is disclosed, based on multi-frame coherent integration. Step 1: Perform multi-frame coherent integration on the passive sonar received data to obtain measured values; Step 2: Establish a state equation including a target presence indicator variable; Step 3: Establish a multi-frame coherent integration measurement equation based on the coherence of the signal phase in the multi-frame data; Step 4: Derive the likelihood ratio function based on multi-frame coherent integration that matches the measurement equation; Step 5: Combine the measured values ​​from Step 1 and the multi-frame coherent detection and tracking model established in Steps 2 to 4, and use a particle filter iterative algorithm to obtain a more robust target line spectrum detection and tracking result. This invention addresses the problem of numerous missed detections in single-frame detection and tracking methods caused by amplitude fluctuations in the received signal in dynamically time-varying marine acoustic channels. It can be applied to most sonar equipment.

The invention discloses a gas source positioning method, an operation control device and a gas source positioning device.The gas source positioning method comprises the steps that an indoor area is scanned through a laser radar to obtain a grid map of the indoor area, initialization processing is carried out on a free area of the grid map to obtain particles representing candidate positions of a gas source, and the particles represent the candidate positions of the gas source; obtaining a gas source distribution state according to the particles; initializing the weight of each grid in the grid map to obtain a map state; the gas source distribution state is sampled and updated through a particle filter, and the map state is analyzed and updated through Bayesian inference; resampling the particles according to the effective particle number and the dynamic threshold value of the particles in the gas source distribution state; and outputting re-sampled candidate sampling points through a density-based noise application spatial clustering algorithm, moving to the candidate sampling points for re-sampling until the particles converge, and outputting the finally estimated gas source position. The gas source can be accurately positioned, and unknown environment exploration can be completed at the same time.

A method may include receiving training data comprising a plurality of RGB-D images of an object at a plurality of time steps, and a plurality of robot actions associated with the object at the plurality of time steps; and optimizing, using the training data, a dynamics function to predict a future state of the object based on a current state of the object and a robot action. A state of the object is estimated as a plurality of particles comprising 3D Gaussians using particle filtering.

The application discloses a kind of vector particle filtering methods for ranging signal SLAM, comprising the following steps: running N groups of particle filters for N ranging signals, estimating signal source position and user positioning;Particle filter carries out filtering estimation, and the value of each particle is updated according to the motion model of user and signal source;The transient weight value of all particles of each group of particle filters is calculated;The transient mean and transient covariance of user positioning state and signal source position state in each group of particle filters are calculated respectively;The auxiliary mean and auxiliary covariance of each group of particle filters are calculated;Each group of particle filters is updated with joint weight: each group of particle filters calculates the local state quantity and local covariance of corresponding user positioning and signal source position;Resampling is carried out on each group of particle filters;Estimate N filters, output the state quantity and covariance of user positioning.This application can reduce the difficulty of particle sampling, improve the real-time performance of algorithm, and is suitable for practical engineering application.