71 results about "One step prediction" patented technology

The invention discloses a multi-modal thinking chain-based multi-modal large language model attribute prediction method, which comprises the following steps of: obtaining a multi-modal large language model, and forming a multi-modal attribute prediction framework together with a mask generator and a scene graph analyzer; designing an inference method of a hierarchical thinking chain, inputting an attribute prediction task into a multi-modal attribute prediction framework, extracting the hierarchy of an attribute table, and generating a hierarchical and formatted inference prompt; decomposing an attribute prediction task into hierarchical sub-tasks, and selecting a prompt method for the next step of prediction of the model; supervising a model prediction process through critical thinking; identifying object categories in the image by adopting a logic check prompt method, and extracting corresponding attribute values in combination with common knowledge in the model; and combining the generated scene graph with common knowledge in the model to check the predicted attributes. According to the method, the context understanding capability, the logic consistency, the task universality and the model reasoning efficiency of the attribute prediction task can be effectively improved.

The invention provides a self-adaptive robust vehicle navigation method and device based on LSTM assistance, and the method comprises the steps: calculating a process noise scale factor through process noise covariance estimation, and dynamically adjusting a process noise covariance matrix Q in a Kalman filter; executing a prediction process of Kalman filtering to obtain a one-step prediction state value and a state prediction covariance matrix; calculating a robust factor and a mahalanobis distance adaptive factor based on the prediction residual vector, and performing expansion adjustment on the prior measurement noise covariance matrix R and the predicted state prediction covariance matrix P; executing an updating process of Kalman filtering to obtain a state optimal estimation and a covariance matrix at the current moment; and when the GNSS loses lock, pseudo GNSS extended Kalman filtering is carried out by using a pseudo GNSS measurement value generated by the LSTM neural network and combining a process noise covariance estimation method. The method can better adapt to the dynamic characteristics of the navigation system in different environments, and provides powerful support for the stability and reliability of the combined navigation system.

The embodiment of the invention discloses a digital key ranging value filtering method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring ranging information of a plurality of UWB anchor points in real time; obtaining the position change trend of the digital key according to the distance measurement information; under the condition that the effective distance measurement value does not exist, a preset target position is obtained according to the positioning result corresponding to the moment t of the previous effective distance measurement value, the position change trend and the unlocking and locking state of the vehicle terminal, and a predicted distance is obtained according to the distance relation between the effective distance measurement value at the moment t and the preset target position; obtaining a prediction duration T according to the prediction distance; in the T, when an effective distance measurement value is not detected, state one-step prediction of a Kalman filtering algorithm is executed to obtain a filtered distance measurement value, and when the effective distance measurement value is detected, the state one-step prediction is terminated, and a filtering estimation value is obtained according to the Kalman filtering algorithm to serve as the filtered distance measurement value; and obtaining a positioning result of the area outside the vehicle according to the filtered distance measurement values of the plurality of anchor points.

The invention relates to the technical field of radar detection, and provides a multi-model filtering target positioning method, system, device, product and medium, and the method comprises the steps: determining a model matching parameter, carrying out the initialization assignment, obtaining an initialization parameter, and obtaining a sampling point set and a weight coefficient through the initialization parameter; performing state one-step prediction to obtain a state one-step prediction value, and performing secondary sampling and measurement one-step prediction to obtain an autocorrelation covariance and a mixed covariance; carrying out SVD operation to obtain a decomposition value set, detecting a target by a target radar to obtain a radar detection value, obtaining a Kalman gain matrix according to the mixed covariance, and calculating state estimation and an initial covariance; updating the model probability to obtain a Markov transition probability, and normalizing the Markov transition probability to obtain a normalized probability; and state estimation and target covariance are obtained through state fusion, so that the target orientation is determined. According to the invention, accurate tracking of the high maneuvering target is realized.

The invention discloses a coordinate system transformation fusion filtering tracking method and system for a dual-base-station radar, and relates to the technical field of radar target tracking. The method comprises the following steps: firstly, establishing a state and observation equation, and determining an initial state of a target based on prior information under Cartesian coordinates; then, obtaining a target motion state through one-step prediction, generating a sigma point by utilizing U transformation, and calculating a mean value and a covariance for constructing state space prediction; thirdly, fusing the measurement data and the prediction data by using a Kalman filter to obtain optimal state estimation and a covariance matrix; and finally, converting the updated state estimation back to the Cartesian coordinate system based on U transformation, and circulating the process until the tracking is finished. According to the method, the nonlinear filtering problem in the updating process can be avoided, and the tracking robustness and precision are improved.

A secure localization method for multi-mobile robots system based on network communication includes: Step 1, establishing a nonlinear dynamic model of a multi-mobile robots system based on network communication; Step 2, designing a secure estimator for the nonlinear dynamic model; Step 3, calculating an upper bound on a one-step prediction error covariance matrix Σi,k+1|k for each mobile robot in the network communication; Step 4, based on Σi,k+1|k, calculating an estimator gain matrix Ki,k+1 for each mobile robot in the network communication; Step 5, substituting the estimator gain matrix Ki,k+1 calculated in Step 4 into Step 2 to obtain a state estimation {circumflex over (x)}i,k+1|k+1 at time k+1; determining whether k+1 reaches a total duration M, that is, if k+1<M, performing Step 6, and if k+1=M, ending; and Step 6, based on Ki,k+1, calculating an upper bound on an estimation error covariance matrix Σi,k+1|k+1 of each mobile robot; let k=k+1, and performing Step 2 until k+1=M is satisfied.

The invention provides a model-free current prediction control method for a permanent magnet synchronous motor based on a Lyapunov function method, and the method comprises the steps: building a q-axis control channel, constructing a discrete prediction current model without motor parameters, and introducing a parameter alternative item vector to represent the influence of the precise parameter change of the motor on a prediction model; performing first-step prediction on the current state by using the discrete prediction current model; constructing a candidate voltage set, calculating a measurable output quantity regression error by utilizing a regression error model according to the current operation state data, estimating a current parameter alternative item vector by utilizing the parameter alternative item vector self-adaptive law based on the regression error, and for each candidate voltage, calculating the current parameter alternative item vector according to the parameter alternative item vector self-adaptive law; performing second-step prediction on the current state by using the discrete prediction current model; calculating a cost function corresponding to each candidate voltage, and selecting a cost candidate voltage as an optimal control voltage; according to the method, the robustness of the system is improved, and the calculation complexity is reduced.

The disclosed techniques automatically identify cyber-security attacks and predict attack next steps. Descriptions of previously observed cyber-attack campaigns are decomposed into attack campaign steps. Real-time security incident signals are generated by cybersecurity software. Attack campaigns are identified by mapping attack campaign steps to security incident signals. Custom-generated telemetry queries are executed to determine if a missing attack campaign step occurred. A machine learning model generates embeddings for attack campaign steps, security incident signals, and telemetry query responses. A security incident signal or a telemetry query response matches an attack campaign step when their embeddings are within a defined distance. A security alert may be raised when most or all of the attack campaign steps of a particular attack campaign are matched. Attack campaign steps that are not matched to security incident signals or telemetry query results are predicted as attack next steps.

The invention provides a method for accurately estimating target geometric features under laser radar point cloud integration. The method is divided into a modeling stage and an operation stage. In the modeling stage, a system state equation and a measurement equation are established according to the relation between the real geometrical characteristics and motion characteristics of a target during point cloud integration, the relation between the geometrical characteristics detected by a sensor and a sampling period and the relation between point cloud integration times. In the operation stage, firstly, original point cloud is preprocessed, a non-ground point cloud set is segmented, and then point cloud integration and clustering are performed to realize target detection, so that a target measurement set is obtained; secondly, carrying out one-step prediction, calculating target similarity, and carrying out target and trajectory association by utilizing the similarity; and finally, the optimal estimation of the geometric features and the motion features of the target is obtained by using Kalman filtering. According to the method, the precision of target geometric feature estimation under the point cloud integration method can be improved, the small target detection performance of the low-line-number laser radar is enhanced, and the cost of a target detection system based on the laser radar is reduced.

The invention belongs to the technical field of hydroelectric generating set vibration prediction, and particularly relates to a hydroelectric generating set mechanical vibration trend prediction method and device, equipment and a medium, and the method comprises the steps: processing a vibration signal of a hydroelectric generating set through employing a variational mode decomposition algorithm; training the deep belief network by using the training set to obtain a trained deep belief network model; using a sparrow search algorithm to optimize parameters of the deep belief network model to obtain a vibration trend prediction model; and collecting a vibration signal of the hydroelectric generating set, processing the vibration signal by using a variational mode decomposition algorithm, inputting the processed vibration signal into the vibration trend prediction model to obtain a vibration prediction value, and taking the obtained vibration prediction value as an input of next prediction to realize multi-step prediction of the vibration trend of the hydroelectric generating set. By means of multi-step prediction, operation and maintenance personnel can know the vibration trend of the hydroelectric generating set in a period of time in the future in advance, potential fault hidden dangers can be found in time, and the operation safety and stability of the hydroelectric generating set are improved.

The invention discloses a digital Kalman gain rapid calculation method for a time-sharing uniformly-accelerated radar target, which comprises the following steps of: further coupling a Kalman gain calculation formula at the kth moment and a state one-step prediction covariance equation at the (k-1) th moment by extracting the characteristics of a prediction equation and an update equation in Kalman filtering, and calculating a matrix HPk in advance to obtain the Kalman gain of the kth moment. And the k-1HT + Rk achieves an optimization effect. According to the method, the distribution rule of inverse matrix elements of a matrix HPk, k-1HT + Rk is deeply mined by utilizing the mathematical characteristics of the matrix HPk, k-1HT + Rk, so that the rapid calculation of the inverse matrix is realized. The improvement greatly improves the calculation speed of the Kalman gain. Compared with a traditional Kalman filtering algorithm, the method shows greater potential in the aspect of parallel calculation, the calculation complexity is effectively reduced, and the execution speed of the Kalman filtering process is remarkably increased. According to the invention, a more efficient solution is provided for real-time tracking and prediction of the space target, and the method has significant advantages especially in a scene where computing resources are limited or high-frequency updating is needed.

This invention discloses a distributed optimization state estimation method for asynchronous sampling rates under stochastic topology. The method is as follows: 1. Establish an asynchronous sampling rate time-varying nonlinear dynamic model with different system state update rates and measurement sampling rates; 2. Transform the dynamic model into a single-rate time-varying nonlinear dynamic model; 3. Design a state estimator under stochastic topology; 4. Calculate the state estimator at q... k The gain matrix K of the estimator at time step i (q k ) and G i (q k ); V. K i (q k ) and G i (q k Substituting this into the estimator, we obtain q. k+1 VI. Calculate the upper bound of the one-step prediction error covariance matrix; VII. [The text abruptly ends here, likely due to an incomplete sentence or missing information.] i (q k ) and G i (q k Substitute the upper bound of the one-step prediction error covariance matrix into the formula to calculate q. k+1 The minimum upper bound of the one-step prediction error covariance matrix at time step q; let q k =q k+1 Execute step three until q is satisfied. k+1 =K. This invention solves the problem that existing state estimation methods cannot simultaneously handle distributed optimization state estimation problems with stochastic nonlinearity and asynchronous sampling rates under stochastic topologies.

The invention discloses an event scheduling distributed estimation method of a renewable micro-grid system under a sensor network framework. The method comprises the following steps: 1, establishing a dynamic model of the renewable micro-grid system under the sensor network framework; 2, constructing a distributed estimator; 3, deriving one-step prediction of the ith sensor node at the kth moment; 4, calculating a one-step prediction error covariance upper bound of the ith sensor node at the kth moment; 5, deducing distributed estimator parameters of the ith sensor node at the (k + 1) th moment through the trace of the upper bound of the minimum estimation error covariance; 6, deducing state estimation of the ith sensor node at the (k + 1) th moment; and 7, solving the upper bound of the estimation error covariance of the ith sensor node at the (k + 1) th moment. According to the method, the problem that the existing state estimation method cannot process the state estimation of the renewable micro-grid system under the sensor network framework with inaccurate occurrence probability loss measurement and a dynamic event triggering mechanism at the same time is solved.

The invention provides an MEMS navigation attitude correction method and device based on motion acceleration estimation, and belongs to the field of MEMS navigation algorithms. The method comprises the following steps: carrying out filtering and moving average processing on real-time angular velocity and acceleration data, adding Markov modeling on a motion acceleration state on the basis of a current navigation system model, and carrying out one-step prediction through state transition. And when the heading attitude filtering admission condition is met, separating the motion acceleration in the measurement matrix, and estimating the current navigation state of the carrier by using Kalman filtering. And finally, east and north mathematical platform deflection angles with motion acceleration influence deducted are used during filtering correction. According to the method, accurate modeling can be carried out on the motion acceleration under the large maneuvering condition, the attitude error correction probability of the MEMS attitude and heading system is reduced, the convergence time after a carrier is converted into stable flight from maneuvering is shortened, and the product performance and the use experience are improved.

The invention discloses a time-varying output coupling complex network state estimation method in a low-reliability communication environment, and belongs to the technical field of distributed state estimation methods. Comprising the following steps: S1, establishing a time-varying output coupling complex network model with random uncertainty probability measurement delay; s2, a periodic communication mechanism is introduced for scheduling; s3, constructing a distributed estimator; s4, deriving one-step prediction # imgabs1 of the i-th node at the s-th moment based on # imgabs0 #; S5, calculating an upper bound xi, s + 1s of one-step prediction error covariance of the i-th node at the s-th moment; s6, deducing a distributed estimator parameter Ki, s + 1 of the ith node at the (s + 1) th moment; s7, deducing state estimation # imgabs2 # S8, solving an estimation error covariance upper bound xi, s + 1s + 1 of the ith node at the (s + 1) th moment; and letting k = k + 1, and returning to S3. According to the method, the problem that the existing state estimation method cannot process complex network state estimation of a low-reliability communication environment and periodic scheduling at the same time is solved, so that the accuracy of the performance of the state estimation algorithm of the problem and the data transmission efficiency are improved.

This invention discloses a one-step snap-fit ​​prediction method for flexible flat cable based on virtual-real fusion and adaptive gradient. First, in the real operating environment of mobile phone flexible flat cable assembly, real tactile images are acquired using a tactile sensor. Then, based on the real tactile images, the Material Point Method (MPM) is used to simulate the real tactile sensation in a virtual environment, generating a virtual tactile image. Real and virtual tactile sensations together constitute a digital twin environment. By performing adaptive gradient calculation on the virtual and real tactile images in the digital twin environment, a one-step prediction model based on adaptive gradients is established. The model uses a CNN+LSTM network to predict the tactile images, generating predicted tactile images, which are then compared with ideal tactile images. Then, by encoding and decoding the tactile images, a one-step snap-fit ​​guidance strategy is obtained and applied in the real operating environment. The one-step prediction model based on adaptive gradients is trained and iterated in the digital twin environment.

The invention relates to the technical field of life prediction of electronic components, in particular to an IGBT life prediction method based on Gaussian process regression and an LSTM neural network. Collecting and emitting peak voltage data of an IGBT device are collected, preprocessed and then decomposed into a plurality of IMFs reflecting measurement noise and a residual error reflecting the overall degradation trend; using an LSTM model to train the residual error, and fitting a residual error sequence; training each IMF by using a GPR model, and fitting an IMFs sequence; using the trained LSTM model to predict a residual value; predicting the value of each IMF by using the trained GPR type; synthesizing the predicted residual value and the value of the IMF to obtain a final predicted value of the set emission peak voltage and uncertainty quantization of the predicted value of the set emission peak voltage; and the predicted collector emission peak voltage is used as the input of the next step of prediction, and the prediction is continued until the decommissioning point of the IGBT is reached. According to the method, the advantages of the LSTM neural network and the GPR model are combined, so that excellent voltage degradation evaluation performance is obtained.

The invention discloses a gate adaptive radar data association method and device. The method comprises the following steps: acquiring one-step predicted position information of a plurality of targets and echo information of a plurality of sensors; generating a first position Euclidean distance feature tensor between the targets and a second position Euclidean distance feature tensor between the targets and echoes according to the one-step prediction position information and the echo information; determining a gate threshold parameter of the target based on the first position Euclidean distance feature tensor and a position Mahalanobis distance feature tensor between the target and the echo; determining an association result of each target in each sensor according to the position mahalanobis distance feature tensor between the target and the echo, the second position Euclidean distance feature tensor and a wave gate threshold parameter; according to the method, the environment of the current tracking scene can be better perceived under the condition of not depending on prior information through self-adaptive adjustment of the gate threshold parameter, the calculation burden is relieved for subsequent data association, and meanwhile, the calculation difficulty is reduced.

The application discloses a kind of adaptive robust Kalman filtering integrated navigation methods based on MVC, mainly include: obtaining sensor real-time data;Carry out Kalman filtering state step prediction;Adaptive robust Kalman filtering estimation parameter matrix and adaptive factor are utilized;Carry out Kalman filtering measurement update;Output integrated navigation result.The application solves the problem that when system process noise is uncertain and system measurement noise is abnormal, the precision of integrated navigation result decreases or even diverges.

The invention discloses a technical acceptance and behavior intention prediction method based on a behavior intention derivative model. The prediction method comprises the following steps: S1, data acquisition: providing necessary input data for behavior intention prediction; s2, multiple interaction modes: providing more-dimensional information support for subsequent behavior intention prediction; s3, a behavior intention analysis model: constructing a behavior intention calculation formula through quantitative analysis of a plurality of influence factors, and providing accurate prediction of the user behavior intention; s4, construction of a prediction model: further predicting future behaviors of the user through analysis and training of user behavior intentions, and providing a basis for decision making; and S5, continuous iteration and updating of the model: performing real-time updating and optimization according to new data by adopting a continuous learning mechanism, and constructing an optimization prediction algorithm formula. The method is high in prediction accuracy, has extremely high applicability, can accurately grasp user requirements, and provides valuable decision support for technical design and popularization strategies.

The invention discloses a differential privacy fusion filtering method of an asynchronous sampling positioning system in a low-reliability environment. The method comprises the following steps: 1, establishing a dynamic model of an asynchronous sampling nonlinear positioning system based on sensor network communication; 2, converting the asynchronous sampling nonlinear positioning system dynamic model into a synchronous sampling nonlinear positioning system dynamic model through a zero-order holding strategy; 3, designing a differential privacy distributed fusion filter; 4, calculating one-step prediction and one-step prediction error covariance; 5, deriving a local filter gain according to the one-step prediction error covariance; 6, substituting the local filter gain and the one-step prediction into the step 3 to obtain fusion filtering; and 7, calculating a local filtering error covariance upper bound. According to the method, the problem that the privacy protection and filtering performance of the positioning system cannot be considered cooperatively in the existing filtering method is solved, and meanwhile, the influence of randomly generated nonlinearity of the sensor on the fusion filtering effect is discussed, so that the estimation precision of the problem is improved.

The present invention discloses a multi-sensor tight integration navigation method for underwater robots based on anti-error filtering. First, a SINS-DVL-USBL tight integration model is constructed to obtain the state equation and measurement equation of the tight integration model. Second, the state update is filtered, and a one-step prediction vector and covariance matrix are calculated based on the state equation to provide a covariance matrix for the measurement update. Then, the measurement noise is estimated based on the Mahalanobis distance to provide a measurement noise matrix for the measurement update. Finally, the measurement update is filtered, and the SINS information is corrected using the DVL and USBL information to obtain the navigation result. The present invention can effectively eliminate outliers and improve the accuracy of robot navigation while reducing the amount of computation. It can also obtain high-precision position information of the underwater robot in complex environments, further improving the operating efficiency of the underwater robot.

The present application belongs to the technical field of ship working condition prediction, and discloses a rotating speed prediction method for a ship gas-electric hybrid power system. A ship gas-electric hybrid power propeller rotating speed prediction model is constructed based on an adaptive neuro-fuzzy inference system (ANFIS); a difference between the obtained predicted rotating speed and the actual rotating speed is obtained; an improved rotating speed prediction model is constructed using the obtained difference and the used rotating speed information; and future rotating speed prediction is performed through the constructed improved rotating speed prediction model. The present application uses the initially constructed rotating speed prediction model to obtain the first-step predicted rotating speed, and provides the difference between the predicted rotating speed and the actual rotating speed for the improved prediction model; the improved rotating speed prediction model is constructed according to the obtained difference and the initial rotating speed information, so as to improve the rotating speed prediction accuracy and achieve the real-time prediction effect within a given time step.

The application discloses a kind of laser radar synthetic wind speed and wind direction processing method based on Kalman filtering, through the data detected by laser radar is calculated by Kalman filtering algorithm, specifically for the weight of sample point and sample point set one-step prediction, system state weighted mean and covariance matrix, measurement point set one-step prediction, measurement system weighted mean and covariance matrix, Kalman filtering gain and update system state value, update covariance value calculation, with prior state input and prior covariance matrix selects sample, obtains new system state weighted mean and covariance as the parameter of Kalman filtering by the nonlinear transformation of inversion algorithm, under the premise that the mean of wind speed and wind direction inversion error at hub is zero and unchanged, reduce the standard deviation of inversion error.

The invention discloses a nonlinear pipeline system state estimation method and system based on unscented Kalman filtering and a storage medium, relates to the field of network control, and aims to solve the problems of limited state estimation precision and high power consumption caused by signal attenuation, nonlinear characteristics and limited communication resources in the existing method. Comprising the following steps: 1, establishing an oil pipeline system dynamic model, and discretizing to obtain a state space model; 2, designing a dynamic event triggering mechanism, and constructing an amplification forwarding relay system to obtain amplification measurement signal data; step 3, setting initial values of state estimation and a covariance matrix; step 4, weighting 2n + 1 Sigma sampling points obtained by unscented transformation according to a weight coefficient to obtain one-step prediction and a one-step prediction error covariance Pj + 1j; 5, designing a filter according to a system phenomenon, calculating an estimation error, calculating an estimation error covariance upper bound matrix by using a matrix inequality, and solving a filtering gain matrix Kj + 1; and step 6, executing the step 4 and the step 5 until the total duration is reached.

The application discloses a kind of optimization distributed filtering method of target tracking system under work cycle scheduling strategy, the method is as follows: one, the dynamic model of target tracking system with state constraint and probability quantization is established;Two, distributed filter design is carried out to dynamic model under work cycle scheduling;Three, the upper bound of one-step prediction error covariance matrix is calculated;Four, filter gain matrix is calculated;Five, it is obtained in distributed filter by substituting, whether the total length of sensor network is reached, if, then execute six, if, then end operation;Six, the upper bound of filtering error covariance matrix is calculated;Let, execute two, until meet. The application solves the problem that existing distributed filtering method cannot simultaneously process sensor network with state constraint, probability quantization and work cycle scheduling, leading to the problem of reduced filtering performance.

The invention provides a double-view feature matching method based on a paired motion field, and the method comprises the following design steps: 1, for a given image feature matching image pair, employing an SIFT algorithm to extract feature points; 2, designing a pairwise motion vector field construction module; 3, designing a feature fusion module; and 4, further predicting the probability of each matching pair serving as a real matching point (namely an inner point) by using the initial matching pair processed by the constructed paired motion vector field and the feature fusion module. And 5, taking the obtained probability set and the corresponding set as input, and estimating an essential matrix by using a weighted eight-point algorithm. And F, iteratively executing the step C to the step E for five times, calculating cross entropy loss according to a predicted classification result and a real category result in each iteration, calculating regression loss in combination with a predicted essential matrix and a real essential matrix to guide network training, and finally obtaining a double-view feature matching model with optimal performance.

The application discloses an optimization method, system and device for predicting uplink throughput based on RSRP and a medium. The method comprises the following steps: constructing a data set comprising RSRP values and corresponding uplink throughputs; constructing an RSRP and uplink throughput mapping relationship to obtain a mapping model; performing multi-step prediction on the time sequence of RSRP based on an improved LSTM model, wherein the improved LSTM model comprises an encoder, an attention mechanism layer, a decoder and a full connection layer; the attention mechanism layer is used to calculate the correlation weight between the current time hidden state and all hidden states of the encoder at each step of the decoder prediction, and generate a weighted sum as the input of the decoder at the current time according to the correlation weight; and the value of the future time slot RSRP is predicted by using the improved LSTM model, and the uplink throughput is predicted in combination with the mapping model. The application predicts the uplink throughput by using the 5G network parameter RSRP based on the deep learning technology, so that the resources can be allocated in advance to meet the demand of the current system on the communication performance.

The application discloses a coordinate system transformation fusion filtering tracking method and system for a dual-base station radar, and relates to the technical field of radar target tracking. The method comprises the following steps: firstly, establishing state and observation equations, and determining an initial state of a target based on prior information in a Cartesian coordinate system. Then, a target motion state is obtained through one-step prediction, sigma points are generated by using U transformation, and a mean value and a covariance of state space prediction are calculated and constructed. Next, measurement data and prediction data are fused by using a Kalman filter to obtain optimal state estimation and a covariance matrix. Finally, the updated state estimation is converted back to the Cartesian coordinate system based on U transformation, and the process is repeated until the tracking is completed. The application can avoid the nonlinear filtering problem in the updating process, and improves the robustness and precision of tracking.

This invention discloses a multi-sensor information fusion target tracking method with a stochastic variable parameter matrix. The method includes the following steps: 1. Establishing a dynamic model of the tracking target in a multi-sensor target tracking system; 2. Designing the predictor and estimator structures; 3. Calculating the upper bound Θ of the one-step prediction error covariance matrix of the i-th sensor at time k+1. i,k+1|k 4. Calculate the estimated gain matrix K of the i-th sensor at time k+1. i,k+1 And the fusion estimation of the tracking target, part five, K i,k+1 Substituting into step two, we obtain the state estimate of the target tracked by the i-th sensor at time k+1. We then determine whether k+1 has reached the estimated total duration MN. If k+1 < MN, proceed to step six; if k+1 = MN, the process ends after calculating the fusion estimate. Step six: Calculate the upper bound of the estimation error covariance Θ. i,k+1|k+1 Let k = k + 1, and continue in step 2 until k + 1 = MN is satisfied. This invention can effectively estimate the target state and has good robustness.