20 results about "Weighted least squares algorithm" patented technology

The present disclosure discloses a seismic inversion method based on joint constraint of a physical model and priori information. The method includes: extracting seismic wavelets based on seismic data, and determining an amplitude scaling factor of the wavelets; counting priori information of impedance parameters; establishing an initial impedance parameter model by using seismic structural interpretation information and logging data; obtaining a simplified approximate equation based on an interface weak elasticity difference hypothesis, forward modeling a seismic gather by using the simplified equation, and calculating an inversion residual; rewriting an objective function into a function related to the impedance parameters by using a generalized linear inversion idea, solving the impedance parameters by using an iterative reweighted least squares algorithm, and updating the impedance parameters; and repeating the above steps until the inversion residual reaches the requirements or reaches the maximum number of iterations, and outputting a final processing result.

The invention relates to the technical field of railway signal enhancement, and discloses a railway signal enhancer ad-hoc network management method and system based on Beidou fusion positioning, and the method comprises the steps: a signal enhancer receives a Beidou satellite signal, calculates the initial coordinate of the signal enhancer, monitors a communication signal from a ground control base station, and transmits the communication signal to the ground control base station; when the communication signal level is lower than a first preset threshold value or the data packet loss rate is higher than a second preset threshold value, a differential relay mode is triggered, one signal enhancer is selected from a plurality of signal enhancers to serve as a temporary main reference, and the differential quality factor of each device serves as an input weight through a weighted least square algorithm; and the signal intensifier which is not elected as the temporary main reference receives the differential correction data sent by the temporary main reference and corrects the initial coordinate of the signal intensifier by using the data, so that the intelligence, the reliability and the stability of system operation and maintenance in a complex railway environment are realized.

This disclosure provides a method, apparatus, device, and storage medium for three-dimensional model reconstruction, relating to the field of artificial intelligence technology. In some embodiments of this disclosure, a measured depth map and a normalized depth map corresponding to a three-dimensional scene are obtained; the measured depth map and the normalized depth map are aligned to obtain a set of effective pixels; based on the set of effective pixels, a target scale factor is determined according to an iterative weighted least squares algorithm and a robust loss function; the coordinates of three-dimensional points in the normalized three-dimensional model are reconstructed according to the target scale factor to obtain a three-dimensional model corresponding to the three-dimensional scene. This disclosure, by aligning the measured depth map and the normalized depth map and combining the target scale factor, achieves automatic mapping from the normalized three-dimensional model to the real physical scale, improving the accuracy of three-dimensional model reconstruction.

The invention discloses a real-time monitoring method for over-pouring and under-pouring of track base concrete based on multi-sensor fusion, and particularly relates to the technical field of quality monitoring of track traffic engineering. The method comprises the following steps: collecting multi-point elevation data of a concrete pouring surface, performing spatial correction on a distance measurement result in combination with a pitch attitude angle, and generating a pouring thickness value matrix under a unified coordinate system; based on the thickness data, constructing a section fitting elevation curve by adopting a weighted least square algorithm, calculating a deviation value between each measuring point and a design reference surface, and realizing real-time judgment and three-color light visual prompt of over-under pouring through a set tolerance threshold value; furthermore, a micro-area anomaly recognition model is constructed according to the deviation change rate of the continuous section, and the potential pouring tendency anomaly can be early warned in advance; according to the method, a real-time decision basis can be provided for template elevation adjustment and pouring rate control, and high-precision, high-time-efficiency and intelligent quality management and control of track base concrete pouring are achieved.

The invention discloses an unmanned aerial vehicle positioning method and an unmanned aerial vehicle positioning device, which are used for solving the technical problem that the existing unmanned aerial vehicle positioning technology does not consider the deviation problem of a constraint weighted least square problem in model conversion, the deviation in positioning is difficult to reduce, and the positioning precision is relatively low. The method comprises the following steps: acquiring a reference position and a reference speed of an unmanned aerial vehicle, and calculating a measurement frequency difference value and a measurement distance difference value according to the reference position and the reference speed of the unmanned aerial vehicle; according to the reference position and the reference speed of the unmanned aerial vehicle, the measurement frequency difference value and the measurement distance difference value, constructing a constraint weighted least square algorithm model; solving the constraint weighted least square algorithm model, and determining the initial speed and the initial position of the unmanned aerial vehicle; based on the measurement frequency difference value, the measurement distance difference value and the initial speed and the initial position of the unmanned aerial vehicle, constructing an unmanned aerial vehicle positioning time-frequency domain error compensation model; and solving the unmanned aerial vehicle positioning time-frequency domain error compensation model, and outputting the target position and the target speed of the unmanned aerial vehicle.

The invention discloses a joint error dynamic correction method based on an optical fiber and wireless sensing co-judgment event, and relates to the technical field of distributed optical fiber sensing, and the method comprises the steps: associating a DAS vibration event with a wireless network acoustic event through high-precision time synchronization, generating a co-judgment event instance, defining a to-be-estimated error parameter, and building a joint observation model; and performing real-time dynamic correction on the event by combining an iterative reweighted least square algorithm with the error parameters and utilizing final error parameter estimation, and obtaining a result through weighted averaging. According to the method, joint modeling and real-time compensation are carried out on the positioning error of the DAS host and the wireless sensor network, so that the fusion positioning result is closer to the real geographic coordinate of the disturbance event, the limitation of a traditional method is overcome, the positioning precision in a complex environment is improved, and the positioning accuracy is improved. Moreover, an iterative reweighted least square algorithm and a forgetting factor mechanism are also adopted, so that the robustness to abnormal events and environmental changes is enhanced, and long-term stable adaptive optimization is realized.

The invention provides a multi-frequency observation data restoration method and device for a Beidou differential router. The method comprises the following steps: acquiring original multi-frequency observation data through the Beidou differential router; constructing an original observation equation according to the original multi-frequency observation data; constructing a single-satellite multi-frequency epoch difference model according to the original observation equation; resolving the single-satellite multi-frequency epoch difference model by adopting a weighted least square algorithm to obtain a resolving result; determining the frequency of missing data in the original multi-frequency observation data and a corresponding missing moment; obtaining observation data of a frequency corresponding to a previous epoch at the missing moment; and according to the resolving result and the observation data of the corresponding frequency of the previous epoch at the missing moment, repairing the missing data at the corresponding frequency at the missing moment. According to the method, the utilization rate of observation data can be effectively improved, positioning precision reduction and ambiguity fixation difficulty caused by effective observation reduction are avoided, and the positioning precision is further improved.

A seismic inversion method under joint constraints of a physical model and prior information, comprising: extracting seismic wavelets on the basis of seismic data, and determining a wavelet amplitude scaling factor; collecting statistics about priori information of impedance parameters; using seismic structure interpretation data and well logging data to establish an initial impedance parameter model; obtaining a linear approximate equation on the basis of an interface weak elasticity difference hypothesis, and using the linear approximate equation to forward simulate a seismic gather and calculate an inversion residual error; using a generalized linear inversion idea to rewrite a target function into a function related to the impedance parameters, using an iterative reweighted least square algorithm to obtain the impedance parameters, and updating the impedance parameters; and repeating the steps above until the inversion residual error reaches the requirement or reaches the maximum number of iterations, and outputting a final processing result. The inversion method can achieve stable and high-precision prediction of impedance parameters.

The invention provides a moving target parameter estimation method based on networking radar space-time measurement. The method comprises the following implementation steps: establishing space-time measurement of networking radar moving target parameter estimation; estimating the positions of the target at all observation moments; estimating the initial speed and acceleration of the target based on a weighted least square algorithm; and obtaining a moving target parameter estimation result based on a Gaussian Newton iteration method. According to the method, time sequence observation is performed on the uniformly accelerated moving target through the networking radar to obtain the time-space measurement information of the target, and the moving target parameters including the initial position, the initial speed and the accelerated speed of the target are estimated, so that the observation dimension and the information utilization rate are effectively increased; the defect that in the prior art, only measurement data under single observation is utilized, but time sequence distance and radial speed measurement data obtained by a radar in continuous observation are not fully utilized is overcome, the Cramer-Rao lower bound of estimation errors is remarkably reduced, and parameter estimation precision and comprehensiveness are improved.

The invention relates to the technical field of robots, in particular to a robot navigation optimization method based on iterative reweighted least squares. The method comprises the following steps: calculating a pre-integration result for an onboard inertial measurement unit, and recurring a physical track; calculating the geometric difference between the cross-node measurement closed loop and the physical track, and generating a cross-node track residual error; calculating an interaction trust value between the nodes in combination with a historical trust sliding window, and constructing a node consistency network; generating environment confidence for all nodes based on navigation sensor data; solving a global optimal navigation constraint weight set through iterative calculation, and generating navigation constraint weights in combination with the environment confidence; and carrying out optimization solution on the space-time pose image according to the navigation constraint weight by adopting an iterative reweighted least square algorithm to obtain a navigation trajectory solution. According to the method, the influence of the pseudo-signal nodes can be effectively inhibited, and the global consistency and reliability of cluster navigation solution are improved, so that a high-precision navigation trajectory solution with global geometric consistency is generated.

The invention discloses an error calibration method of a magnetometer and a magnetometer calibration method. The error calibration method of the magnetometer comprises the following steps: obtaining a data set; preprocessing the original data of the magnetometer in the data set to obtain magnetic field data; constructing a total error model of the magnetometer; based on a weighted least square algorithm, according to the attitude angle in the data set and the noise variance of the magnetic field data, constructing a diagonal weight matrix inversely proportional to the noise variance; and performing linearization processing on the total error model of the magnetometer, combining the diagonal weight matrix, taking a minimum weighted residual sum of squares as a target function, iteratively updating parameters to be estimated of the total error model of the magnetometer by using the data set, and repeating the iteration process until a preset convergence judgment condition is met, so as to obtain the total error model of the magnetometer. Obtaining calibration parameters of the magnetometer total error model; by adopting the technical scheme, the magnetic field measurement error is effectively reduced, the calibration parameter can be directly applied to the real-time error compensation of the magnetometer in the working stage, and the measurement precision is ensured.

The present application relates to the technical field of railway signal enhancement, and discloses a railway signal enhancer ad hoc network management method and system based on Beidou integrated positioning, which comprises the following steps: receiving Beidou satellite signals by a signal enhancer, solving the initial coordinates of the signal enhancer, monitoring communication signals from a ground control base station by the signal enhancer, triggering a differential relay mode when the communication signal level is lower than a first preset threshold or the data packet loss rate is higher than a second preset threshold, selecting one from a plurality of signal enhancers as a temporary master reference, the selection is carried out by a weighted least squares algorithm with the differential quality factor of each device as the input weight, receiving the differential correction data sent by the temporary master reference by the signal enhancer which is not selected as the temporary master reference, and correcting the initial coordinates of the signal enhancer by using the data, thereby realizing the intelligence, reliability and stability of system operation and maintenance in complex railway environment.

The invention relates to a tunnel positioning method and device based on Beidou and visual target fusion and a medium, and belongs to the technical field of high-precision positioning and navigation.According to the tunnel positioning method based on Beidou and visual target fusion, a unified tunnel coordinate system is established, and high-precision visual targets are arranged in a tunnel in the driving direction; an initial positioning result of a vehicle in a tunnel under a tunnel coordinate system is determined by identifying a target number and a pixel coordinate in the tunnel, a vehicle initial state vector of the vehicle under the tunnel coordinate system is determined according to the initial positioning result and a reference coordinate, IMU prediction and visual target prediction are carried out respectively, an IMU observation value and a visual observation value are obtained, and the vehicle initial state vector of the vehicle under the tunnel coordinate system is determined. IMU dead reckoning and visual observation are fused based on a weighted least square algorithm, centimeter-level continuous positioning is achieved, and continuous and high-precision positioning of the vehicle in the tunnel environment is achieved.

The invention relates to an iterative reweighted least square-based robust noise covariance estimation method, which comprises the following steps of: 1, constructing a state space model of a linear time-invariant system, collecting measurement data in a system operation process to calculate an innovation sequence, and estimating an observation vector formed by auto-covariance of the system; step 2, on the basis of the observation vector, constructing a linear regression equation set of an innovation covariance and a noise covariance by using a system state space model and a steady-state Kalman gain; step 3, acquiring a robust noise covariance estimated value for the linear regression equation set by adopting an iterative reweighted least square algorithm; and step 4, feeding back the robust noise covariance estimation value to the Kalman filter, carrying out Kalman filter parameter updating and circulation, and realizing on-line adaptive adjustment of noise statistical parameters. According to the method, the robustness of abnormal values is remarkably improved in noise statistical parameter estimation, and the accuracy and stability of Kalman filtering can still be ensured in an environment containing noise and abnormal data.

Aspects of the present disclosure include methods for spectrally resolving light from fluorophores with overlapping fluorescence spectra in a sample using adaptive measurement variance. The method, according to certain embodiments, includes detecting light from particles in the sample containing multiple fluorophores with overlapping fluorescence spectra using a light detection system, determining the measurement variance of the detected light for each particle, and spectrally resolving the light from each fluorophore in the sample using a weighted least-squares algorithm that uses the measurement variance determined for each particle. In some embodiments, the spectral unmixing matrix is ​​calculated by the weighted least-squares algorithm using an adaptive measurement variance model. Systems and integrated circuit devices (e.g., field-programmable gate arrays) for implementing the subject methods are also described. Non-transitory computer-readable storage media are also provided.

The application discloses a multi-vehicle positioning method and system based on vehicle networking communication and cooperative sensing, which comprises the following steps: using a target detection algorithm combining motion compensation and deep learning to correct and calculate the original point cloud data output by the laser radar installed on the vehicle to be positioned, so as to obtain the accurate relative distance between the vehicle to be positioned and each selected anchor vehicle within the predetermined communication range; the vehicle to be positioned and the anchor vehicle are networked vehicles or intelligent networked vehicles; based on the accurate relative distance between the vehicle to be positioned and each selected anchor vehicle, combined with the vehicle center position broadcast by each selected anchor vehicle, a nonlinear distance observation equation is constructed; the nonlinear distance observation equation is solved by using an iterative linearization weighted least square algorithm, so as to obtain the position increment of the vehicle to be positioned which changes in real time, and then the global absolute coordinate of the vehicle to be positioned is calculated. The application can improve the positioning accuracy, robustness and real-time performance in a complex traffic environment.

A method for calibrating and determining the error of a magnetometer is disclosed. The error calibration method includes the following steps: obtaining a dataset; preprocessing the raw magnetometer data in the dataset to obtain magnetic field data; constructing a total error model of the magnetometer; based on a weighted least squares algorithm, constructing a diagonal weight matrix inversely proportional to the noise variance of the attitude angle and magnetic field data in the dataset; linearizing the total error model of the magnetometer, and using the diagonal weight matrix, iteratively updating the parameters to be estimated in the total error model of the magnetometer with the objective function of minimizing the sum of squared weighted residuals, repeating the iterative process until a preset convergence criterion is met, to obtain the calibration parameters of the total error model of the magnetometer; the above technical solution effectively reduces the magnetic field measurement error, and the calibration parameters can be directly applied to real-time error compensation during the working stage of the magnetometer, ensuring measurement accuracy.

The application relates to a robust noise covariance estimation method based on an iterative reweighted least square, which comprises the following steps: step 1, constructing a state space model of a linear time-invariant system, collecting measurement data in the system operation process, calculating an innovation sequence, and estimating an observation vector composed of self-covariance; step 2, based on the observation vector, using the state space model of the system and a steady-state Kalman gain, constructing a linear regression equation group of the innovation covariance and the noise covariance; step 3, using an iterative reweighted least square algorithm, obtaining a robust noise covariance estimation value from the linear regression equation group; and step 4, feeding back the robust noise covariance estimation value to a Kalman filter, performing Kalman filter parameter updating and circulation, and realizing online adaptive adjustment of noise statistical parameters. The application significantly improves the robustness to abnormal values in noise statistical parameter estimation, and can guarantee the accuracy and stability of Kalman filtering in an environment containing noise and abnormal data.

The invention provides a method for updating the depth of a channel and a storage medium, and the method comprises the steps: firstly carrying a depth finder on a ship, and collecting discrete new depth sounding points containing coordinates, a water depth value and a timestamp in real time; then generating a sounding surface by using a concave shell boundary extraction algorithm and a triangulated irregular network interpolation algorithm; uniformly converting the newly generated sounding surface and the old water depth surface into a point set form for projection registration to obtain a registration point pair corresponding to the space; the registration point pairs are processed through a Huber loss function and an iterative reweighted least square algorithm, and the real-time draught depth of the depth finder is obtained through inversion; and finally, superposing the draft obtained by inversion with the sounding surface, and outputting a new version of water depth surface. According to the method, traditional periodic off-line water depth survey is converted into on-ship instant on-line updating, the problem of water depth data lag caused by water level dynamic change and terrain deposition is effectively solved, and the current trend of data is remarkably improved.