66results about How to "Achieve precise estimates" patented technology

The invention relates to an SOC estimation method for a series-wound battery pack having an equalization circuit. According to the estimation method, adaptive extended Kalman filtering SOC estimation is carried out on the single batteries with the lowest voltage or the highest voltage in the charge-discharge stage so as to obtain the SOC of the series-wound battery pack. For eliminating influence on the measurement signal from noise, and for analyzing the unstable and sharply-changed voltage and current signals, wavelet transform is performed before the implementation of the adaptive extended Kalman filtering so as to realize denoising and analyzing of the signal; and the condition that the curves of the battery parameters and the open circuit voltage-state of charge (OCV-SOC) are changed along with the changes of temperatures and the changes of the SOC is taken into consideration. By adoption of the estimation method, online updating of the parameters can be performed, so that the SOC estimation precision can be improved; the SOC of the battery pack can be accurately estimated under the premise of ensuring the safety operation of the series-wound battery pack; therefore, the estimation method is applicable to active equalization and passive equalization; and meanwhile, by adoption of the estimation method, the influence from noise can be effectively eliminated, and the voltage and current signals can be effectively analyzed.

The invention relates to a low-altitude unmanned aerial vehicle vision positioning method based on a rapid robust feature. The low-altitude unmanned aerial vehicle vision positioning method based on the rapid robust feature can be suitable for rotation and size variation of aerial sequential images of a low-altitude unmanned aerial vehicle and noise interference and achieve precise estimation of positions of aircrafts. Firstly, SURF scale space is built, extreme points are positioned by using a rapid Hessian matrix, and 64-dimention SURF feature descriptors of an aviation image are calculated; then feature point matching is completed on the basis of a Hessian matrix track; and finally, outliers are removed by using a random sampling consensus (RANSAC) method so as to achieve precise estimation of position parameters. Transformation parameters between a reference image and a real-time image are obtained through the RANSAC method. After estimation of partial parameters on the basis of RANSAC is completed, and outer points are removed, inner points meeting matching requirements are figured out, and transformation results of estimation parameters of the real-time image based on the RANSAC and positioning results of the center of the real-time image on reference image can be obtained.

The invention relates to an aviation sequence image position estimating method based on SURF (Speeded Up Robust Features). The aviation sequence image position estimating method can adapt to rotation of an aviation sequence image, dimension change and noise interference, realize accurate estimation of an airplane. The aviation sequence image position estimating method comprises the following steps of: firstly, structuring SURF dimension space, locating an extreme point by using a quick Hessian matrix, calculating a 64-dimensional SURF describer of the aviation image; then finishing point match on the basis of a Hessian matrix track; at last, eliminating outliers by using a random sample consensus algorithm (RANSAC) method, realizing accurate estimation of a position parameter. The invention aims at a severe working environment of an aircraft and aims to improve precision of visual navigation; compared with a traditional visual navigation method, the method can effectively overcome the shortcoming of nonsensitive to dimension, rotation, light and other factors in the process of position estimation by an unmanned plane, and obviously improve the matching precision of a real-time image and a reference image; and therefore, the method has an important actual meaning for a visual navigation project application of the unmanned plane.

The invention belongs to the GNSS data processing and ionospheric modeling technology field and discloses an LEO satellite-enhanced GNSS ionospheric normalization and fusion modeling method. The method comprises the following steps of extracting and obtaining an ionospheric total electron content observation value containing pseudo-range / phase hardware delay deviation based on original observation data of a GNSS satellite and an LEO satellite; determining various hardware delay deviation parameters and obtaining a "clean" ionospheric TEC observation value; normalizing the TEC observation value of the LEO satellite and converting the LEO satellite TEC observation value which can only detect partial paths into a full path; and realizing ionospheric unknown parameter estimation of the deviation with an LEO satellite TEC observation value system to construct a high-precision and high-resolution ionospheric model. In the invention, a problem of GNSS / LEO multi-source ionospheric TEC information normalization and time varying systematic deviation estimation is solved, and construction of a high-precision and high-resolution global ionospheric model is realized.

The invention discloses a phase test device and a control method thereof, and relates to the technical field of digital signal processing. The phase test device comprises a conditioning circuit, an A / D converter, a digital signal processor, an output driver and a display circuit of the output driver, wherein an analog signal is accessed to the conditioning circuit, the conditioning circuit adjusts the amplitude of the analog signal according to a first threshold to obtain a first analog signal, the first analog signal is sampled by the A / D converter to obtain a sample sequence, the sample sequence is accessed into the digital signal processor through a parallel digital input manner, an estimated phase value of the sample sequence is obtained through processing of the digital signal processor, and the estimated phase value of the sample sequence is output and displayed by the output driver and the display circuit of the output driver. The phase test device can realize accurate estimation of instantaneous phase of the center sample point by using shorter sample data; the phase test device can realize phase estimation without frequency estimation; without frequency estimation, error of frequency estimation is avoided, the cost of hardware and the requirement on complexity are reduced, and the requirements in actual application are met.

The invention discloses a frequency domain channel estimation method and system. The method includes that: S1, a receiving terminal carries out FFT conversion to convert the received data into frequency domain, and frequency domain data of pilot frequency of each antenna is extracted at corresponding carrier position of pilot frequency in frequency domain signal received by each antenna; S2, the frequency domain data of pilot frequency of each antenna is matched with local pilot frequency of any user to be processed; S3, the matched result is converted to time domain by IFFT conversion, and multi-user separation is carried out in time domain to obtain related information of time domain pilot frequency channel response of different user channels; S4, FFT conversion is carried out on the related information, the related information is converted to frequency domain, so as to obtain frequency domain pilot frequency channel estimation of different user channels corresponding to each receiving antenna. The invention can effectively separate each user when multiple user pilot frequency sequences occupy the same video resource, thus realizing accurate estimation of each user channel response.

The invention belongs to the technical field of multi-hop networks, and discloses a distance-corrected chaotic particle swarm multi-dimensional scaling wireless sensor network positioning method. During a diversity matrix building stage, all node pairs adopt recursive measurement and RSSI to perform weighted correction on distance, rather than utilization of 1 to replace the distance between the node pairs when the distance between the nodes cannot be measured, a node hole problem is avoided, and a distance estimation error is reduced. During a coordinate conversion stage, a chaotic particle swarm optimization algorithm is used for optimizing a conversion parameter, the problems of local optimal solution of the particle swarm and misconvergence are avoided, influence caused by the parameter is reduced, and node positioning accuracy is effectively improved. A simulation result shows that the distance-corrected chaotic particle swarm algorithm can significantly improve node positioning accuracy, and has better robustness and adaptation to irregular networks.

The invention discloses a state prediction and estimation method for an unmanned vehicle, and the method comprises the steps: employing a provided vehicle torque correction speed prediction formula, determining the optimal parameters of the speed prediction formula through the characteristics of acceleration, acceleration derivative and the like, and carrying out the training through a genetic algorithm neural network, introducing historical data of a current driving condition of a vehicle to update and optimize acceleration and an acceleration derivative in time, fusing a kinematics method and a dynamics method to estimate a road slope angle, using a least square method to estimate vehicle mass, and improving estimation precision of the vehicle mass and sprung mass. Estimating the vertical force of the wheels by using an information fusion method, introducing a road adhesion coefficient, and performing joint estimation on the lateral force and the state of the vehicle by using a nonlinear estimation method; finally, a vehicle state rolling estimation method in which vehicle variable parameters and variable working conditions are introduced is provided, and the precision of vehicle speed prediction under the conditions of vehicle variable parameters and variable working conditions is improved.

The invention discloses a frequency agility radar target information reconstruction method based on coherent accumulation. The method comprises the following steps: establishing a target echo signal model of the frequency agile radar based on coherent accumulation, and an optimization model taking a phase error and scene sparseness as optimization targets, and solving the optimization model through phase self-focusing algorithm reconstruction and an alternating direction multiplier method so as to obtain self-adaptive phase difference and speed and distance information of a target, and achieving reconstruction of the target information. According to the method, accurate target echo phase error estimation and observation scene reconstruction can be carried out, the problem that the coherent accumulation performance of the coherent frequency agility radar is reduced due to the phase error is effectively solved, good coherent accumulation performance is provided, and accurate estimation of the distance and the speed of the target is achieved.

The invention relates to a carrier frequency offset estimation method based on a convolution neural network, belonging to the technical field of signal processing, which solves the problem that the frequency offset in the prior art cannot be estimated accurately and is difficult to adapt to all digital communication signal modulation patterns. A carrier frequency offset estimation method based onconvolution neural network includes the following steps: acquiring signal sampling data; constructing a convolution neural network based on the signal sampling data; training the constructed convolution neural network to obtain a convolution neural network model based on carrier frequency offset; the signal carrier frequency offset is estimated using the convolution neural network model based on the carrier frequency offset. Accurate estimation of carrier frequency offset is realized, and it can be adapted to different modulation signals.

The invention discloses a visual target tracking method based on deep residual network characteristics. The visual target tracking method comprises the following steps: 1, selecting a characteristic layer of a deep residual network and calculating a weight; 2, extracting features of the first frame of actual input image; 3, constructing a response and initial position filter of the characteristicsof the first frame of actual input image; 4, performing scale sampling and fHOG feature extraction on the first frame of actual input image; 5, constructing an initial scale filter; 6, feature extraction of the second frame of actual input image; 7, position filtering; 8, weighting a position filtering response graph and positioning a target; 9, performing scale sampling and fHOG feature extraction on the target image; 10, performing scale filtering and scale estimation on the target feature vector; 11, updating the filter; And 12, inputting a next frame of actual input image, regarding the next frame of actual input image as a second frame of actual input image, and repeating the step 6. The method is high in tracking precision and success rate, adapts to target scale changes, and achieves the robust tracking of the target.

The present invention discloses a fractional order KiBaM model parameter identification method and system of a power battery, wherein the identification method comprises the steps of 1 initializing a linear parameter vector, utilizing an iterative learning identification method to identify the fractional order of a fractional order KiBaM model, and obtaining a fractional order initial value to thereby identify a non-linear parameter vector; 2 utilizing the identified non-linear parameter vector and combining the fractional order to identity the linear parameter vector and further determine the fractional order KiBaM model of the power battery; 3 acquiring the actual output of the power battery, calculating and determining whether the norm of the deviation of the output of the determined fractional order KiBaM model of the power battery and the acquired actual output of the power battery reaches a minimum value, if yes, ending to identify, and obtaining an optimal parameter of the fractional order KiBaM model of the power battery; otherwise returning to the step 1.

The invention provides an atomic spin gyroscope / magnetometer tightly-integrated attitude determination method based on an ant colony PF (Particle Filter) algorithm and relates to an SERF (spin-exchange relaxation free) atomic spin gyroscope / magnetometer integrated attitude determination method. The method comprises the following steps of: firstly, acquiring inertial angular velocity information by utilizing the atomic spin gyroscope, then carrying out gyroscope error compensation, and carrying out an attitude algorithm to acquire attitude information of a carrier; secondly, acquiring terrestrial magnetism measurement information by utilizing an atom magnetometer, and carrying out terrestrial magnetism matching on the terrestrial magnetism measurement information, so as to acquire terrestrial magnetism vector information; finally, merging the terrestrial magnetism vector information and the attitude information of the carrier in a tightly-integrated manner by utilizing the ant colony particle filter algorithm to solve nonlinearity and noise non-Gaussian problems of a system, working out the attitude information of the high precision carrier, estimating the gyroscopic drift, carrying out feedback and correction on the attitude of the carrier, and compensating the gyroscopic drift; at last, realizing online correction on random gyroscope errors of an SERF-based atomic spin gyroscope / magnetometer tightly-integrated attitude determination system, and finishing the long-time and high-precision integrated attitude determination of a moving carrier.

The invention discloses a lithium ion energy storage battery SOC online estimation method, which comprises the steps of obtaining rated parameters of a to-be-detected lithium ion energy storage battery, and establishing an equivalent circuit model of the to-be-detected lithium ion energy storage battery, identifying model parameters of the established equivalent circuit model on line, establishingan improved adaptive extended Kalman filter of the lithium ion energy storage battery according to the established equivalent circuit model and model parameters obtained by online identification, andinputting the model parameters obtained by online identification into an improved adaptive extended Kalman filter of the lithium ion energy storage battery, and carrying out online estimation on theSOC of the lithium ion energy storage battery. Parameters of the first-order RC equivalent circuit model are identified in real time through a recursive least square method with a forgetting factor, and then the parameters are input into the improved adaptive extended Kalman filter, so that accurate estimation of the SOC of the lithium ion energy storage battery is completed.

The invention discloses a flight target altitude estimation method based on dual-frequency RCS information fusion. The method comprises steps: through fusing high-frequency ground wave radar dual-frequency signal echo intensity information, an estimation model in which the high-frequency ground wave radar target flight altitude and the RCS information can be totally observed and fused in a single base system is built. In a high-frequency ground wave radar dual-frequency working mode, low-altitude flight target RCS information at two working frequencies are combined and transferred, fusion of dual-frequency signal echo information is realized, observability of the altitude estimation model is improved, and the flight target altitude and the RCS are precisely estimated.

The invention relates to an InSAR interferometric phase unwrapping method based on a semi-parametric model. The method is used for acquiring an unwrapping phase from an interferometric phase in the InSAR or DInSAR data processing process. The method specifically includes the following steps that firstly, a semi-parametric phase unwrapping model is established; secondly, the semi-parametric phase unwrapping model is resolved, and the unwrapping phase is estimated. Compared with the prior art, the method takes system errors in the InSAR / DInSAR interferometric phase into consideration, non-parametric components adopting the semi-parametric model are considered, modeling and compensation are performed on the non-parametric components, and therefore influences of the method on evaluation of the unwrapping phase are eliminated, the defects of a traditional method based on a least-square phase unwrapping algorithm are made up for, influences of the system errors on a phase unwrapping result are reduced, and the precision of terrain estimation and deformation monitoring is improved.

The embodiment of the invention relates to a target state estimation method and device and an unmanned aerial vehicle, and the method comprises the steps: obtaining a target image collected by a camera device, and carrying out the image recognition of the target image, so as to obtain a target image region of a target in the target image; obtaining an original point cloud of the surrounding environment of the unmanned aerial vehicle, and obtaining a first target position according to the target image area and the original point cloud; acquiring attitude information of the camera device, acquiring a first image position based on the target image area, and acquiring a second target position according to the first image position and the attitude information; acquiring a second image positionbased on the target image area; initializing a target state according to the first target position or the second target position; and taking the first target position and / or the second target positionand / or the second image position as a measurement value of an extended Kalman filter to obtain a target state. According to the embodiment of the invention, strong hypothesis conditions are not needed, and the state estimation precision is high.

The invention discloses a surgical navigation system of mandibular angle osteotomy, which comprises the following steps: S1, establishing an osteotomy surface prediction model learning version based on a multitask convolutional neural network and according to the relevant data of past mandibular angle osteotomy patients, carrying out training by using the relevant data of a newly enrolled group ofmandibular angle osteotomy patients to obtain a stable osteotomy surface prediction model, and then superimposing data sets in a danger area to obtain an osteotomy surface prediction model stabilization version; S2, inputting the relevant information of mandibular angle osteotomy patients into an osteotomy surface prediction model, and predicting the change range of a post-operation face 3D effect; S3, drawing a projected three-dimensional image, which can be seen through and fit with a surgical field real time image, on the screen of a lens according to the maximum osteotomy amount and pre-operation CT; and S4, carrying out function superimposition on an osteotomy surface prediction model continuously, testing and improving continuously, and perfecting the surgical navigation system. Thesystem predicts a post-operation effect through establishing models, real-time rendering is performed during operation, and the operation accuracy is improved.

The invention discloses a retinal artery vein diameter ratio automatic measuring method and a device. The retinal artery vein diameter ratio automatic measuring device can automatically calculate thecorresponding AVR indexes, increase the accuracy and automation degree of AVR calculation, realize accurate measurement of the AVR indexes, and accelerate clinical disease diagnosis and treatment speed. The method comprises the following steps: (1) acquiring a retina image by using fundus imaging equipment; (2) carrying out optic disc positioning and blood vessel segmentation; (3) based on the center line, carrying out blood vessel topology analysis; and (4) based on a deep convolutional network, classifying the blood vessel tree into arteries or veins, calculating the diameter and category ofsampling point positions, and completing the measurement of the retinal artery-vein diameter ratio AVR.

The invention discloses a UAV (Unmanned Aerial Vehicle) positioning method based on multi-source direction finding information fusion, and the method comprises the following steps of acquiring direction finding information of the UAVs by the stations deployed at different detection locations, wherein the direction finding information of the UAVs comprises the signal strength, the signal interference noise ratio and the signal direction of arrival of the UAVs; when the measurement and control center receives the direction finding information of UAVs of at least 3 stations, initially estimatingpositions of the target UAVs according to the directions of arrival measured by any two stations and the positions of the stations; and performing linear weighted sum on the calculated positions of all the target UAVs by combining the signal interference noise ratio sent by the stations to obtain the final positions of the target UAVs. The invention provides a practical and rapid UAV positioning method, which fully utilizes the existing measurement information, takes into account the position information of different base stations and the signal interference noise ratio of the received signalto realize accurate estimation of the positions of the target UAVs, and has simple process and low complexity in the calculation of the positions of the target UAVs.

The invention discloses a quantum random number generator independent of measurement end devices. The quantum random number generator includes a credible random source and an incredible measurement end; the credible random source randomly works in a random number generation mode and a safety verifying mode; in the random number generation mode, the credible random source sends a fixed quantum states to the incredible measurement end, and the incredible measurement end performs measurement; in the safety verifying mode, the credible random source sends different quantum states to the incredible measurement end, real-time safety verifying is performed according to a measurement result output by the incredible measurement end, the minimum entropy of original random data is estimated according to the measurement result, and random numbers independent of the measurement end devices are generated. The scheme of the invention can effectively eliminate potential safety hazard due to incredibility and imperfection of measurement devices.

The invention discloses an unmanned aerial vehicle SAR self-focusing method based on aerial carrier track deviation estimation. According to the method, the aerial carrier track deviation is estimatedon the basis of echo data, and the accurate estimation of the aerial carrier track deviation is realized by utilizing a coarse-to-fine step-by-step self-focusing strategy, so that the two-dimensionalspace-variant property of motion errors is fundamentally solved, the low-speed unmanned aerial vehicle-mounted SAR self-focusing is realized, and a high-resolution unmanned aerial vehicle-mounted SARimage is obtained.

The invention discloses an MIMO-SAR three-dimensional imaging self-focusing method based on motion parameter estimation. The method comprises: selecting a feature region based on a coarse imaging result, and carrying out two-dimensional spectrum analysis on a local feature region; obtaining a local space-invariant two-dimensional slope distance error through contrast ratio optimal estimation; and solving translation and attitude errors of MIMO-SAR through a motion error transfer model and a least square method, so that three-dimensional imaging self-focusing is realized. By using the method, MIMO-SAR self-focusing three-dimensional imaging can be realized under the condition that navigation precision cannot meet the imaging requirement due to no high-precision navigation information or relatively high working frequency band, and a well-focused three-dimensional radar image is obtained.

The invention belongs to the GNSS data processing and ionospheric modeling technology field and discloses an LEO satellite-enhanced GNSS ionospheric normalization and fusion modeling method. The method comprises the following steps of extracting and obtaining an ionospheric total electron content observation value containing pseudo-range / phase hardware delay deviation based on original observation data of a GNSS satellite and an LEO satellite; determining various hardware delay deviation parameters and obtaining a "clean" ionospheric TEC observation value; normalizing the TEC observation value of the LEO satellite and converting the LEO satellite TEC observation value which can only detect partial paths into a full path; and realizing ionospheric unknown parameter estimation of the deviation with an LEO satellite TEC observation value system to construct a high-precision and high-resolution ionospheric model. In the invention, a problem of GNSS / LEO multi-source ionospheric TEC information normalization and time varying systematic deviation estimation is solved, and construction of a high-precision and high-resolution global ionospheric model is realized.

The embodiment of the invention provides a carrier frequency estimation method and device. The estimation method comprises the following steps of: obtaining a to-be-estimated signal received by the demodulation end, and inputting the to-be-estimated signal into a Duffing vibration subsystem in a critical chaotic state to carry out step size iterative calculation to obtain an output phase track, and if the phase track is in a periodic intermittent chaotic state, calculating the carrier frequency of the to-be-estimated signal according to the step size in the periodic intermittent chaotic state.In the embodiment of the invention, in non-cooperative digital communication, the complexity of a carrier frequency estimation algorithm is reduced, accurate estimation of the carrier frequency of the signal to be estimated is realized, the accuracy of signal demodulation is improved, and the communication quality is ensured.

The invention discloses an elevator fault diagnosis method. The method comprises the steps that real time data of elevator components are collected, and a theoretical temperature change curve is constructed, wherein the elevator components comprise a traction machine, a lift car portal crane, a speed governor, a buffer, a controller, a transformer and a frequency converter; actual temperature of the elevator components is collected, and real time temperature change curves are generated according to the real time temperature; the real time temperature change curves are compared with the theoretical temperature change curve, and abnormal real time temperature change curves are found; and faulted elevator components are extracted according to the abnormal real time temperature change curves. The invention further discloses an elevator fault diagnosis system. By means of the elevator fault diagnosis method and system, the theoretical temperature change curve can be constructed by collecting the real time data of the elevator components, and therefore a polling person can be better assisted in rapidly judging whether the elevator components are abnormal or not.

The invention discloses an indoor high-precision positioning method for an intelligent warehouse management system, and the method comprises the following steps: S1, constructing a sample set with a distance label, deploying an indoor position prediction model in the intelligent warehouse management system, training the indoor position prediction model through the sample set with the distance label, and obtaining a high-precision indoor position prediction model; and s2, according to a high-precision indoor position prediction model, calculating a positioning coordinate of a target point through the distance. According to the invention, the problem of inaccurate indoor positioning is solved.

The invention discloses a carrier frequency offset estimation method based on zero subcarrier phase fitting, which comprises the following steps of: S1, determining a sending interval according to a custom function to control a Wi-Fi signal sending end to send Wi-Fi frames, and obtaining CSI (Channel State Information) data of each Wi-Fi frame; s2, phase data interpolation is carried out on the CSI data of each Wi-Fi frame through a Wi-Fi signal receiving end, and a phase difference sequence of adjacent frames is calculated; s3, reconstructing a sending time sequence of the Wi-Fi signal sending end through the Wi-Fi signal receiving end and a self-defined function; and S4, performing first-order linear fitting according to the sending time sequence to complete carrier frequency offset estimation. According to the invention, the influence of other error factors in the CSI is avoided through the use of the zero sub-carrier, the problem of multiple candidate values of the CFO is avoided through the non-equal interval sending strategy, and finally the accurate estimation of the CFO based on the CSI is realized.

The invention provides a dual-current observer-based permanent-magnet synchronous motor control method. By the dual-current observer-based permanent-magnet synchronous motor control method, a motor encoder-free control technology is achieved. The dual-current observer-based permanent-magnet synchronous motor control method comprises the steps of firstly, building a permanent-magnet synchronous motor mathematical model; secondly, designing two current observers, and further obtaining a counter electromotive value from the two current observers; thirdly, verifying system stability by employing aLyapunov function; and finally, determining a system a parameter according to a simulation result. A sliding-mode variable structure method is employed during the design process of the current observers, the current estimation accuracy is effectively improved, the influence of external interference is greatly reduced by the dual current observers, and accurate estimation of a speed and a positionof the motor is achieved; and due to the application of a mechanical encoder in the permanent-magnet synchronous motor, the cost is increased, the reliability is reduced, thus, the various defects ofthe prior art are effectively overcome, and the dual-current observer-based permanent-magnet synchronous motor control method has high application value.

The invention discloses a sequential space-time adaptive processing parameter estimation method. The method mainly solves the problems of grid mismatching, high computational complexity and the like existing in space-time two-dimensional parameter joint estimation in the prior art. The method comprises the steps that an electromagnetic pulse emitter and a uniform linear array are deployed; the linear array is used for receiving echoes reflected by peripheral objects and conducting space-time signal modeling; a sampling covariance matrix of space-time signals is calculated; a space-time power spectrum is initialized; a space-time relative probability spectrum is established and prior knowledge is extracted; time domain parameter estimation is carried out on the basis of rotation invarianceof a pulse sequence; space domain parameter estimation is carried out on the basis of rotation invariance of the uniform linear array; meshing-free space-time parameter estimation parameters are matched to finally obtain space-time parameter estimation results. According to the method, the efficiency and precision of target characteristic parameter estimation in the space-time adaptive processingprocess are improved, and the method can be applied to the fields of airborne radar detection tracking, unmanned aerial vehicle communication and the like.