142results about How to "Guaranteed estimation accuracy" patented technology

The invention belongs to the technical field of precision high-efficiency numerical control machining, and relates to a contour error real-time estimation method based on spatial arc approximation in a spline curve interpolation process. According to the method, a parameter initial value for Newton iteration on a parameter curve is calculated based on a first-order Taylor series expansion method before a perpendicular foot is calculated in each Newton iteration, and then a single iteration parameter final value is calculated according to the parameter initial value by using a Newton iteration method. In order to avoid greatly increasing the calculation time of the algorithm, circulation is ended under double constraints of an iteration terminating precision condition and a maximum iteration number condition to obtain a parameter estimation value of the perpendicular foot, and a contour error vector estimation value is calculated. The method still can effectively guarantee the estimation precision of a contour error when the follow-up error is relatively large, and has great significance for improving the spline curve interpolation machining contour precision of a high-grade numerically-controlled machine tool.

The invention discloses a contrast optimization self-focusing method based on a hypersonic platform synthetic aperture radar (SAR), which mainly solves the problems of dependence on initial closed interval division and low convergence rate in the prior art. The method comprises the following steps of: 1) transmitting linear frequency modulation pulse signals to the ground by using a radar; 2) receiving radar echo signals by using an antenna and performing distance-to-pulse compression on the echo signals; 3) selecting M distance units with highest signal intensity from the echo signals subjected to the distance-to-pulse compression and performing Doppler frequency modulation rate estimation on the selected distance units; 4) calculating the Doppler frequency modulation rate of each distance unit of the echo signals according to the optimal Doppler frequency modulation rate estimation value; and 5) constructing reference function according to the Doppler frequency modulation rate and performing orientation-to-pulse compression on the echo signals subjected to the distance-to-pulse compression to acquire the final SAR imaging result. The method has the advantages of high convergence rate and no dependence on initial closed interval division and can be applied to motion compensation of hypersonic platform SAR imaging.

The invention relates to a single-image human face posture estimation method based on a depth value, and the method comprises the following steps: obtaining a human face with a to-be-estimated posture; converting the human face with a to-be-estimated posture into a gray scale image, and carrying out the noise reduction and standardization processing; introducing an affine transformation invariance initialization principle to build an initial model, improving a conventional active shape model method, and extracting the two-dimensional values of ten human face feature points; solving the depth values of ten human face feature points through the perspective imaging principle through employing an iteration algorithm; estimating a human face posture estimation target function; carrying out the primary estimation of the human face posture through four human face feature points, enabling the primary estimation value to serve as the input value for the least square optimization calculation of a target function, and outputting a human face posture result of secondary estimation; carrying out the further optimization of a least square optimization calculation result through employing an M-estimation algorithm, and outputting a final estimation result. The invention overcomes a pathological problem of posture estimation, and is good in posture estimation precision and robustness.

The invention discloses a millimeter-wave system channel estimation method low in complexity. The method includes the following steps: firstly, a receiving end uses a first-order codebook to estimatea pair of azimuths AOA and AOD, and then establishes a set on the basis of the defined quantization precision estimated the second time and the number of quantization angles estimated the second time;secondly, the receiving end uses the set obtained in the first step and also a second-order codebook to conduct azimuth estimation at the second-order precision; and thirdly, the receiving end utilizes a least square method to estimate fading of channels corresponding to the azimuths, and determines whether channel estimation is carried out according to a result. Through utilization of sparsity of a millimeter-wave system channel and redundancy of a sparse dictionary in a compression sensing technology, a channel estimation process is quantified through angles to be decomposed into two-complexity channel estimation processes. According to the method, the number of searching of the sparse dictionary is reduced, the millimeter-wave system channel estimation efficiency is increased, and thechannel estimation complexity of channel estimation is reduced effectively while the estimation precision is guaranteed.

The invention relates to a multifunctional control system and a control method of a battery used in a substation, which is characterized in that the multifunctional control system comprises a currentacquisition module, an input end of the current acquisition module is connected with a current sensor, and an output end is connected with an input end of an A/D converter. The input end of the temperature acquisition module is connected with the temperature sensor, and the output end is connected with the input end of the A/D converter. The output of A/D converter is connected with CPU through I/O interface. The CPU is also connected with human-computer interaction module, discharge load and voltage balancing module through I/O interface, and the discharge load and voltage balancing module are electrically connected with each single battery respectively. The control method comprises the following steps of: balancing control of a battery pack; SOC estimation of battery pack; measurement ofinternal resistance of battery pack.

The invention discloses a method for estimating quick global motion based on motion vector cancellation and a difference principle. The method comprises: firstly, inputting the current frame and a reference frame of a video image, and acquiring a motion vector of each pixel point of the video image by an optic flow method to form a motion vector field; secondly, constructing a global motion model of the video image on the motion vector field; thirdly, estimating a translation parameter T of global motion by utilization of the characteristic that motion vectors in diagonal and symmetrical positions on the motion vector field can cancel other motion components except a translation motion component; fourthly, estimating a conversion parameter A of the global motion by utilization of the difference principle of motion vector pairs on the same row and motion vector pairs on the same column on the motion vector field; and fifthly, substituting the translation parameter T and the conversion parameter A into the global motion model, and completing global motion estimation of the video image. The method greatly shortens the computation time under the condition of guaranteeing the estimation precision, and can be applied to a video processing system with higher real-time requirement.

The invention relates to a consistency clock synchronization relative frequency offset estimation method based on a weighted median, and belongs to the technical field of wireless sensor networks. Themethod is oriented to a wireless sensor network scene in which communication time delay is exponentially and randomly distributed; a relative clock relation between nodes is considered, a Laplace variable model is obtained through timestamp data processing, then a weighted median method is adopted to estimate relative frequency offset, logic clock parameters of the nodes are compensated in combination with an average consistency method, and finally the effect that logic clocks of the nodes in a network are kept synchronous is achieved. The invention provides two relative frequency offset estimation methods, namely an interval-based weighted median estimation method and a threshold-based weighted median estimation method. The time stamp interval is estimated by searching an optimization value of the time stamp interval, and the method is suitable for scenes with high requirements on relative frequency offset estimation precision; the latter reduces the storage requirement of the estimation method by setting a threshold value, and is suitable for a node storage resource limited scene.

The invention discloses a method for estimating the attitude angles of a rescuing and obstacle-clearing vehicle. According to the method, the rescuing and obstacle-clearing vehicle is subjected to dynamics modeling according to the work characteristics of the rescuing and obstacle-clearing vehicle; the estimated values of the attitude angles, such as side rake angle, pitch angle and the like, of the rescuing and obstacle-clearing vehicle are obtained by a recursive least-square method, and can be applied to attitude monitoring and dangerous attitude early warning of the rescuing and obstacle-clearing vehicle, so the method has the significant advantages of high precision, low cost, high real-time property and the like.

The invention discloses a method for tracking targets of a wireless sensor network on the basis of information consistency weight optimization. The method includes (1), initializing prior information vectors and prior information matrixes of various nodes; (2), computing observation vectors of the various nodes; (3), computing local observation vectors and local observation matrixes of the various nodes; (4), broadcasting information of the various nodes to neighbor nodes; (5), computing fused local observation vectors and fused local observation matrixes of the various nodes; (6), optimizing consistency weights; (7), computing consistency-fused local information vectors and consistency-fused local information matrixes of the various nodes; (8), acquiring information vector consistency estimation and information matrix consistency estimation of the various nodes; (9), acquiring target state predicted values and target state estimated values of the various nodes; (10), predicting and updating the prior information vectors and the prior information matrixes of the various nodes. The method has the advantage that the state estimation accuracy and the state estimation consistency of the various nodes in the sensor network are improved.

The invention discloses a power grid harmonic source positioning algorithm based on wide-area monitoring. The positioning algorithm comprises the following steps: applying power distribution line measurement to harmonic state estimation and harmonic source positioning, taking the voltage current and like state variables collected by a wide-area monitoring system as the basis, proposing a specificoptimization algorithm according to the weighted least square method, and realizing the power grid harmonic source positioning by utilizing the phasor difference method. The harmonic source positioning based on the wide-area measurement system sufficiently plays the wide-area low-cost characteristics, the operation process is optimized, the estimation precision is guaranteed, the full-network harmonic voltage state is obtained, and the harmonic source detection sensitivity and the positioning accuracy are much better than the traditional detection. And the method disclosed by the invention hasimportant significance for the analysis processing of the power grid harmonic source.

The invention discloses a non-cooperative signal sensing system and a multi-dimensional parameter estimation method of a satellite-borne passive SAR and belongs to the technical field of radars. The invention provides a non-cooperative SAR signal sensing system and the non-cooperative SAR signal multi-dimensional parameter estimation method. The non-cooperative SAR signal sensing system comprises a zenith sensing antenna, a direct signal array receiving unit, a ground detection antenna and an echo signal array receiving unit, and receives non-cooperative SAR direct signals and echo signals. The non-cooperative SAR signal multi-dimensional parameter estimation method involves a direct-arrival signal time-frequency-space characteristic estimation unit and an antenna beam pointing parameter estimation unit and achieves multi-dimensional parameter estimation and antenna pointing estimation of non-cooperative SAR signals. Compared with a traditional satellite-borne SAR system, the satellite-borne passive SAR does not need to transmit any signal, can sense and capture available non-cooperative SAR signal sources in space in real time, and uses the non-cooperative SAR signal sources to perform double-station SAR imaging detection.

The present invention discloses a lithium battery SOC estimation method for preventing from filter divergence, and relates to the technical field of electric car lithium battery state estimation. Themethod mainly comprises the steps of: I, obtaining a lithium battery SOC starting value through an open-circuit voltage method; II, considering influences of system noise and observation noise, and employing an improved Sage-Husa adaptive extended Kalman filtering algorithm to estimate a real-time SOC value; and III, employing filter divergence criterion for the previous step to perform filter divergence determination, when the filter divergence determination does not meet a criterion condition, constructing an index freezing factor at a Kalman gain matrix to avoid filter divergence. On the basis of the extended Kalman filtering algorithm, a noise estimation value device is added to ensure the estimation precision of the SOC, avoid the generation of filter divergence in a severe work condition and improve the stability of the system estimation.

The invention provides an OFDM system fast-change channel estimation method based on double weighing. The method includes the following steps that pilot signals are extracted from receiving data of a system, and the channel response at a pilot frequency position is estimated; for each pilot frequency sub-carrier, a predicated value of the channel response at the data position of the corresponding pilot frequency sub-carrier is obtained through forward prediction; the predicted values of the channel responses are used for conducting balance and judgment on the received data, feedback information is obtained, the predicted values of the channel responses are corrected, and the channel responses of the whole pilot frequency sub-carriers are obtained; the channel response of each pilot frequency sub-carrier is used for conducting balance and judgment on the receive data, and error parameters of the channel response of each pilot frequency sub-carrier are obtained; interpolation coefficients of each pilot frequency sub-carrier are calculated according to the error parameters, interpolation in the frequency direction is performed, and the channel response of the system is obtained. According to the OFDM system fast-change channel estimation method based on double weighing, effective tracking of a fast time-varying channel can be achieved, and the accuracy of channel estimation is improved.

The invention discloses a wireless sensor network target tracking method based on informational consistency. The method includes: (1) initiating a prior information vector and a prior information matrix of each node; (2) calculating an observation vector of each node; (3) calculating a local observation vector and a local observation matrix of each node; (4) enabling each node to broadcast information to adjacent nodes; (5) calculating a fused local observation vector and a fused local observation matrix of each node; (6) calculating a consistency fusion local information vector and consistency fusion local information matrix of each node; (7) acquiring an information vector consistency estimation and an information matrix consistency estimation of each node; (8) acquiring a predictive value and an estimation value of a target state of each node; and (9) predicting and updating the prior information vector and the prior information matrix of each node. Information filtering and consistency strategies are combined, and accordingly accuracy and consistency of state estimation of each node in a sensor network are improved.

The invention discloses a quad-rotor torque model/directional gyro/magnetic sensor fault-tolerant integrated navigation method. The method comprises the following steps: periodically reading information of a quad-rotor aircraft airborne sensor at a moment k; executing a fault detecting filter and a fault location strategy, and judging faults of a z-axis gyro, a torque model and a magnetic sensor;determining a state equation of each sub-filter, and performing data fusion; according to a fault detecting result, performing fault isolation; executing a global filter, and performing the data fusion on the sub-filter without the fault, to obtain course angle information; resetting state quantities of each sub-filter and the fault detecting filter, and executing a system reset strategy. The method is capable of, through the torque model of a quad-rotor aircraft, forming redundancy of the z-axis gyro, realizing fault detection of the z-axis gyro, the magnetic sensor and the torque model and the system reset, and acquiring accurate estimation of an attitude angle and a course angle still while the gyro is failure.

The invention discloses a rotor position and rotating speed detection method based on counter potential open-loop estimation, and belongs to the technical field of closed cycle power generation. The method comprises the following implementation steps: firstly, a linear tracking differentiator is introduced to preprocess winding current to suppress noise, then a variable parameter low-pass filter is adopted to filter back potential of a motor, delay of PWM sampling is compensated at a high rotating speed, and finally, a dead zone compensation unit is introduced to compensate voltage waveform distortion caused by an inverter dead zone. Stable power generation of the power generator in the closed circulation system can be achieved, the motor body does not need to be provided with a position sensor, and the reliability of the system is improved.

The invention discloses a direction of arrival estimation method based on grid partial refinement. Through learning and fission processes, in the fission process, new grid points are generated to refine the grid. The learning process is continuously close to the direction of arrival, grids near the real direction of arrival are finely divided, and grid areas far away from the direction of arrivalare coarsely divided, so that partial refinement of the grids is realized, the estimation precision is ensured, and compared with a previous off-grid DOA estimation algorithm, the number of grid points is greatly reduced, and the calculation amount is reduced accordingly. According to the method, the information source number does not need to serve as a priori, grid division is as sparse as possible, the grid number is reduced, therefore, the calculation complexity is reduced, and the algorithm consumes less time. And under the condition of very sparse initial lattice point division, the estimation precision of the algorithm is ensured through minimum interval threshold self-definition.