33 results about "Error surface" patented technology

The invention discloses a device and method of interferometry for a large-scale optical plane. Wherein, the interferometry device comprises an adjustment platform for the two-axis linear motion of an interferometer, a two-dimensional tilt adjustment platform for the measured plane mirror located at the front of the adjustment platform, a laser wave-front interferometer installed in the adjustment platform for the two-axis linear motion of an interferometer and a control computer contained the program of measured data processing algorithm, which is connected to the laser wave-front interferometer. With the device, a plurality of obtained error surface maps of partial region can be spliced to an error surface map containing medium-high frequency (MHF) band on the whole caliber according to the measured data processing algorithm through the control computer, which includes a method for determining the initial position, an algorithm for extracting data from the overlap region and an algorithm for splicing the area data. The invention is a device and method of interferometry for a large-scale optical plane with low cost, high precision and high efficiency.

The invention discloses a correction method of an electronic compass in an underwater vehicle, which comprises the following steps of: inquiring magnetic declination, and carrying out corrected navigation on the water surface of the vehicle; recording navigation data by a navigation sensor; exporting correction navigation data from a communication module; searching a hard magnetic interference parameter, solving the error surface plot between a correction track and a reference track under the condition of (2N+1)<2> disperse hard magnetic interference parameters, wherein the minimum point corresponding to the surface plot is the search result; according to the set correction precision requirement, repeating and reducing the search range and grid, and re-searching to obtain a more precise hard magnetic interference parameter evaluation value; and finally, applying the searched magnetic parameter evaluation value in the navigation of the underwater vehicle. The method disclosed by the invention has the advantages that the operation is simple, the influence of the outer magnetic interference to the electronic compass in the underwater vehicle can be effectively eliminated and thus, the corrected electronic compass can return back to the corrected attitude angle.

A fast search algorithm for loop alignment of a feed forward amplifier is disclosed. The algorithm controls a processor that adjusts, digitally, the gain and phase of the loop alignment based on power measurements at the input and output of cancellation combiners, to find the optimal setting. A “non-collinear descent” algorithm is used to search, iteratively, for the minimum within an error surface. For loop alignment, the error surface is defined by the set of measurement points comprising the alignment settings and the associated cancellation residuals. For the case of first loop alignment, the cancellation residual is measured using the ratio of two power detectors located at the input and after the cancellation (error) combiner. For second loop alignment, cancellation is estimated using the residual pilot power detected at the output of the amplifier. The preferred alignment method uses three successive measurements to estimate the gradient direction with respect to gain and phase shifter settings. The actual descent direction is selected to be close to the gradient direction without being collinear with the most recent alignment adjustments. Quantization of the descent direction simplifies the implementation as well as the enforcement of the non-collinearity constraint on successive alignment settings. Two different step size selection approaches are disclosed, however, any standard step size selection approach may be employed.

Given that decoder side motion vector refinement / derivation is a normative aspect of a coding system, the encoder will also have to perform the same error surface technique in order to not have any drift between the encoder's reconstruction and the decoder's reconstruction. Hence, all aspects of all embodiments are applicable to both encoding and decoding systems. In template matching, the refinement movement occurs only in the reference starting from the sub-pixel accurate center that is derived based on the explicitly signaled merge index or implicitly through cost evaluations. In bilateral matching (with or without averaged template), the refinements start in the reference lists L0 and L1 starting from the respective sub-pixel accurate centers that are derived based on the explicitly signaled merge index or implicitly through cost evaluations.

The embodiment of the invention discloses a convergence method of a high-dimensional depth learning model and a device. The method includes the steps of performing one unit iteration on a model according to a first position of the error surface to determine a solution of the model at a second position of the error surface; determining a gradient and a curvature of the second position relative to the error surface according to one unit iteration, and determining an accuracy increasing efficiency and a model error of the solution of the model according to the first position and the second position; determining whether the second position is a saddle point or a high noise point of the error surface based on the gradient, the curvature, the accuracy increasing efficiency and the model error; and when the second position of the error surface is a saddle point or a high noise point, adjusting the batch number of the next unit iteration. Embodiments of the present invention determine the batch number of the model at the next batch iteration based on the saddle point or high noise. Therefore, it is possible to optimize the accuracy increasing efficiency and the escape saddle point of the model by adjusting the gradient and stochastic noise estimation in the iteration process.

The invention, which belongs to the field of the information technology, relates to an energy consumption prediction method for improving an energy utilization rate of an iron and steel enterprise. A neural network algorithm with error feedback correction learning is used for carrying out prediction. In order to solve a problem of slow convergence speed of the traditional learning method, a momentum is added into an adjusting formula of a weight value and a threshold value, thereby accelerating a training speed of a network. With an adaptive adjusting learning rate, a learning rate of training at an error curved surface flat area is increased; and a learning rate is reduced at a drastic error changing area. On the basis of time-varying-weight-based parameter combination optimization method, distribution of an initial weight value and an initial threshold value is optimized, and good searching space is localized in solution space to avoid a local minimal point. According to the energy consumption prediction method, characteristics of energy consumption of all procedures are taken into consideration; the energy utilization rate is improved; and the energy costs are reduced and thus total production cost is lowered.

The invention provides a method for automatically compensating the error of the measurement surface change of a network analyzer, step 1: calibrate the network analyzer or call the calibration state; step 2: change the measurement surface, and send a measurement surface automatic compensation command; step 3: call the measurement surface to automatically Compensation fitting algorithm, obtain fitting curve parameters and calculate compensation data; Step 4: store parameters and compensation data. Using the above scheme, the calibration compensation algorithm is realized by using the curve fitting method, and the error introduced by the change of the measurement reference surface can be removed without user intervention. The obtained compensation data is simple to calculate and fast to run, improving the user's measurement efficiency.

The invention relates to a method for determining assembly-oriented rectangular plane shape error evaluation parameters and belongs to the field of error evaluation. The method comprises: generating a cutting surface by non-Gaussian plane simulation method, acquiring a shape error surface via wavelet filtering, and extracting characterization parameters of surface shape error distribution; subjecting the shape error surface to simulation assembly through contact algorithm, calculating assembly precision of spatial orientation changes of a second part after assembly, determining a relation between surface characterization parameters and post-assembly part assembly precision through correlation analysis, thereby determining shape error evaluation parameters and enabling quantitative description for the relation between the surface distribution characterization parameters and assembly precision. Plane characteristic parameters having important impact on assembly precision are obtained, and scientific basis is provided for improving machining quality of mating planes and optimizing assembly process to improve assembly precision.

The invention relates to a calculation method for the inverse problem of ampacity. By comparing the cable surface temperature monitored in real time and the cable surface temperature data in the dynamic ampacity analysis, the error of the parameters in the cable and the laying environment is corrected to obtain the corrected parameter values. The calculation result is accurate, which effectively improves the reliability and accuracy of the ampacity analysis software.

The invention discloses a high-frequency error detection device and method in the large caliber large relative aperture non-spherical mirror, the device including the five-axis movement adjustment platform with the interferometer focusing platform, the side swing reflecting mirror side swing platform located in front of the interferometer focusing platform, and the measured non-spherical mirror 3D movement adjustment platform located below the side swing reflection mirror side swing platform, and the multiple points supporting machine with the laser wave surface interferometer, the side swing reflection mirrors, and the measured non-spherical mirror installed on the corresponding platforms, and the main control computer with built-in detection data-processing algorithm program connecting with the laser wave surface interferometer. The device uses the main control computer to process the detection data-processing algorithm, which can combine the detected multiple part regions error surface maps into error surface map with medium or high frequency in full caliber, including the initial pose determining method, the overlapping regional data extraction algorithm and the regional data stitching algorithm. The invention is a high-frequency error detection device and method with low-cost, high-precision, high-efficiency in the large caliber large relative aperture non-spherical mirror.

Symbol sampling in a high time delay spread interference environment includes acquiring (602) a time varying baseband waveform. The waveform has a signal amplitude that varies between one of a plurality of symbol states. The waveform is sampled (603) at a rate of m times the symbol rate. During an evaluation time, an error value is calculated (604, 606) for each of m data sample positions. Each of the error values comprises an average distance between the measured value of the waveform as indicated by the data sample and a closest known symbol value. The error values are used to create an error surface. Thereafter, the error surface is modeled as a quadratic and an optimal sample time is determined (608, 610, 612) based on finding the time location where the quadratic surface is minimum. A sinc interpolator is then used to resample the data.

The embodiment of the invention discloses a convergence method and device for a high-dimensional deep learning model. The method includes performing a unit iteration on the model according to the first position of the error surface to determine a second position of the solution of the model on the error surface; according to a unit iteration, determining the gradient and curvature of the second position relative to the error surface, And determine the precision growth efficiency and model error of the solution of the model according to the first position and the second position; According to the gradient, curvature, precision growth efficiency and model error, determine whether the second position is a saddle point or a high noise point on the error surface; When the second position of the error surface is a saddle point or a high noise point, the batch number of the next unit iteration is adjusted. In the embodiment of the present invention, the batch number of the model in the next batch of iterations is determined according to the saddle point or high noise. In this way, by adjusting the gradient and random estimation noise in the iterative process, the accuracy growth efficiency of the solution of the model can be optimized and the saddle point can be escaped.

A hypersonic vehicle attitude adaptive recursive control method and system, the steps are as follows: (1) Combining the control target, the aircraft re-entry attitude dynamics analysis and the unmodeled dynamics and disturbance analysis, an attitude error characteristic model with an expanded state is established , where the expanded state is used to describe the unmodeled dynamics and disturbances of the system; (2) construct a fuzzy neural network estimator to estimate the expanded state in the feature model, and combine the structure of the feature model to jointly design the parameters of the feature model and the automatic parameters of the neural network Adaptive law; (3) Based on the error surface, a recursive adaptive control law is designed to further improve the robustness of the control system. The method of the invention has better adaptability to uncertainty, has the advantages of strong robustness, high control precision, etc., and is suitable for accurate tracking of attitudes of hypersonic aircraft in complex unmodeled dynamics and strong interference environments.

A method for background calibration of sampler offsets in an Analog to Digital Converter (ADC), according to which one of the samplers of the ADC is established as a reference sampler, whose threshold and timing offsets will be the criterion for adjusting threshold offsets and timing offsets of all other samplers. Then each of the other samplers of the ADC, one at a time, is calibrated by selecting an uncalibrated sampler and establishing it as the current Sampler Under Calibration (SUC); disregarding contribution of the SUC to the output of the ADC; adjusting the threshold of the SUC to be identical to the threshold of the reference sampler; performing one-bit cross-correlation between the reference sampler and the SUC; establishing an error surface representing the threshold offset and timing offset of the SUC with respect to the reference sampler; adjusting the threshold and the timing of the SUC to be equal to the threshold and timing of the reference sampler; restoring level of the SUC to its original threshold with respect to the overall ADC and restoring contribution of the SUC to the output of the ADC.

The embodiment of the invention provides a solid modeling method and device for a joint surface error. The method comprises the following steps of establishing a geometrical characteristic error surface of the joint surface based on the geometrical characteristic error points of the joint surface in a machine tool structure; based on geometrical feature error surface and ideal solid model, establishing the geometrical feature error model of joint surface, in which the ideal solid model is machine tool structure model without geometrical feature error points constructed by computer. The invention provides the solid modeling method and device for the joint surface error, which adopts the geometric characteristic error points of the joint surface of a machine tool structure to construct a geometric characteristic error surface, by combining the geometric feature error surface with the ideal solid model, the geometrical feature error model is established, and the geometrical feature errormodel can truly reflect the joint surface feature error between the actual parts and the ideal solid model constructed by computer, which provides a basis for the machine tool structure optimization design and the accurate calculation and analysis of assembly characteristics prediction.

The invention discloses a device and method of interferometry for a large-scale optical plane. Wherein, the interferometry device comprises an adjustment platform for the two-axis linear motion of an interferometer, a two-dimensional tilt adjustment platform for the measured plane mirror located at the front of the adjustment platform, a laser wave-front interferometer installed in the adjustmentplatform for the two-axis linear motion of an interferometer and a control computer contained the program of measured data processing algorithm, which is connected to the laser wave-front interferometer. With the device, a plurality of obtained error surface maps of partial region can be spliced to an error surface map containing medium-high frequency (MHF) band on the whole caliber according to the measured data processing algorithm through the control computer, which includes a method for determining the initial position, an algorithm for extracting data from the overlap region and an algorithm for splicing the area data. The invention is a device and method of interferometry for a large-scale optical plane with low cost, high precision and high efficiency.