12 results about "Maximum error" patented technology

The invention relates to a measuring head measurement compensation method and system based on multi-dimensional error data. The method comprises the steps of collecting actual coordinate values of the X axis, the Y axis, the Z axis and the four space diagonal directions of a workpiece, obtaining actual coordinate deviation values in the corresponding directions through calculation so as to obtain actually-measured multi-dimensional error data of the workpiece, obtaining an error trend fitting line of the multi-dimensional error data through calculation, and combining the actually-measured multi-dimensional error data to obtain an error trend fitting line of the workpiece. Constructing a composite error coupling model, calling a pre-stored compensation parameter library in real time according to the composite error coupling model, dynamically correcting the measuring parameters of the measuring head to generate self-adaptive compensation parameters, and performing secondary measurement verification on the compensated measurement data, when the maximum error value is smaller than a preset error threshold value and the average error value meets a preset precision threshold value, compensation is completed, and otherwise, model construction and parameter compensation are executed again. According to the invention, through integration of mechanical error acquisition and dynamic parameter correction compensation, precision optimization of the measuring head under a complex working condition is realized.

A method is disclosed of selecting users for multi user multiple-input multiple-output (MU-MIMO) communication from an initial set of potential users. The method comprises repeating the following steps in successive iterations until a stopping criterion is met: determining a channel correlation metric for each user in the set of potential users, reducing the set of potential users by exclusion of a user based on the channel correlation metric, and calculating a performance metric of the set of potential users. The method also comprises selecting users corresponding to one of the sets of potential users, wherein the selection is based on a comparison of the calculated performance metrics of the sets of potential users. The channel correlation metric for a user may comprise one or more of: a channel filter norm for the user, a channel norm for the user, a channel gain for the user, pair-wise correlations between the user and one or more other users of the set of potential users, and a channel eigenvalue for the user. The performance metric may comprise one or more of: a sum-rate, a per-user-rate, an average error rate, a maximum error rate, a per-user error rate, and a sum-correlation. Corresponding apparatus, network node and computer program product are also disclosed.

The application discloses a kind of based on best fitting accuracy reflector antenna actuator layout method, comprising: obtaining the average of reflector deformation under the condition of looking up and pointing flat, and the deformation of reflector at different pitch angles is pre-adjusted;The pre-adjusted deformation reflector is obtained by translation and rotation to best approximate ideal reflector, obtain the maximum error distribution range, determine the layout scheme of actuator corresponding to reflector;Determine the adjustment target surface when reflector antenna is in the condition of looking up, obtain the adjustment amount of actuator when reflector antenna is in the condition of looking up;Obtain the adjustment accuracy when reflector antenna is in the condition of looking up;Update the preset error threshold value, and correspondingly update adjustment accuracy, until the curve of adjustment accuracy under all conditions with preset error threshold value is obtained;According to the adjustment accuracy of engineering requirement, the corresponding error threshold value is obtained on the curve, and the layout scheme of actuator is obtained.The application can obtain the best actuator layout scheme.

The application belongs to the technical field of structure optimization design, and discloses a microstructure proxy model establishing method, application and equipment, which comprises the following steps: (1) constructing a sample point database and then constructing an initial proxy model; (2) selecting each sample point in the sample point database in turn by using an orthogonal least square method and calculating the error caused by each sample point, then selecting the sample point corresponding to the maximum error as the first sample point of the final proxy model; (3) selecting the remaining sample points in turn, making the feature vectors of the sample points orthogonal to the feature vectors of the selected sample points, calculating the error caused by the sample points, then selecting the sample point corresponding to the maximum error in this round as the next sample point, and repeating the selection until the sum of the errors of the sample points of the final proxy model is greater than or equal to a threshold value; (4) constructing the final proxy model based on the sample points of the final proxy model and the corresponding weights. The application shortens the solving time.

According to the function waveform generation method and system based on expression compilation and adaptive segmented sampling, the combination of an upper computer, an ARM, an FPGA and a DAC is adopted, and the upper computer inputs a segmented function set and parameters; the ARM processor executes expression compilation and completes segmentation coverage degree and continuity verification; in each section, according to second-order difference curvature approximation, under the constraint of an error upper limit epsilon, adaptively determining a step length for sampling, then performing uniform resampling at a global equidistant position, and performing linear calibration and quantization according to y '= alphay + beta to generate a waveform lookup table; a DDS in the FPGA calculates a frequency control word FCW and a phase control word POW according to a work clock and a target frequency, a phase accumulator looks up a table to output a digital waveform, and finally the digital waveform is converted into an analog signal by a DAC. According to the scheme, the maximum error of the amplitude is controlled, meanwhile, the sample point and table scale is remarkably reduced, the bandwidth is reduced, the frequency and the phase are independently adjustable, and the expression ability and the freedom degree of waveform generation are greatly improved through a segmented expression.

The invention discloses a channel noise distribution probability classification sorting and guess decoding complexity limit calculation method and system, and the method comprises the steps: S1, obtaining a binary channel noise sample sequence through a receiving end, and carrying out the classification sorting of the channel noise distribution probability and the guess decoding complexity limit calculation; calculating a noise parameter of each sample sequence; s2, calculating distribution probabilities corresponding to channel noise samples according to the calculated noise parameters of each sample sequence, and sorting the distribution probabilities from large to small; and S3, calculating the limit of the guess decoding complexity according to the sorted distribution probability and the preset maximum error correction capability. According to the invention, the decoding complexity is reduced, the consumption of computing resources is optimized, and the decoding process is more efficient. Therefore, more possible error modes are sequenced preferentially, and the success rate of error correction is improved.

Provided are an RS code decoding method and device. In the present application, a concomitant formula is determined according to a code word to be decoded, the code word being an RS encoded code word (S301); an error position polynomial and a minimum length satisfying a key equation are determined according to the concomitant formula (S302); if the minimum length satisfying the key equation is greater than a maximum error correction capability of the RS encoding (S303), an uncorrectable error is detected, and decoding of the code word is abandoned, otherwise, an error position is determined according to the error position polynomial; if the number of the determined error positions is not equal to the minimum length satisfying the key equation (S305), an uncorrectable error is detected, and decoding of the code word is abandoned, otherwise, the code word is corrected according to the error positions, and a decoding result of the code word is obtained (S307). The present application can detect uncorrectable errors, thereby improving system performance and reliability.

The invention discloses a sampling test method and device for a flash memory chip and a medium. The method comprises the steps that the maximum error bit number of the unit code length of the flash memory chip and the maximum page error rate are determined; dividing the blocks into a plurality of same areas; for each area of all the blocks, sorting the unit code length error bit numbers from large to small, selecting the unit code length error bit numbers of a first proportion, and recording a page set where the unit code length error bit numbers are located; for the same area of all the blocks, selecting the page with the occurrence frequency exceeding a first threshold value; combining all the areas to obtain a block test page address combination; performing erasing, writing and reading operations on the block test page address combination of all blocks of the flash memory chip to obtain the maximum error bit number of unit code length and the maximum page error rate; and if the absolute value of the front and back difference values is within the preset threshold value, testing the flash memory chips of the same specification and batch by using the block test page address combination. According to the invention, the test time and cost can be effectively reduced while the test result is ensured to be reliable.

The invention relates to a method for realizing servo and system linkage debugging in a numerical control system. The method comprises the following steps: inputting linkage debugging parameters to the numerical control system; the driver parameters of all the feed shafts are automatically set to be in a load characteristic learning mode, the feed shafts are controlled to execute preset reciprocating motion, and the inertia ratios of the shafts are obtained; automatically setting a standard mode, and setting an initial rigidity value; performing linkage circle drawing motion with the other feeding shafts in sequence; fitting an actual motion trail circle according to feedback data of the encoder, and calculating to obtain a whole circle error; and judging whether the whole circle error is smaller than or equal to the maximum linkage error value, and if so, completing linkage debugging. By the adoption of the method and device for achieving servo and system linkage debugging in the numerical control system, the processor and the computer readable storage medium of the processor, servo debugging can be conducted under the condition that manual intervention is not needed, linkage errors of all shafts can be automatically optimized, and guarantee is provided for the machining precision of a machine tool.

The invention provides a Huber loss function configuration method for a deep learning model, and the method comprises the steps: collecting an error between the model output of a current batch of samples in a current iteration round and a real label in a training process of the deep learning model, and forming an absolute error set; the absolute error set and the historical error set are combined, the last K elements of the combined set are obtained, a current window error cache is formed, and K is a positive integer; determining the median of the current window error cache; determining a target hyper-parameter according to the historical hyper-parameter and the median; and updating a Huber loss function according to the current hyper-parameter, and carrying out training based on a next batch of samples. According to the method, the typical value of the error distribution is captured through the median of the sliding window, and the error distribution change is quickly responded based on the typical value, so that the dynamic adaptation of the hyper-parameter is realized, the outlier interference is avoided, the anti-abnormal value robustness is realized, the prediction precision of the traditional Huber is effectively improved, and the maximum error is reduced.

This invention relates to a method for reducing the peak-to-average power ratio (PAPR) of an OFDM system. It involves extending reserved subcarriers to the data subcarrier frequency band, constraining the maximum error vector of the system, and reducing the PAPR. The method includes: first, establishing an primal optimization problem with the OFDM signal PAPR as the objective function and the maximum error vector on the data subcarriers and the power of the reserved subcarriers as constraints; second, decomposing the primal problem into three subproblems using the ADMM algorithm, alternately solving each subproblem and iterating repeatedly until the globally optimal solution is obtained. Compared with existing technologies, this invention has advantages such as avoiding excessive interference that leads to an increase in the system's bit error rate and achieving better PAPR reduction.

A system and method for an estimation of error bounds for file size calculations using MinHash in deduplication systems. The system includes one or more processors to determine a similarity score between the first file and the second file. The one or more processors are further to determine a size estimation of a combination of the first and second files based on the similarity score. Finally, the one or more processors are to determine a maximum error for the size estimation of the combination of the first and second files, wherein the first and second file are to be combined via deduplication and have at least one shared data segment.