49 results about "Negative error" patented technology

A system and method for generating a neural network ensemble. Conventional algorithms are used to train a number of neural networks having error diversity, for example by having a different number of hidden nodes in each network. A genetic algorithm having a multi-objective fitness function is used to select one or more ensembles. The fitness function includes a negative error correlation objective to insure diversity among the ensemble members. A genetic algorithm may be used to select weighting factors for the multi-objective function. In one application, a trained model may be used to produce synthetic open hole logs in response to inputs of cased hole log data.

A reduced chop rate analog to digital converter technique including selectively weighting input samples to a digital filter, alternately inverting the polarity of an input error into positive and negative error components; and generating the positive and negative error components in a plurality of time response intervals of the digital filter in which the sum of the weights of the positive and negative error components are substantially equal.

An object of the present invention is to provide a method for encoding signals by separating a signal by using a common multiplier so as to improve the efficiency of compression of audio signals and a method for calculating a common multiplier of a set of numeric values including errors. A common multiplier A, of the input signal samples x(i) in each frame is estimated. Each sample x(i) is divided by A, the quotient is rounded to the number of significant digits, and the rounded quotient y(i) is encoded. The quotient y(i) is multiplied by A and the product is subtracted from x(i) to obtain an error signal z(i), which is then encoded. The multiplier A is encoded and the code of y(i), the code of z(i), and the code of A are combined together and outputted. To estimate the common multiplier, each numeric value is divided by a determined representative value. Irreducible fractions between the maximum positive error and the maximum negative error are obtained by dividing each numeric value by a correct representative value using rational approximation. Among the obtained irreducible fractions, the one having the smallest common denominator of the numeric values is obtained and the common denominator is corrected using the representative value to obtain a common multiplier.

The present method relates to the manual assistance for the automated indexing (100) of a collection of images using facial recognition. In a first automated indexing step, automated indexing of faces within a collection of images is performed, creating sets of faces each of which comprises faces that are determined by the automated process to be representative of the same person. In a second splitting step (200), sets are displayed to an operator, who determines whether there are false-positive associations within a set. If false-positive associations are found, the faces representing different people are manually split into different sets, hi this way, there will be no false-positive associations within the collection of images, hi a third merging step (300), sets that have some degree of similarity are presented to the operator, who determines whether the two sets comprise representations of the same person. If so, the two sets are manually merged, thereby eliminating false-negative errors. In this way, all of the faces in the image collection can be completely and accurately indexed.

The present method relates to the manual assistance for the automated indexing (100) of a collection of images using facial recognition. In a first automated indexing step, automated indexing of faces within a collection of images is performed, creating sets of faces each of which comprises faces that are determined by the automated process to be representative of the same person. In a second splitting step (200), sets are displayed to an operator, who determines whether there are false-positive associations within a set. If false-positive associations are found, the faces representing different people are manually split into different sets, hi this way, there will be no false-positive associations within the collection of images, hi a third merging step (300), sets that have some degree of similarity are presented to the operator, who determines whether the two sets comprise representations of the same person. If so, the two sets are manually merged, thereby eliminating false-negative errors. In this way, all of the faces in the image collection can be completely and accurately indexed.

A system and method for generating a neural network ensemble. Conventional algorithms are used to train a number of neural networks having error diversity, for example by having a different number of hidden nodes in each network. A genetic algorithm having a multi-objective fitness function is used to select one or more ensembles. The fitness function includes a negative error correlation objective to insure diversity among the ensemble members. A genetic algorithm may be used to select weighting factors for the multi-objective function. In one application, a trained model may be used to produce synthetic open hole logs in response to inputs of cased hole log data.

A system for measuring the weight of a seat occupant is used to control airbag deployment. The system includes a seat bottom that receives a vertical seat occupant weight force. The seat is divided into four quadrants, a front right quadrant, a left front quadrant, a left rear quadrant, and a right rear quadrant. Each quadrant includes a sensor that generates a weight signal corresponding in magnitude to the seat occupant weight in the respective quadrant. Each weight signal is comprised of a vertical force component and an error component induced by braking or deceleration forces acting on the seat bottom in a non-vertical direction. The sensors are orientated within their respective quadrants such that two sensors generate a positive error reading and two sensors generate a negative error reading. When the weight signals from each quadrant are added together to determine the total seat occupant weight, the error is eliminated as the positive and negative error components cancel each other out. The seat occupant weight is determined by adding the vertical force components from each of the weight signals together.

The present invention relates to a method for detecting malicious code patterns in consideration of control and data flows. In the method of the present invention, a malicious code pattern is detected by determining whether values of tokens (variables or constants) included in two sentences to be examined will be identical to each other during execution of the sentences, and the determination on whether the values of the tokens will be identical to each other during the execution is made through classification into four cases: a case where both tokens in two sentences are constants, a case where one of tokens of two sentences is a constant and the other token is a variable, a case where both tokens of two sentences are variables and have the same name and range, and a case where both tokens of two sentences are variables but do not have the same name and range. According to the present invention, it is possible to exclude a false positive error that may occur in conventional comparison of variable names and to lower a false negative error rate, thereby improving the accuracy of detection of malicious behaviors.

The invention relates to a self-adaptive digital DC / DC control method which has fast dynamic response and a converter. The self-adaptive control method related by the invention is implemented in a secondary side controller. In the control method, error detected is positive (negative) when sag (uprush) happens to an output voltage, and a first PID with good steady-state characteristics is switched to a second PID with good dynamic characteristics when the absolute value of the error exceeds a preset threshold value, the output voltage begins to enter a steady state when the error turns to be negative (positive), and the second PID is switched to the first PID when alternative changes between positive and negative errors are continuously detected, for which a steady state is deemed to be realized. The circuit structure of the DC / DC converter comprises a main power topology, an isolating transformer, a primary side controller, the secondary side controller, a driving circuit, a voltage and current detecting channel, a feedback and sampling channel, a primary side assisting source, a secondary side assisting source, a parameter memorizer and a communication interface. The control method of the invention has reasonable design and the converter is simple and practical.

The invention discloses a method for realizing a negative error frame in a CAN protocol and a device thereof. The invention determines the length of a delimiter of the negative error frame by changing condition of error data flow, carries out correction to the negative error frame or resets to an automatic machine in CAN protocol state and guarantees that the beginning time of the sending and receiving of the negative error node and all the nodes is synchronous, thus solving the fault that the negative error node is possible to be in actual off-line, improving the reliability of the CAN protocol communication and the utilization ratio of bandwidth, improving the safety of automobiles applying the CAN protocol, leads the network design tool software which is widely used currently to have a reliable theoretical basis and guaranteeing the continuity of industrial production.

The invention discloses a full-scan and high-throughput rapid screening detection method for illicit drugs in a health food. Data-dependent acquisition (IDA-MS) and data-independent acquisition (SWATH-MS) high-resolution mass spectrometry technologies are combined with Chemspider and an ACS compound structure database, illegally added western medicine ingredients in the health food are rapidly screened and analyzed according to the accurate parent ion and fragment molecular weights of primary mass spectrometry and secondary mass spectrometry and isotope peaks and other high-resolution mass spectrometry data information in the absence of a standard substance, and illicit drug ingredients in the health food are qualitatively identified under an unknown target substance condition in order toovercome the non-target substance detection problem of the health food. The method has the advantages of high-throughput analysis of the target substances, small matrix effect, high accuracy, simplicity in pretreatment, fastness, accuracy, small false negative error and small false positive error.

The invention belongs to weight sensing technology field. The flotage weight sensor is made up of flotage balancer, weighing liquid leakage preventing mechanism, shit signal transmitting device and magnetic centering suspending mechanism. It converts the linear relation of gravity value based on the flotage theory into prominent shift value, and outputs with electric signal. It has no technology deficiencies such as slow change and zero point, and sensitivity excursion. The invention also provides a method for flotage weight sensor in PID error control when in dynamic weight metering. What is detected and controlled by the method is error of the weighted material to the set weight in dynamic weighing process. The positive and negative error value measured by the method, they counteracted in the dynamic weighing process, thus the accuracy of weighing can be upgraded greatly.

A method for controlling the bit-rate of a bitstream of encoded video signals at a variable bitrate, the bitstream being generated by compressing a video sequence of moving pictures, wherein each picture comprises a plurality of macroblocks of pixels compressed by any of transform coding, temporal prediction, bi-dimensional motion compensated interpolation or combinations thereof, to produce any of I and / or P and / or B frames, the method involving quantization of said macroblocks effected as a function of a quantization parameter. The method includes defining a target bit-rate as well as maximum positive and negative error values between the target bit-rate and an average value of the current bit-rate of the bitstream, controlling the current bit-rate in order to constrain it between said maximum positive and negative error values, and defining an allowed range of variation for updating at least one reference parameter representative of the average value of the quantization parameter over each picture, wherein said allowed range is determined as a function of the target bit-rate and the maximum positive and negative error values.

First order gradient errors are canceled with no current source splitting by placing consecutive current sources symmetrically around the center of the array. Consecutive elements that correspond to small input amplitudes (mid-scale codes) make a smaller spatial jump than those correspond to larger signal amplitudes. Both linear and second order gradients are reduced by splitting each current cell into two and placing sub-elements symmetrically with respect to the center of the array to address the linear gradient effect. To address second order gradients, current element placement follows a pattern such that consecutive element pairs are chosen with one of the pair being placed with respect to the zero error contour of the second order gradient so as to have a positive error and the second of the pair being placed so as to have a negative error resulting in reduced second order error accumulation.

The invention discloses a composite prestressing force ceramic lining cylinder sleeve and a manufacturing method thereof; the cylinder sleeve comprises a metal outer sleeve and a ceramic lining, wherein the ceramic lining is sheathed in the metal outer sleeve by a heating assembly method, after cooling, the metal outer sleeve applies pressure stress to the ceramic lining and increases the mechanical strength of the ceramic so that the ceramic can be used for manufacturing the cylinder sleeve; wherein the inner diameter of the metal outer sleeve is identical to the outer diameter of the ceramic lining; the mismatching tolerance of the inner diameter of the metal sleeve is negative error, and the matching tolerance of the outer diameter of the ceramic lining is positive error. When the composite prestressing force ceramic lining cylinder sleeve provided by the invention is used, the exterior of the cylinder sleeve has no need to be cooled by a cooling water system, therefore, the radiation loss of the cylinder sleeve is reduced, the temperature of an air cylinder combustor is increased so that fuels are more rapidly and completely burned, and the heat engine efficiency of an internal combustion engine is improved; the manufacturing method of the composite prestressing force ceramic lining cylinder sleeve is simple, has higher requirements on the machining precision and can be realized by a modern numerical control machining technology.

The invention provides a hierarchical algorithm of selective laser sintering, which is used for establishing a precise product model. A method comprises the following steps: (1) reading and displaying an STL file; (2) locally encrypting an STL model; (3) selecting a layered slice plane; (4) sorting the layered slice plane; (5) compensating a positive and negative error; and (6) outputting a model. According to the hierarchical algorithm, through establishing a dynamic facet table for different tangent planes, the usage mount of a memory is lowered, time required by topological information is reduced, and the efficiency of hierarchical processing is improved; and meanwhile, through the method of locally encrypting the model and compensating the positive and negative error, the precision of the model is greatly increased, and a strong technical support is provided for the improvement of the precision of part manufacturing.

A reduced chop rate analog to digital converter technique including selectively weighting input samples to a digital filter, alternately inverting the polarity of an input error into positive and negative error components; and generating the positive and negative error components in a plurality of time response intervals of the digital filter in which the sum of the weights of the positive and negative error components are substantially equal.

Systems and methods described herein enable a residual error image to be added and rendered in an existing HTML canvas using native primitives. A current image and a residual error image containing pixel value differences between the current image and an updated image are received. A positive residual error image and a negative error image are generated from the residual error image. The positive residual error image is added to the current image to generate a partially updated image, and the partially updated image is xored to generate an inverse image. The negative residual error image is added to the inverse image to generate an updated inverse image, and xoring the inverse image to revert the inverse image to generate the updated image.

The invention discloses a photovoltaic power supply power prediction method based on positive and negative error variable weights. The method specifically includes the following implementation steps that firstly, n optical power prediction models are set, N prediction time points exist, and y(t) is a true value of photovoltaic power generation power at the tth moment; then, a prediction result is subtracted with the true value, and the weight of the tth moment of the ith model is calculated; finally, a final prediction result is obtained through n prediction values. The method solves the problems that prediction precision of a prediction method in the prior art can not meet requirements, and the prediction effect is poor.

We disclose a system, which performs a duty-cycle correction operation for an injection-locked phase-locked loop (PLL). The system first obtains a pattern of positive and negative error pulses at rising and falling edges of a reference clock signal for the injection-locked PLL, wherein the pattern specifies deviations of the reference clock signal from a 50% duty cycle. The system multiplies the pattern of positive and negative error pulses by a duty-cycle distortion (DCD) template, which specifies a sign of a duty-cycle error for the reference clock signal, to calculate duty-cycle distortion values. The system then accumulates the duty-cycle distortion values to produce a duty-cycle-error amplitude. Next, the system multiplies the duty-cycle-error amplitude by the DCD template to produce a duty-cycle correction signal. Finally, the system uses the duty-cycle correction signal to compensate for timing errors in the injection-locked PLL, which are caused by duty-cycle variations in the reference clock signal.

The invention discloses a pulse health data management system based on Internet-of-things communication. The system comprises wearable equipment, an Internet-of-things communication module, a smart handheld terminal, a cluster server end, a medical personnel end and a management personnel end; the wearable equipment collects human body pulse signals of a user through a built-in pulse sensor, processes and analyzes the signals through a built-in processor to form heart rate data and wirelessly sends the heart rate data to the smart handheld terminal through the Internet-of-things communicationmodule; the smart handheld terminal is wirelessly connected with the cluster server end through a wide area network and uploads the heart rate data to the cluster server end. The completed health datamanagement system is formed through pulse signal collection, processing, viewing, statistics and analysis, the measured heart rate data is reliable, the positive and negative errors are small compared with professional equipment in hospitals, and the stability, the interference resistance and the overload capability are higher.

An apparatus for modifying an output signal indicative of a downhole parameter that may include a carrier conveyable in a wellbore; a negative error compensator; and an output signal device. The negative error compensator may be configured to modify the output of the device to increase or decrease a characteristic of the output signal from the output signal device. Also, a method for modifying an output signal indicative of a downhole parameter that may include modifying a characteristic of an output signal produced by a output signal device in a wellbore using a negative error compensator.

The invention discloses a double voltage detecting method for a transformer drum type winding. The method comprises the following specific steps that testing voltage is changed into a 380V voltage added to one side of the transformer drum type winding, and the other side of the transformer drum type winding is connected with a 0.2-level measuring voltmeter; the transformer drum type winding obtains variable-ratio error between a first side voltage V1 obtained from one-time actual measurement of the voltage meter and a second side voltage V2 obtained after secondary actual measurement; the obtained variable-ratio error is compared with the national standard, and when the measurement error value exceeds a standard scale, if the measurement error value is larger than 0 positive error, the low-voltage turns will be lack of N turns or the high-voltage turns will be redundant with iN turns, and if the measurement error value is less than 0 negative error, the low-voltage turns will be redundant with N turns or the high-voltage turns will be lack of iN turns. By the adoption of the technical scheme, the test precision can be guaranteed, and the safety of the operation detection test can be ensured; whether the transformation ratio error is within a prescribed range or not can be obtained through comparison with the national standard, and the number of and the position of the error ofthe transformer winding turns can be found out.

A multi-stage variable gain amplifier is disclosed that utilizes overlapping gain curves to compensate for log-linear errors. Each gain stage is configured to approximate a log-linear response with a sinusoidal error term such that a portion of the curve has positive errors and a portion of the curve has negative errors. In operation, the control signal inputs for the gain stages are driven such that the gain curves overlap and positive errors in each stage are offset by negative errors in the adjacent stage. The resulting combined gain is a more log-linear response.

The invention relates to a sampling-free automatic equalizing apparatus for batteries. The apparatus comprises a main circuit, a voltage difference detection unit and a negative feedback control system, wherein the negative feedback control system comprises a positive error comparison unit and a negative error comparison unit. The invention provides the sampling-free automatic equalizing apparatus for the batteries; the voltage of the single battery does not need to be sampled by a voltage sensor; instead, the automatic equalizing of the two adjacent single batteries is realized through an analog circuit; in addition, equalizing current can be adjusted in real time, and the magnitude of the equalizing current is decreased with the decreasing of the inconsistency of the two adjacent single batteries in the equalizing process, so that the equalizing reliability is improved; by adoption of the apparatus, a voltage sampling link and a battery management system control link for the single batteries are omitted, so that the software and hardware complexity of the system is reduced, and high transportability is realized; in addition, the automatic equalizing of the two adjacent single batteries can be realized; the equalizing current can be adjusted in real time; and the equalizing reliability is improved.

First order gradient errors are canceled with no current source splitting by placing consecutive current sources symmetrically around the center of the array. Consecutive elements that correspond to small input amplitudes (mid-scale codes) make a smaller spatial jump than those correspond to larger signal amplitudes. Both linear and second order gradients are reduced by splitting each current cell into two and placing sub-elements symmetrically with respect to the center of the array to address the linear gradient effect. To address second order gradients, current element placement follows a pattern such that consecutive element pairs are chosen with one of the pair being placed with respect to the zero error contour of the second order gradient so as to have a positive error and the second of the pair being placed so as to have a negative error resulting in reduced second order error accumulation.

The invention relates to electronic ballast production on-line testing equipment, in particular to an HID (high intensity discharge) electronic ballast on-line tester which mainly tests constant power control portion in an electronic ballast DC-AC (direct current-alternating current) conversion circuit. The specific technical scheme includes that the HID electronic ballast on-line tester comprises a tester body and a test circuit arranged on the tester body and is characterized in that the test circuit comprises an oscilloscope, a programmable controller, an amplifying / comparison circuit and a dummy load, the oscilloscope is used for testing waveforms of output Ga and Gb of a circuit on a tested constant power control panel, the programmable controller comprises a program starting button, a positive error out-of-tolerance signal I01, a negative error out-of-tolerance signal I02 and an output executing relay J which are positioned at an input end, the amplifying / comparison circuit comprises a current amplifying portion and a comparator COMP, the dummy load is formed by connecting a capacitor with a resistor and used for substituting for a bulb portion on the tested constant power control panel.

The invention discloses a method for realizing a negative error frame in a CAN protocol and a device thereof. The invention determines the length of a delimiter of the negative error frame by changing condition of error data flow, carries out correction to the negative error frame or resets to an automatic machine in CAN protocol state and guarantees that the beginning time of the sending and receiving of the negative error node and all the nodes is synchronous, thus solving the fault that the negative error node is possible to be in actual off-line, improving the reliability of the CAN protocol communication and the utilization ratio of bandwidth, improving the safety of automobiles applying the CAN protocol, leads the network design tool software which is widely used currently to have areliable theoretical basis and guaranteeing the continuity of industrial production.

An information encoding method in which a common multiplier A of the input signal samples x(i) in each frame is estimated. Each sample x(i) is divided by A, the quotient is rounded to the number of significant digits, and the rounded quotient y(i) is encoded. The y(i) is multiplied by A and the product is subtracted from x(i) to obtain an error signal z(i), which is encoded. The A is encoded and the codes of y(i), z(i), and A are combined together and outputted. To estimate A, each numeric value is divided by a determined representative value. Irreducible fractions between the maximum positive error and the maximum negative error are obtained by dividing each numeric value by a correct representative value using rational approximation. Among the obtained irreducible fractions, the one having the smallest common denominator of the numeric values is obtained, which is corrected using the representative value to obtain A.