18 results about "Modulation factor" patented technology

The invention discloses an electromagnetic field simulation method under the coupling action of tidal waves and sea waves. The method comprises the steps that input parameters of the tidal waves and the sea waves are set; configuring environmental parameters and geomagnetic conditions; superposing the tidal wave velocity field and the sea wave disturbance velocity field to form a total velocity field; the tidal wave part and the sea wave part are linearly coupled in time and space to obtain a tidal wave-sea wave coupling field evolved along with a tidal phase; in the tidal wave-sea wave coupling field, electromagnetic response calculation is executed on each frequency-direction sampling point; general field synthesis and modulation effect analysis: introducing a tidal wave modulation factor to clearly reflect amplitude periodic change and spectrum reconstruction characteristics of an electromagnetic field signal along with a tidal phase; and outputting a time sequence, an amplitude spectrum and a spatial distribution result of the comprehensive electromagnetic field. According to the method, the authenticity and applicability of sea wave induction electromagnetic field simulation can be improved, and a new technical means is provided for electromagnetic noise recognition, sea wave parameter inversion and complex ocean dynamics process research.

The invention discloses a squint synthetic aperture sonar imaging method based on a dynamic slant-range space-variant model, and relates to the technical field of synthetic aperture sonar imaging, and the method comprises the steps: constructing a dynamic slant-range model; inputting the converted data into the dynamic oblique variation model to obtain a residual space-variant distance bending amount, and compensating the space-variant distance bending amount by using a first filter; according to the dynamic slope distance model, expanding the azimuth time offset into an azimuth frequency modulation factor; based on the azimuth frequency modulation factor and the azimuth time offset, designing a second filter by using an improved extended nonlinear frequency modulation scaling algorithm, and filtering the corrected data by using the second filter; and carrying out Fourier transform on the filtered data, compressing the transformed data by using a third filter, and then carrying out inverse Fourier transform to obtain a focused synthetic aperture sonar image. According to the invention, the imaging performance and stability of the imaging algorithm in a low signal-to-noise ratio and strong multipath interference environment can be improved.

The invention belongs to the technical field of multi-modal semantic feature alignment, and provides a multi-modal feature hierarchical alignment method and system based on an adaptive temperature coefficient, and the technical scheme is as follows: carrying out feature alignment based on obtained multi-modal data and a trained multi-modal feature hierarchical alignment model; wherein the training process of the multi-modal feature hierarchical alignment model comprises the following steps: respectively constructing a coding network of each modal data, and extracting the feature representation of each layer in the forward propagation process of the coding network; setting a temperature coefficient for each layer, and calculating the contrast loss between different modal features of the layer based on the temperature coefficient of each layer; calculating a specific modulation factor of each layer according to a temperature coefficient dynamic attenuation strategy sensed by a preset characteristic attribute in combination with the characteristic representation of each mode of each layer, and adjusting the temperature coefficient of the corresponding layer based on the specific modulation factor of each layer; and fusing the comparison loss of all levels to obtain a total target loss function. Therefore, accurate alignment of multi-modal features is realized.

This invention proposes a collaborative reconstruction method for multimodal images from unmanned aerial vehicles (UAVs) for all-weather perception. First, it uses an Adaptive Degradation Perceptual Block (ADPB) to perform spectral analysis on degraded visible light features to perceive the degradation type, and utilizes a learnable frequency mask to decouple features, providing a clear, high-fidelity structural prior for subsequent modal interactions. Next, it leverages the multi-scale decomposition capability of wavelet transform to decouple multi-frequency sub-band signals, and learns local contraction mappings within each sub-band through deep convolution, effectively preserving structural integrity while suppressing blur and noise. Finally, addressing the challenge of simultaneous degradation and cross-modal information fusion in multimodal images, it generates dynamic modulation factors in spatial and channel dimensions to guide adaptive complementary interaction and deep fusion of cross-modal features, strengthening the representation capabilities of global and local features. This significantly improves the reconstruction quality and multi-task generalization ability of the unified multi-task recovery model in complex scenarios.

An electric motor device includes an electric motor, an inverter, a noise reducer, a control unit, and a metal housing. The controller controls switching elements using a PWM signal generated from a voltage instruction value and a carrier frequency. The control unit includes a calculation portion that calculates a modulation factor of the voltage instruction value and a setting portion that sets the carrier frequency in accordance with the modulation factor. The carrier frequency set by the setting portion is higher when the modulation factor is less than a predetermined modulation factor than when the modulation factor is greater than or equal to the predetermined modulation factor.

A motor control device for controlling the driving of an AC motor connected to an inverter having a plurality of switching elements comprises: a dq-axis voltage command generation unit for generating a dq-axis voltage command value for the AC motor for each predetermined calculation period; a modulation rate command generation unit for generating one or more modulation rate command values on the basis of the dq-axis voltage command values, the phase of the AC motor, and the number of switches the switching element performs within the calculation period; and a duty command generation unit for generating a duty command value, for generating a gate signal for controlling the operation of the switching element, on the basis of the modulation rate command value. The modulation rate command generation unit switches the generation method for the modulation rate command value according to the number of switches when the number of switches is equal to or greater than a predetermined number. As a result, it is possible to achieve a voltage output closer to the command value while suppressing an increase in calculation load.

The embodiment of the invention discloses a video processing method and device, a medium and equipment. The method comprises the following steps: acquiring a current decoding feature corresponding to a current to-be-reconstructed frame of a target video and a historical decoding feature corresponding to a historical reconstructed frame of the target video; determining a current quality modulation factor reflecting the quality of the historical decoding feature according to the current decoding feature and the historical decoding feature; acquiring a historical target hidden state corresponding to the historical reconstruction frame, and performing quality modulation on the historical target hidden state through the current quality modulation factor to obtain a historical modulated hidden state; according to the method, the current target hidden state corresponding to the current to-be-reconstructed frame is determined according to the current decoding feature and the historical post-modulation hidden state, and the current to-be-reconstructed frame is generated according to the current target hidden state, so that the forward contribution of the high-quality historical decoding feature can be enhanced, the invalid interference of the low-quality historical decoding feature can be suppressed, and finally the decoding performance is improved.

A motor control device according to one aspect of the present disclosure comprises a reception unit, a detection unit, and a control unit. The reception unit receives a switching of a travel mode from a first mode in which overmodulation control is performed to a second mode in which non-overmodulation control is performed. The detection unit detects the speed of the motor. The control unit decreases the modulation rate of the overmodulation control so as to decelerate the speed at a predetermined reduction rate when the switching is received by the reception unit in a state in which overmodulation control is performed while the speed exceeds an upper limit speed of the second mode in the first mode, and performs non-overmodulation control by setting the modulation factor to 1 or less and starts flux weakening control when the speed is equal to or less than the upper limit speed of the second mode.

The invention provides a calibration method, a measurement method and a device for 3D structured light multi-range measurement, and relates to the technical field of AOI industrial visual inspection, and the method comprises the steps: determining a frequency doubling sine stripe for calibration and a Gray code corresponding to the frequency doubling sine stripe according to camera information and DLP information; acquiring a fringe image obtained by projecting the frequency-doubled sine fringes when the DLP moves in a preset range in the vertical direction, and performing multi-step phase shift calculation processing based on the fringe image to obtain a relative phase of the fringe image; determining a target relative phase from the frequency-doubled sine fringes based on the modulation degree through a preset frequency-doubled algorithm, and combining the target relative phase with the Gray code to determine an absolute phase; and performing phase conversion processing on the absolute phase to obtain a target phase, determining a non-parametric fitting parameter of the preset phase height contrast model through the target phase, and storing the non-parametric fitting parameter as a calibration file. According to the invention, the measurement efficiency and accuracy can be significantly improved.

The invention relates to the technical field of power system monitoring and load management, and particularly discloses a non-intrusive target load identification method based on cloud-end scene migration, which introduces a physical consistency data enhancement mechanism at a cloud end, and simulates a modulation effect of voltage sag on a background load by calculating a dynamic coupling modulation factor. And a high-fidelity enhanced triple data set is generated, so that the problem that the coupling characteristic of a circuit is neglected in traditional data enhancement is solved. On this basis, a brand invariance feature extractor is trained by using hierarchical metric learning, and class essential features are learned by optimizing an inter-class and inter-brand distance forced model. Finally, the feature extractor and the classification head are assembled into a generalization model to be deployed to the end side, reasoning is carried out by freezing parameters of the feature extractor, the cross-brand generalization problem caused by physical heterogeneity of the electric appliance is effectively solved, and accurate identification of the electric appliance with no brand at the end side is achieved.

This invention discloses a method for predicting the unsteady acoustic quality of electric drives based on intelligent sensors. Specifically, it includes: acquiring noise signals and extracting sound pressure level, loudness, sharpness, speech intelligibility, and priority speech interference level to form a feature vector sequence and dataset; constructing a neural network to determine nodes and form network parameter vectors; establishing an information entropy budget allocation model, mapping each feature to an information budget variable vector, setting constraints to construct an objective function containing budget constraint terms and prediction error terms, and embedding the budget variables into the feature representation; constructing a particle swarm, uniformly encoding the information budget vector and network parameter vector into particle state vectors and then initializing them; introducing an information budget modulation factor during iteration to weight the velocity update component, achieving co-evolution of the two; updating the optimal particle based on fitness, outputting the optimal information budget and network parameters to construct a prediction model; and inputting the feature vector sequence to output the prediction result. This invention assigns independent budgets to each acoustic feature, dynamically co-optimizes, and achieves high accuracy and fast response.

Method and system for accurate measurement of blood oxygen saturation levels of a subject by manipulating respective DC components of first and second detection electrical signals associated with a detected light response of a body tissue being illuminated by a first and second wavelengths, and determining a respective correction factor for adjusting a modulation factor derived from the AC and DC components of first and second manipulated electrical signals (referred throughout the application as diminished electrical signals).

The application discloses a high-precision space target orbit prediction method based on numerical fitting, and relates to the technical field of space target orbit prediction.The frequency domain decomposition of the application is combined with a harmonic compensation model to completely eliminate seasonal system drift caused by the gravitation of the sun and the moon; high-frequency components are stripped of attitude noise through light pressure parameterization and smooth difference to avoid the transmission of perturbation coupling error; a harmonic order adaptive mechanism captures the main disturbance term with the least amount of calculation; least square batch estimation realizes one-time solution of coefficients; template library preloading technology effectively compresses the time consumption of new target initialization; in addition, the application introduces a geomagnetic modulation factor to dynamically suppress compensation distortion during magnetic storms, and a sliding window mechanism guarantees real-time updating of coefficients, and stable prediction is still maintained under extreme events; photometric curve inversion is adopted instead of attitude measurement to make non-cooperative target prediction possible; tangential component directional compensation ensures the stability of the orbit plane and reduces satellite position fuel consumption.

This invention relates to an efficient fine-tuning method for translation parameters of large model based on hierarchical activation difference perception, belonging to the field of natural language processing and large model fine-tuning technology. To address the problems of pre-training bias dominance and hierarchical functional imbalance in existing large language models (LLMs) for low-resource machine translation tasks, this invention includes: firstly, accurately locating model modules sensitive to the machine translation task using cumulative activation intensity and cross-layer activation variance; then, dividing the model hierarchy into task-adaptive layers and legacy inert layers based on activation difference indices; introducing learnable latent gating parameters to generate hierarchical-specific modulation factors; and during efficient parameter fine-tuning, using the modulation factors to amplify the signal in the task-adaptive layer while suppressing the legacy inert layers. This invention effectively improves translation performance from Chinese to low-resource languages ​​such as Lao and Khmer by structurally differentiating the optimization direction.

A field-weakening controller determines a d-axis current command correction value in field-weakening control according to a modulation factor that correlates with the ratio of the voltage amplitude applied to the inverter to the power supply voltage. A d-axis current final value calculator outputs a d-axis current command final value calculated using a d-axis current command value and a d-axis current command correction value. A Current controller performs feedback control of the d-axis current command final value and the q-axis current command value. A d-axis current adjuster adjusts the d-axis current command final value based on the q-axis current command value so as to suppress an influence of the frequency component of the q-axis current command value in a specific frequency band for which at least a lower limit frequency is determined.

The application discloses a modulation method of minimum DC side capacitor current effective value of a three-phase inverter in a full range, comprising the following steps: S1, determining a three-vector combination corresponding to a predetermined modulation method; grouping according to an intersection of a synthesizable area of the three-vector combination on a space vector alpha-beta plane and a first sector; S2, determining a predetermined modulation method of minimum DC side capacitor current effective value of each group based on minimum DC side instantaneous inverter current effective value in each group; S3, globally comparing the determined predetermined modulation method, and obtaining the modulation method of minimum DC side capacitor current effective value of the three-phase inverter in the full range based on minimum DC side instantaneous inverter current effective value under different phase angles, modulation factors and power factor angles. The modulation method can realize minimum DC capacitor current effective value under the full range power factor angle and modulation factor on the basis of traversing all possible three-vector combination modulation methods.

The invention relates to a magnetic resonance imaging method and an adaptive magnetic resonance imaging method. The method comprises the following steps: respectively acquiring effective coverage ranges of a plurality of coil units in a magnetic resonance imaging range; obtaining a modulation factor corresponding to each coil unit according to the effective coverage range and the receiving sensitivity of the coil unit; and performing channel merging reconstruction on the magnetic resonance signals received by the plurality of coil units according to the modulation factor to obtain a magnetic resonance image. By adopting the method, the effects of high signal-to-noise ratio and low peripheral artifacts can be achieved during magnetic resonance imaging.

The invention provides a single-phase cascade type optical storage hybrid system power coordination control method and device, and relates to the technical field of photovoltaic system power control, and the method comprises the steps: according to the reference power and the DC side voltage of a photovoltaic module unit, and the reference amplitude of the output voltage of a cascade converter, calculating the power of a single-phase cascade type optical storage hybrid system; determining a modulation factor of the cascade converter when the power of the battery unit is zero; judging the size of the modulation factor and a target threshold value, judging the size of the state of charge of the battery unit and the size of the state of charge threshold value range, and determining a comparison result; and based on the comparison result, determining a corresponding power coordination control strategy to control the operation mode of the single-phase cascaded optical storage hybrid system, and the method solves the technical problems of complex system structure and high over-modulation risk in the prior art. The technical effects of simplifying the hardware architecture, maintaining the system performance and reducing the over-modulation risk are achieved.