155 results about "Linear prediction" patented technology

The invention discloses a photovoltaic prediction method and system based on a bidirectional Mama structure and combined with a sky image, and the method comprises the steps: obtaining a continuous sky image sequence, extracting dynamic visual features and static visual features, and obtaining image sequence features through the integration of a convolution layer, space adaptive adjustment and global average pooling; performing time coding processing and linear mapping on the historical photovoltaic power sequence to obtain photovoltaic sequence features consistent with image sequence feature dimensions; a cross-source alignment integration mechanism is adopted for processing, a bidirectional Mamba structure is constructed, and a final image data source hiding state and a final photovoltaic data source hiding state after long and short time sequence dependency analysis are output; a cross-source depth feature correlation enhanced gating integration mechanism is adopted for processing, and final integration features are obtained through self-adaptive gating integration; and outputting a future photovoltaic power prediction value through the linear prediction layer. According to the method, the accuracy, the robustness and the calculation efficiency of ultra-short-term photovoltaic prediction can be remarkably improved.

The invention belongs to the technical field of deep learning and time sequence analysis, and particularly relates to a multivariable time sequence prediction method and device based on spatial-temporal feature fusion. The method comprises the following steps: acquiring multivariable traffic time series data, and processing the traffic time series data; dividing the processed traffic time sequence data into overlappable patches, and generating a patch embedding sequence through linear mapping; applying bimodal time attention to the patch embedding sequence to obtain fused attention features; based on the learnable node embedding matrix, generating a time-varying adjacency matrix through dynamic graph construction, executing graph convolution to obtain a time domain graph propagation result, executing fast Fourier transform, multiplying by a learnable scaling factor and then performing inverse transformation to obtain an inverse transformation time frequency result; adding the time domain graph propagation result and the inverse transformation time frequency result to obtain a final spatio-temporal characteristic; and flattening the final spatio-temporal characteristics, and generating predicted values of the variables in a future prediction window through a linear prediction head.

The invention relates to the field of biological medicine, and discloses a method and system for analyzing cough sound by using disease characteristics to diagnose respiratory diseases, and the method comprises the steps: deploying a six-microphone annular array to achieve the precise positioning and triggering of a sound source; self-adaptive spectral subtraction and Wiener filtering cascade are adopted to enhance the audio; segmenting a cough segment based on energy envelope; fusing the Mel-cepstrum, the linear prediction residual error, the harmonic energy ratio and the transient zero-crossing rate to construct a pathological feature matrix; extracting local, medium-range and global time sequence features through a three-branch parallel convolutional network; inputting a disease specific classifier to discriminate asthma, pneumonia and laryngitis respectively, and applying a feature decoupling regular term to improve interpretability. The system correspondingly realizes the modularized processing flow. According to the method, the cough sound collection quality and the disease subtype recognition accuracy in a complex environment are improved, meanwhile, the thermodynamic diagram is output to assist clinical decision making, and the diagnosis credibility and practicability are enhanced.

Disclosed in the present invention are a data-driven linear MPC-based assessment method and apparatus for a frequency regulation capability of a wind farm, and a control method and apparatus. The assessment method comprises the steps of: offline acquiring historical operation data sets of wind turbines, and training a linear predictive control model, wherein the linear predictive control model is obtained by applying a dimensionality-increasing transformation process to a wind farm frequency modulation dynamic nonlinear model on the basis of a Koopman operator theory; acquiring real-time operation data of the wind turbines, and using the acquired real-time operation data to solve for a linear MPC optimization model, to obtain droop coefficients of the wind farm, wherein the linear MPC optimization model is constructed on the basis of the trained linear predictive control model and by taking the maximum droop coefficient of the wind farm as an objective and setting a safety rotational speed constraint condition; and assessing a frequency regulation capability of the wind farm on the basis of the obtained droop coefficients of all the wind turbines in the wind farm. The present invention has the advantages of ease of implementation, high assessment efficiency and accuracy, and strong scalability, and the like.

The invention discloses a semi-supervised segmentation method combining double segmentation head frequency decoupling learning and entropy change pseudo label screening, is applied to the field of image processing, and aims to solve the problems that in an existing semi-supervised semantic segmentation method, pseudo label generation quality is difficult to guarantee, a model is single in enhanced disturbance response, and the efficiency is low. The problems of high false label error and unstable training caused by difficulty in distinguishing global structure change and local detail change are solved; the student model adopts a double-segmentation-head structure and is composed of a prototype prediction head and a linear prediction head, and function division is realized through a frequency domain decomposition mechanism: the prototype head receives low-frequency component characteristics of an encoder, the linear head receives high-frequency component characteristics of the encoder, and for a label-free sample, after random enhancement is applied, the original prediction head and the linear prediction head are divided into two segments; and respectively predicting category distribution of samples before and after enhancement by two segmentation heads of the teacher model, and calculating a change ratio of average prediction entropy. And based on the entropy change ratio before and after enhancement, synthesizing the reaction of the prototype head and the linear head, and jointly evaluating the prediction stability after enhancement.

An image blur correction device acquires a blur detection signal, and separates it into a high frequency band signal component and a low frequency band signal component using an HPF and an LPF. A high frequency side prediction processing unit acquires an output of the HPF, and a low frequency side prediction processing unit acquires an output of the LPF via a down-sampler. A prediction processing unit updates a filter coefficient of a prediction filter using an adaptive algorithm, and performs linear prediction on the blur detection signal. An adder adds an output of the high frequency side prediction processing unit to an output of the low frequency side prediction processing unit via a up-sampler, and outputs a superimposed output signal. Image blur correction control is performed on the basis of a superimposed output signal.

The invention discloses a resource quota adjustment method, device and equipment and a readable storage medium, and is applied to the technical field of distributed storage, and the method comprises the steps: determining current traffic data corresponding to a current service request; performing linear prediction based on the current traffic data to obtain initial predicted traffic, and performing nonlinear residual correction on the predicted traffic to obtain target predicted traffic; and configuring the resource quota based on the target predicted traffic to obtain a target pre-application resource quota. According to the invention, prediction is carried out by using the prediction model based on the current traffic data to obtain the initial predicted traffic, nonlinear residual correction is carried out on the predicted traffic to obtain the target predicted traffic, and the target predicted traffic is obtained through prediction and error compensation, so that the predicted traffic is more accurate, and the prediction efficiency is improved. Therefore, the target pre-applied resource quota obtained by performing pre-application on the resource quota based on the target predicted traffic subsequently is more accurate, the burst traffic resistance of the distributed storage system is improved, and the subsequent service quality is improved.

The present invention discloses a continuous speech formant extraction method based on peak selection, comprising: performing a preprocessing operation on a single frame of input speech; using a linear prediction method to preliminarily estimate the peak value in the spectral envelope of the speech frame; establishing a reference point and a formant trough, and then using a peak selection method to establish a mapping relationship between the peak value and the reference point; using the mapping relationship between the peak value and the reference point and the formant trough to determine the formant of the speech frame; and performing formant estimation on the continuous speech: dividing the continuous speech into frames according to different frame numbers, using the above algorithm to loop 100 times to obtain the formant parameters under different frame number tests, averaging the results after 100 loops, and obtaining the final result after smoothing. The method of the present invention can eliminate the influence of merged peaks and false peaks, and has a fast convergence speed and strong robustness.

The invention provides a cloud sensing shared information fusion method for a vehicle and road cloud cooperative system. The method comprises the steps that the vehicle and road cloud cooperative system comprises a cloud, a vehicle end and a road end; the cloud end obtains environment sensing information from other terminals through network communication; a space-time two-dimensional calibration system is constructed, and the space-time two-dimensional calibration system is adopted to calibrate the environmental perception information; dividing a unified global ID for each target by adopting a global target feature matching method, and performing redundancy removal processing on the calibrated environmental perception information by adopting a confidence evaluation comprehensive redundancy removal mechanism; performing pre-estimation compensation on the environment perception information after redundancy removal by adopting a model based on state estimation and kinematics; transmitting the compensated information to a corresponding terminal by adopting a sensing information routing sharing strategy based on the global ID to complete information sharing; according to the method, a resampling optimization strategy and a collaborative prediction correction algorithm are adopted, compared with a traditional linear prediction compensation method, situation information real-time performance and prediction accuracy are remarkably improved, and timeliness and reliability of follow-up decisions are effectively guaranteed.

The present application relates to a kind of place recommendation method based on hypergraph neural network and diffusion model, interest point recommendation model is constructed, interest point recommendation model successively includes local trajectory flow hypergraph module, space-time feature coding, multi-dimensional feature fusion network, global hypergraph representation learning module, feature optimization module, aggregation layer, frequency domain learning layer and linear prediction layer.Analysis of the long trajectory of user, and it is divided into space-time region, constructs three global hypergraphs, aims at comprehensively capturing the overall behavior pattern of user.In order to better optimize trajectory intention representation, propose feature optimization module based on improved diffusion model.Introduce multi-dimensional global representation to ensure a more stable and controllable reverse process, and use feature normalization and improved Transform network, enhanced diffusion model is more suitable for recommendation system.Loss function is designed to train interest point recommendation model, and the interest point recommendation model trained is used to recommend next interest point for new user.

The invention relates to an image forgery positioning detection method based on collaborative difference optimization and multi-modal perception, and belongs to the technical field of computer vision and multimedia evidence obtaining. The method comprises the following steps: obtaining an RGB image and carrying out random data enhancement; extracting a noise map by using a noise extractor; the RGB image and the noise graph generate multi-scale features through a pre-trained cross-modal information encoder; the multi-scale features are input into an exception encoder, the exception encoder comprises a multi-layer perceptron, a linear fusion block and a linear prediction block, multi-layer mapping features are obtained through processing of the multi-layer perceptron, and an initial prediction map is output; inputting the original RGB image into a ViT model to extract global features, and inputting the global features and the multilayer mapping features into a cross-scale feature enhancement module for interactive optimization to obtain a refined positioning map; fusing the initial prediction map and the refined positioning map to obtain a final prediction map; and iterative optimization is carried out through the loss function. According to the invention, the detection sensitivity of tampered areas with different sizes can be improved.

The invention relates to the technical field of module control, and discloses a multi-directional operating servo sliding table module control method and device, and the method comprises the steps: carrying out the motion state collection of an X axis, a Y axis and a Z axis of a servo sliding table module, and obtaining multi-dimensional inter-axis coupling data containing a lead screw lead error, a guide rail friction characteristic and motor torque fluctuation; constructing a tensor compensation model for describing XYZ three-axis coupling strength distribution based on the multi-dimensional inter-axis coupling data; carrying out CP decomposition dimension reduction based on the tensor compensation model to obtain dimension reduction coupling relation data; executing multi-axis synchronous linear prediction according to the dimension reduction coupling relation data to obtain multi-axis coordination trajectory planning data; according to the method, nonlinear control law calculation is performed based on multi-axis coordination trajectory planning data to obtain an optimal control instruction sequence, so that redundant information and noise components in original coupling data are effectively removed, the robustness of a control system to environmental interference and parameter change is improved, and stable control performance under different working conditions is ensured.

The invention relates to the field of data processing, and provides a data center short-term power load prediction method, system and device and a medium, and the method comprises the steps: obtaining a to-be-predicted time period, and obtaining time series data according to the to-be-predicted time period; inputting the time sequence data into a preset SARIMA and BiLSTM composite model to obtain a linear predicted value and a linear residual sequence; normalizing the linear residual sequence to obtain a data set, and generating a nonlinear predicted value and a nonlinear residual sequence according to the data set; and dynamically combining the linear predicted value and the nonlinear predicted value into a load predicted value, and outputting the load predicted value and a nonlinear residual sequence. According to the method, the SARIMA model and the BiLSTM model are combined, the accuracy of data center load prediction can be improved, and the method is suitable for complex data center load prediction tasks.

This application relates to a real-time power supply and demand forecasting method and system based on a cloud-native architecture. The method includes: deploying a data access service in a cloud-native cluster using containerized microservices to receive real-time supply and demand data streams and historical time-series data from a power trading system; writing the data streams to distributed storage and pushing them to the forecasting pipeline via a message queue; performing timestamp alignment, missing data handling, normalization, and smoothing / denoising on the supply and demand data by a preprocessing service to obtain a low-noise supply and demand sequence; updating model parameters in a rolling window by an ARIMA forecasting service and outputting linear forecast values ​​as the first forecast result; calculating the forecast residuals based on the first forecast result and the actual observations, constructing residual time-series samples, and outputting residual forecast values ​​by an LSTM forecasting service; and superimposing the first forecast result and the residual forecast values ​​by a fusion service to obtain the real-time supply and demand forecast result and publishing it to the real-time trading business interface.

The application discloses a linear prediction method based on orthogonal compensation, relates to the technical field of data processing and machine learning, and calculates pixel offset between two adjacent images after initial linear compensation and corresponding offset data through an image matching algorithm; offset data is used as a difference value between an observation value and a prediction value, and a reference value of reference data is used as a state quantity; if the offset is normal, an interactive multiple model based on Bayesian theory is used to fuse real-time data, a prediction value of the interactive multiple model based on Bayesian theory is used as input, linear data error of a system linear data error model is used as output, and the system linear data error model is subjected to model training; if the offset is abnormal, a compensation data information is used to calculate a reference value, so that modified data information is obtained, and when an abnormal type that has not been sufficiently learned is encountered, the recognition capability can be greatly reduced, and even a missed detection or a false detection can occur.

The application provides an audio signal alignment method and device, a medium and a terminal, and relates to the technical field of audio processing. The method comprises the following steps: calculating a short-time frequency domain representation of an audio signal, wherein the audio signal comprises a first audio signal and a second audio signal; calculating a frequency response of a time-varying linear prediction error filter according to the short-time frequency domain representation of the audio signal, and calculating a short-time spectrum of a linear prediction error; extracting a low-frequency coefficient from the short-time spectrum and calculating a short-time spectrum of a corresponding linear prediction error envelope; calculating a short-time cross-power spectrum of the linear prediction error envelope corresponding to the audio signal according to the short-time spectrum of the linear prediction error envelope; calculating a frame-based time delay between the two audio signals according to the short-time cross-power spectrum of the linear prediction error envelope; and performing a variable-speed constant-pitch processing on the first audio signal or the second audio signal according to the calculation result of the frame-based time delay, so as to obtain an aligned audio signal. The present scheme can improve the accuracy of time delay estimation.

The application discloses a bone conduction speech conversion method based on spectral envelope mapping, comprising the following steps: pre-processing and linear prediction analysis of the bone conduction speech signal, and calculating the LP filter coefficient; mapping the LSF coefficient of the air conduction speech signal corresponding to the bone conduction speech signal by using the trained neural network; controlling the minimum error of the original bone conduction speech signal and the synthesized air conduction speech signal according to the timbre weighting characteristic of the bone conduction speech characteristic; estimating the integer pitch of the bone conduction speech signal through the timbre weighting filter, obtaining the reference signal by passing the linear prediction residual signal through the timbre weighting filter, estimating the fractional pitch to obtain the adaptive codebook vector; obtaining the new reference signal by subtracting the adaptive codebook vector from the reference signal, searching for the optimal excitation in the fixed codebook; synthesizing the air conduction speech signal by combining the optimal excitation and the LP filter of the air conduction speech signal, and correcting the air conduction speech signal.

The present invention discloses a fine-grained collaborative prediction method for electric loads, the method comprising: obtaining multivariate time series input data; slicing the input data to obtain multiple data patches; processing the data patches through dynamic weight variables and a short-term time series feature extraction module to obtain enhanced feature representations; inputting the enhanced feature representations into an encoder to generate coding features; inputting the data patches into a linear prediction layer to obtain a first prediction component; generating a second prediction component based on the enhanced feature representation and the coding features; and obtaining a collaborative prediction result for electric loads based on the fusion of the first prediction component and the second prediction component. The present invention can effectively collaboratively model the complex dependencies of time and variable dimensions through AVSTFE, channel-independent Transformer, and dual-path prediction heads, overcoming the limitation of existing methods that are difficult to balance the two, thereby significantly improving the prediction accuracy of multivariate time series, especially fine-grained electric loads.

An audio decoder is disclosed. In one example, the audio decoder is for providing a decoded audio information on the basis of an encoded audio information includes a linear-prediction-domain decoder configured to provide a first decoded audio information on the basis of an audio frame encoded in a linear prediction domain, a frequency domain decoder configured to provide a second decoded audio information on the basis of an audio frame encoded in a frequency domain, and a transition processor. The transition processor is configured to obtain a zero-input-response of a linear predictive filtering, wherein an initial state of the linear predictive filtering is defined depending on the first decoded audio information and the second decoded audio information, and modify the second decoded audio information depending on the zero-input-response, to obtain a smooth transition between the first and the modified second decoded audio information.

The application provides an artificial intelligence-based carbon analysis and diagnosis method and system, which comprises the following steps: identifying and reconstructing abnormal data from original energy consumption time series data; dividing the standardized energy consumption data set into time blocks in time sequence, mapping the potential representation of the time block into a discrete token sequence through a time convolution network encoder; inputting the discrete token sequence into a preset BERT model for self-supervised pre-training; fine-tuning the linear prediction layer using labeled historical energy consumption data, outputting energy consumption or carbon emission prediction values, and performing energy efficiency analysis and diagnosis based on the energy consumption or carbon emission prediction values. The application can effectively process multi-source heterogeneous device energy consumption data, improve the prediction accuracy and generalization ability in cross-device type scenarios, and provide reliable diagnosis basis for device-level carbon management.

The application discloses a voltage stability margin calculation method considering a control mode of a photovoltaic power station, and adds modeling consideration of a reactive power-voltage control mode of the photovoltaic power station in voltage stability margin calculation. In a continuous power flow calculation process, parameterized power flow equations are obtained by extending node voltage with the fastest voltage drop. Meanwhile, in a prediction link in the continuous power flow calculation process, a hybrid prediction method is adopted. In the case that voltage margin is large, nonlinear prediction is adopted to accelerate the calculation speed. In the case that voltage stability limit is approached, linear prediction is adopted. The voltage stability margin calculation method fully considers the influence of the control mode of the large-scale photovoltaic power station on voltage stability, and improves the calculation efficiency and the calculation precision.

Various embodiments provide a method and an apparatus for determining a weighting factor during stereo signal encoding. In those embodiments, a parameter value corresponding to the encoding mode of the to-be-encoded signal is determining based on an encoding mode of a to-be-encoded signal in a stereo signal and a correspondence between an encoding mode and a parameter value. Based on the determined parameter value and an energy spectrum of a linear prediction filter corresponding to an original line spectral frequency parameter of the to-be-encoded signal is a weighting factor for calculating a distance between the original line spectral frequency parameter and a target original line spectral frequency parameter is calculated.

An audio decoder is disclosed. In one example, the audio decoder is for providing a decoded audio information on the basis of an encoded audio information includes a linear-prediction-domain decoder configured to provide a first decoded audio information on the basis of an audio frame encoded in a linear prediction domain, a frequency domain decoder configured to provide a second decoded audio information on the basis of an audio frame encoded in a frequency domain, and a transition processor. The transition processor is configured to obtain a zero-input-response of a linear predictive filtering, wherein an initial state of the linear predictive filtering is defined depending on the first decoded audio information and the second decoded audio information, and modify the second decoded audio information depending on the zero-input-response, to obtain a smooth transition between the first and the modified second decoded audio information.

Encoder for coding an audio signal comprising a band-limited signal portion and an extended-band signal portion, the encoder comprising: a baseband encoder configured to encode the band-limited signal portion, wherein the baseband encoder comprises at least one learnable layer; and a bandwidth extension encoder comprising a linear prediction entity for performing linear prediction on the extended band signal portion.

The invention discloses a speckle noise detection and removal method based on sound wave-laser cross-medium communication, and the method comprises the steps: analyzing and recognizing a noise position through a sliding window based on a double-feature joint detection mechanism of a kurtosis ratio and a Teager energy operator; adaptive boundary expansion is carried out on each detected noise position, and adjacent noise positions are combined to form continuous noise segments; processing the noise segments in parallel by adopting two methods of bidirectional linear prediction and empirical mode decomposition, and respectively generating prediction signals of the noise segments; according to the local signal-to-noise ratio and boundary continuity of the prediction signals, fusion weights of the two prediction signals are calculated in a self-adaptive mode, the two prediction signals are subjected to weighted fusion, and a fusion signal is obtained; endpoint continuity matching is carried out on the fusion signal through linear transformation, smooth connection of the repair section and the original signal is ensured, and speckle noise removal is completed. According to the method, speckle noise caused by dynamic water surface disturbance can be detected and effectively suppressed, and the integrity and definition of signal transmission are remarkably improved.

The invention discloses a medical image partial multi-label classification method and a related device, and the method comprises the steps: obtaining a training set, and determining the credible label distribution of each training sample in the training set based on a K-nearest neighbor label attention mechanism; training a preset multi-label classifier through the training set and the credible label distribution to obtain a target multi-label classifier; obtaining a to-be-predicted medical image, and inputting the to-be-predicted medical image into the target multi-label classifier; performing label semantic extraction according to the to-be-predicted medical image through an encoder to obtain a label semantic matrix; performing label interaction enhancement operation according to the label semantic matrix through a decoder to obtain label semantic embedding; and linear prediction is carried out through a linear prediction layer according to label semantic embedding, and prediction label distribution of the to-be-predicted medical image is generated. According to the method, the relationship between the medical image features and the label semantics and the label interaction can be captured, so that the label classification precision in the multi-label classification task of the medical image is effectively improved.

An intra prediction method using a cross component linear prediction mode (CCLM) includes determining a luma block corresponding to a current chroma block, obtaining luma reference pixels of the luma block according to determining L available chroma reference pixels of the current chroma block, wherein the obtained luma reference pixels of the luma block are down-sampled luma reference pixels, calculating linear model coefficients according to the luma reference pixels and chroma reference pixels corresponding to the luma reference pixels, and obtaining a prediction value of the current chroma block according to the linear model coefficients and values of a down-sampled luma block of the luma block.

A hybrid learning method for intracranial electroencephalogram (EEG) signal processing based on contrastive learning and mask reconstruction is proposed. After acquiring EEG signals offline and constructing a training set, an IntraBraM network is built, comprising a convolutional network block encoder, a Transformer encoder, a decoder, and a linear prediction layer. Following contrastive and mask reconstruction training, the trained IntraBraM is used for real-time EEG signal classification online. This invention addresses both the pre-training paradigm and model structure design, creating a pre-training paradigm and model structure adapted to the intracranial EEG modality. This enhances the model's performance on downstream tasks within the intracranial EEG modality, provides robustness to the spatial coordinates of adversarial electrodes, and improves generalization ability across subject settings.

The present application relates to the field of ocean engineering and simulation modeling technology, especially to a kind of underwater vehicle digital twin synchronization method and system based on predictive compensation, the method comprises the original attitude data of underwater vehicle is acquired, dead zone threshold is dynamically adjusted based on the motion energy of vehicle, and the original attitude data is preprocessed according to dead zone threshold, the received data packet is mapped to the original position by time and space sliding window cache pool according to generation time stamp, and standard state frame is generated after missing data is filled;Standard state frame is pushed to digital twin visualization terminal;Visual terminal carries out predictive compensation rendering, completes digital twin synchronization, including combining hydrodynamic damping model and weighted fusion algorithm, drive virtual model smooth motion, linear prediction interpolation driving algorithm is used in digital twin visualization terminal in the present application, ensure the continuous motion trajectory of underwater vehicle virtual model, significantly improve the reality and operability of shore-based monitoring.

The invention relates to the technical field of manhole cover monitoring, in particular to a manhole cover comprehensive environment monitoring system based on the Internet of Things, which is characterized in that a manhole cover displacement value is calculated point by point in a time window and is subjected to weighted stacking with a preorder trend curve to form a fusion sequence, and then the fusion sequence is input into a long short-term memory network to complete nonlinear prediction; an output result is compared with a threshold value interval section by section, it is ensured that abnormity judgment is achieved in a continuous fragment, through combination of time sequence prediction and nonlinear learning, unified modeling is achieved on displacement trend change and environment trend, the continuity and accuracy of abnormity judgment are improved, the displacement and the inclination angle of the well lid are synchronously judged in a trend set, and the safety of the well lid is improved. According to the method, the initial signal is triggered, the signal and the gas sudden change difference value are accumulated, features are extracted through the convolutional neural network, matching is completed, and the high-risk signal is output, so that sensitive capture of sudden abnormity is enhanced, the probability of missing report is reduced, and the response timeliness under the sudden situation is improved.