29results about How to "Reduce correlation" patented technology

The invention discloses a photovoltaic cleaning robot based on automatic path planning and an operation method thereof, relates to the technical field of robots, and solves the problem of cleaning blind spots in an existing cleaning robot and the endless loop risk of a path planning algorithm. According to the invention, the surface information of the photovoltaic panel is accurately captured through the multispectral acquisition module, a high-resolution and high-definition comprehensive image is generated through the data fusion processing module, then dirt is accurately identified on different scales by using a multi-scale pyramid structure and an adaptive feature fusion mechanism, and an attention mechanism and environmental perception data are combined, so that the real-time performance of the system is improved. The recognition threshold is dynamically adjusted, and misrecognition and missing detection are reduced; and finally, enabling the robot to autonomously learn in a complex environment and optimize an obstacle avoidance strategy through a dynamic obstacle avoidance algorithm based on deep reinforcement learning. The intelligent level and the cleaning effect of an existing photovoltaic cleaning robot are greatly improved.

The invention relates to the technical field of channel estimation, and discloses an efficient OTFS sparse channel estimation method based on fast sparse Bayesian, which comprises the following steps: S1, forming a 2D data block in a time delay Doppler domain, and dividing the time delay Doppler domain; s2, acquiring a receiving signal of a time delay Doppler domain; s3, constructing a sparse sensing matrix, and optimizing the sensing matrix by using an adaptive particle swarm optimization algorithm; s4, setting a judgment condition for successful channel recovery; s5, channel posterior distribution and hyper-parameters are updated by adopting rapid sparse Bayesian based on Laplacian prior; and S6, judging whether iteration is continued or not according to a set judgment condition. According to the method, the sensing matrix is optimized by using the adaptive particle swarm optimization algorithm, the system performance is improved, the posterior distribution and the hyper-parameter of the channel are updated by using the fast sparse Bayesian method, the calculation complexity is reduced, and the calculation resources are saved.

The invention relates to the technical field of electronic component data processing, in particular to a database construction method and system based on electronic components. The method comprises the following steps: carrying out component information acquisition through an electronic component transaction platform to obtain an original component data set; performing component relation mapping construction according to the original component data set to obtain a component association graph; performing multi-dimensional attribute extraction on the component association map to obtain component attribute vectors; performing component similarity calculation and clustering analysis according to the component attribute vector to obtain a component cluster; and extracting pin information according to the component cluster to obtain a cluster pin data set. According to the method, multi-dimensional condition screening, sorting and result explanation are realized, an accurate, reliable and easily understood component replacement recommendation scheme is provided for a user, the user is helped to quickly find a proper component, and the electronic design efficiency is improved.

The invention discloses a method for predicting the moisture of an outlet material by using a time convolutional neural network, and the method comprises the steps: collecting the historical process data of a process manufacturing production line, and constructing a multivariate time series data set; preprocessing the multivariate time series data set to obtain preprocessed data, and dividing the preprocessed data into a training set and a test set according to an acquisition cycle sequence; introducing the time convolutional neural network into a dual attention mechanism to build a DA-TCN prediction model; training the DA-TCN model by using the training set to obtain a trained DA-TCN prediction model; and verifying the test set by using the trained prediction model DA-TCN to obtain a DA-TCN prediction model passing the verification for predicting the moisture of the outlet material. According to the method, the DA-TCN prediction model used for predicting the outlet material moisture is established by introducing the time convolutional neural network into a dual attention mechanism, and the method is further used for predicting the outlet material moisture in the sheet cut tobacco drying process.

This application belongs to the field of underwater acoustic array signal processing technology. This application provides a tensor-enhanced azimuth estimation method suitable for high-order acoustic field sensor arrays. The embodiments of this disclosure first construct a received tensor from the received data of the high-order acoustic field sensor array; then, multilinear low-rank modeling is performed on the received tensor, and the optimal first subspace factor matrix in the array dimension and the optimal second subspace factor matrix in the channel dimension are extracted using the HOOI method; next, the received tensor and the angle domain dictionary are synchronously projected into a low-dimensional tensor quantum space; finally, the sparse power spectrum in the angle domain is estimated using a multi-shot sparse Bayesian learning model, thereby obtaining the target azimuth. This method can suppress noise components in each dimension while preserving multidimensional coupling relationships, reducing the data dimensionality of subsequent sparse processing, reducing the influence of dictionary correlation and noise, and obtaining a sharper azimuth spectrum and higher resolution under conditions of low signal-to-noise ratio, low snapshot, and adjacent targets.

This invention discloses an optical disc and an optical disc device. The optical disc includes a groove bottom channel and a groove top channel formed by spiral grooves. Both the groove bottom channel and the groove top channel have a plurality of recording blocks capable of recording data. Each recording block contains a predetermined number of frames, and each frame has a frame synchronization field of length T at a predetermined position. In adjacent groove bottom channels and groove top channels along the radial direction of the optical disc, the recording block in the groove bottom channel is the first recording block, and the recording block in the groove top channel is the second recording block. The first recording block and the second recording block are offset from each other by an angular offset length along the circumference of the optical disc. The optical disc disclosed in this invention has the advantages of a small reading signal amplitude during optical disc reading, thus eliminating the need for a large dynamic range and very high SNR performance in the reading signal processing circuit, thereby simplifying the implementation of the reading signal processing circuit.

The invention provides a multi-round dialogue intention classification method and device, and the method comprises the steps: defining and extracting a single-valued slot which is a core keyword contained in an FAQ intention; constructing intention clusters based on the single-valued slots, wherein the intention clusters of the same entity type and the intention clusters of the same single-valued slots are constructed; a single-valued slot inheriting slot mechanism is introduced into multiple rounds of dialogues, and different dynamic score adding strategies are performed on the intention clusters of the same entity type and the intention clusters with the same single-valued slot based on the recognition result that the hit intention has the single-valued slot; and performing intention classification again according to the intention score after score addition. The intention recognition accuracy and effectiveness can be improved.

The disclosed embodiments provide an animation synthesis method, an animation synthesis device, an electronic device and a storage medium. The method includes: determining the play time of each of a plurality of motion effect elements, each of which has a corresponding initial configuration file, the initial configuration file including initial external time parameter information, initial resource information and initial layer information; modifying the initial external time parameter information in the initial configuration file corresponding to each motion effect element according to the play time of each motion effect element, and obtaining the modified external time parameter information; generating new resource information and new layer information according to the modified external time parameter information, initial resource information and initial layer information; generating a new configuration file according to the modified external time parameter information, new resource information and new layer information; determining a target configuration file according to a plurality of new configuration files; calling the target configuration file to synthesize the animation.

The present application provides a method and device for estimating the DOA of a low-orbit satellite based on a phased array. The method provided by the present application includes: adjusting the calibrated fully connected multi-beam phased array based on multi-channel phase weighting to reconstruct a steering matrix; the steering matrix characterizes the response of the antenna unit of the phased array in multiple directions; wherein the phase weight corresponding to each channel of the antenna unit of the phased array is changed based on the multi-channel phase weighting, and the steering vector of the received signal in the antenna unit is reconstructed based on the phase weight; the steering matrix is ​​determined based on the steering vector; the steering matrix is ​​eigen-decomposed, and the DOA estimation result is determined based on the decomposition result; the motion trajectory of the low-orbit satellite is determined based on the DOA estimation result, and the DOA estimation result is updated based on the target tracking algorithm. The method and device provided by the present application are used to realize the DOA estimation of low-orbit satellites in maritime communication scenarios based on multi-beam phased arrays.

The invention relates to the technical field of audio signal processing, and provides an echo cancellation method and device, audio equipment and a storage medium, and the method comprises the steps: obtaining beam signals corresponding to different angles in a target angle range; selecting a first candidate beam signal from the beam signals, wherein the correlation between the first candidate beam signal and an input signal of the audio input device is greater than a first preset threshold; selecting a second candidate beam signal from the first candidate beam signals, wherein the correlation between the second candidate beam signal and the recovery reference signal is smaller than a second preset threshold; a preliminary beam signal is selected from the second candidate beam signals, a target beam signal is determined based on the preliminary beam signal, and the correlation between the preliminary beam signal and the beam signal corresponding to the angle of the audio output device is smaller than the correlation between other second candidate beam signals and the beam signal corresponding to the angle of the audio output device. According to the embodiment of the invention, the echo cancellation effect is ensured under the condition of limited computing resources.

The present invention discloses an array radar anti-deceptive interference method based on blind source separation. Based on the blind source separation technology, the present invention combines the wave arrival angle information of the received signal, that is, recovers the multi-source signal waveform of the excitation array antenna through blind source separation, which can effectively realize the separation of interference and target, and then uses the interference identification technology to effectively extract the target signal. This method can perform interference identification on each signal channel after blind source separation, and adaptively suppress deceptive interference. From the simulation results, the present invention has a significant anti-interference effect on new deceptive interference such as dense false target interference and forwarding slice interference.

This invention discloses a sparse Bayesian target azimuth estimation method based on far-field dictionary reconstruction under strong near-field interference. The method includes subspace decomposition of the observation data to construct a signal subspace projection operator; projecting and normalizing the far-field dictionary to obtain a projected reconstruction dictionary, which replaces the far-field dictionary in sparse Bayesian learning, and jointly performing sparse Bayesian iterative estimation with the near-field dictionary; finally, outputting the far-field target azimuth estimation result. This method effectively partitions the subspace and constructs a projection matrix accordingly to suppress the impact of strong near-field interference on far-field azimuth estimation, significantly improving the far-field target azimuth estimation capability of the sparse Bayesian method under strong near-field interference conditions.

A user equipment (“UE”) in a wireless communication network can generate a padded identifier by inserting a string of padding bits in a field of an identifier associated with the UE. The UE can further encrypt the padded identifier to generate a hidden padded identifier. The UE can further transmit the hidden padded identifier to a network node operating in the wireless communication network.

The invention relates to a separation device for brake dust generated by a wheel brake of a motor vehicle, the separation device being designed to be arranged on a wheel of the motor vehicle in a rotationally fixed manner and the separation device having a flow path, the flow path has at least one first part with a first flow direction and a second part with a second flow direction, the first part adjoins the second part in a boundary region, and the second flow direction forms an angle of at least 45 DEG, preferably at least 60 DEG and particularly preferably at least 90 DEG with the first flow direction, wherein the separating device has a pressure difference generating element for generating a pressure difference between an inlet pressure at the inlet of the flow path and an outlet pressure at the outlet of the flow path, and wherein a collecting section for braking dust is arranged between the boundary region and the outlet of the flow path. The invention further relates to a wheel for a motor vehicle having such a separating device and to a motor vehicle having such a separating device or having such a wheel.

The present invention relates to a controllable scaling convolution processing device and method based on random calculation. By sharing the pseudo-random number sequence generated by the random number source in the random number generation module, one round of convolution calculation uses only one linear feedback shift register, which greatly reduces the hardware cost and calculation overhead. At the same time, the random number generation module performs operations such as shifting and swapping the pseudo-random number sequence through shuffling operations, which effectively weakens the correlation between data, thereby greatly reducing the precision error caused by the correlation in the random calculation. In the convolution calculation module, a JK trigger is used to replace a binary divider to generate a control signal of a multiplexer in the convolution calculation, without relying on a large number of AND gates or XOR gates, which greatly simplifies the complexity of the convolution calculation architecture and further reduces the hardware cost, area occupancy and power consumption; finally, the data processing module completes the conversion and recovery of binary numbers.

The invention relates to the technical field of computer vision application, in particular to a high-voltage transmission line broken strand defect identification and detection method based on SAE, and the method comprises the steps: carrying out the unsupervised feature learning of an obtained transmission line image, obtaining the high-level local features, carrying out the feature selection of the high-level local visual features through a correlation analysis algorithm, and carrying out the recognition and detection of the broken strand defect of a high-voltage transmission line. The correlation between the features is reduced; then, a CNN model is utilized to carry out convolution on a to-be-recognized target object in a target field by using a local high-level feature vector learned by the SAE, and global features of the target object are obtained; and finally, the global feature response is sent to a classifier, and classification and identification of the target object are completed. Compared with the prior art, the identification accuracy is improved under the condition that the number of aerial images is limited.

The present application relates to the field of data encryption technology, and specifically to an industrial Internet data processing method and system based on big data, the method comprising: obtaining each data sequence of each industrial area in an industrial park; for any two data sequences, respectively recorded as the first and second data sequences, randomly extracting U sequences of the same length corresponding to the same time series interval from the first and second data sequences; obtaining predicted correlations according to the sequences at each random extraction number in the first and second data sequences; obtaining a set of frequent sequences between the first and second data sequences according to the distribution of predicted correlations between the data sequences; constructing a binary tree of the first data sequence according to the set of frequent sequences between the data sequences; generating an initial plaintext state matrix according to the characteristics of the binary trees of each data sequence, and improving the AES encryption algorithm to achieve optimized security protection for big data encryption. The present application improves the security of industrial data encryption.

The invention discloses a generation system and a generation method supporting supersampling noise and interference signals, and belongs to the technical field of digital signal transmission. The generation system comprises a parameter configuration module, a parallel noise generation module, a parallel interference signal generation module and a selection output module. The generation system is based on a parallel processing architecture of an FPGA, noise power, interference signals and output parameters are dynamically set through a parameter configuration module, and unused resources are cut off to optimize power consumption; the parallel noise generation module generates M paths of AWGN vector signals in real time; the parallel interference signal generation module supports real-time generation of various signals such as modulation, sweep-frequency comb-shaped signals and composite interference signals; and finally, the selection output module selects and outputs as required. According to the invention, the limitation of the FPGA main frequency on the sampling rate is broken through, large-bandwidth, high-resolution and continuous-phase signal generation is realized, and the problems that the flexibility of the traditional pre-storage scheme is insufficient, and the period and the signal bandwidth depend on the storage capacity are effectively solved.

The application discloses a high-speed data transmission method and device based on a silicon photon chip and a medium, relates to the technical field of data transmission, and comprises the following steps: a plurality of wavelength mode physical channels are constructed by utilizing a micro comb light source, a wavelength demultiplexing structure and a double-mode waveguide on a silicon photon chip; a training sequence is sent under default transmission of a programmable interference network, channel gain and noise are estimated to obtain state information; the number of virtual channels is determined and combination parameters are calculated according to the state information; virtual-real channel mapping is established, and on-chip optical path configuration is completed; a service forward error correction code is encoded on the sending side, and a multidimensional optical signal is generated by cross-virtual channel bit interleaving in the encoding block; after demultiplexing and equalization, the signal is transformed to a virtual channel domain on the receiving side, soft decision demodulation and error correction decoding are performed, a quality index is calculated, and channel state is updated; and linear combination parameter matrix and modulation parameters are further jointly optimized to realize low-error and high-throughput adaptive transmission under a dynamic link.

The present invention belongs to the technical field of social software development, and specifically, is a friend recommendation system and method based on ternary closure, comprising a data collection module, a data processing module, a weight calculation module, a state matching module, and a friend recommendation module; the data collection module determines whether to update data according to the user's geographical location and time; the data processing module pre-processes the user's activity data to cluster the user's activity data; the weight calculation module calculates the weight of the edge that may be formed between the recommended user and other users, adds the calculated weight value to a queue to be recommended, introduces an open coefficient, updates the weight value, and then adds it to the queue to be recommended; the state matching module introduces a state perception mechanism before recommendation to match the user's state with the state of other users; the friend recommendation module selects m users for recommendation for the user, and for each user, the weights of other users after state matching with the user are sorted in order of size, and the first m users are selected from them for recommendation.

The invention relates to a side channel attack resisting method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring an input message and an elliptic curve parameter; extracting verification points from the elliptic curve parameters, and judging whether the verification points meet the elliptic curve equation or not; generating a first random number, and performing blind processing on the first random number and the verification point to obtain a product of the first random number and the verification point; generating a signature based on the verification point and the elliptic curve parameter; and outputting the signature and the input message under the condition that the signature verification is passed. By combining multiple blinding and random number masking, the privacy and integrity of signature generation and message transmission are ensured, and the security protection capability of the system is remarkably improved.

The present invention relates to a medical imaging device and a medical imaging system, comprising: an adapter for connecting an endoscope; a handle, wherein the handle is arranged on the adapter; the image processing unit is positioned in the handle and is connected with the adapter, so that the image processing unit is allowed to receive an image transmitted by the endoscope; one end of the flexible cable is arranged on the handle and is electrically connected to the image processing unit; the flexible cable is installed on the handle, the button unit is installed on the flexible cable and electrically connected with the flexible cable, and the button unit and the handle are arranged in a spaced mode in the extending direction of the flexible cable.

The application discloses a hyperspectral image target intelligent detection method based on deep learning and belongs to the field of remote sensing image processing. The method comprises the following steps: selecting a background training set with large difference and representativeness between each other through a coarse detection and de-correlated background pixel selection strategy; generating sufficient target training sets by using a target spectrum band random replacement target training sample data enhancement strategy; mixing and disordering the background and target training sets, and inputting the background and target training sets into upper and lower branches of a multi-depth-multi-branch target detection network respectively for training; inputting a to-be-detected pixel and a prior target into the upper and lower branches of the trained network respectively for target detection, and outputting the network to form a final detection result. By using the application, target pixels in a hyperspectral image can be detected and highlighted, and background pixels can be effectively suppressed, so that a detection result with high detection precision, high background suppression degree and high target-background separation degree is obtained.

The present disclosure relates to the field of artificial intelligence technology, and in particular to a method, device and storage medium for a human-machine collaborative topic classification search mode, the method comprising: extracting a feature vector from an input query text, the feature vector comprising a plurality of input features; adjusting a preset sparsification model according to the feature vector to obtain an adjusted sparsification model, the adjustment comprising pruning a weight matrix of the sparsification model, the weight matrix comprising a plurality of weights, the weights being used to indicate the contribution of the corresponding input features to the output result of the sparsification model; according to the feature vector, obtaining the corresponding topic classification search result through the output of the adjusted sparsification model. The embodiment of the present disclosure dynamically adjusts and optimizes the sparsification model according to the feature vector of the query text, which involves dynamically pruning the weight matrix, thereby outputting the search results using the adjusted sparsification model to ensure the relevance and accuracy of the search results.

The application discloses a commodity sorting method, an electronic device, a storage medium and a computer program product, relates to the technical field of information retrieval and recommendation, and comprises the following steps: receiving a search request, obtaining an associated commodity set matched with the search request from a preset commodity library; filtering a first target commodity from the associated commodity set and filling the first target commodity into a first commodity recommendation page, wherein the first target commodity is an associated commodity at a first gear in the associated commodity set; mixing and sorting a second target commodity in the associated commodity set, and filling the second target commodity into a second commodity recommendation page, wherein the second target commodity is an associated commodity at a second gear in the associated commodity set, the commodity correlation represented by the second gear is lower than that of the first gear, and the priority of the second commodity recommendation page is lower than that of the first commodity recommendation page. The application solves the technical problem of poor sorting effect.

The embodiment of the invention relates to the field of optical engineering, and provides a differential spectrum reconstruction method, system and terminal. The method comprises the following steps: generating a broadband filter array imitating soil random shape pores; processing transmission characteristics of filtering units in the broadband filter array simulating the soil random-shaped pores by adopting a differential calculation structure to obtain a reconstructed measurement matrix; and reconstructing a spectrum by using a compressed sensing algorithm based on the reconstructed measurement matrix. According to the invention, the spectrum recovery fidelity as high as 98% is realized, and the resolution of 2 nanometers is realized.

The invention relates to the technical field of intelligent monitoring, and discloses a competitive analysis-based intelligent drug flow monitoring method and system. The method comprises the following steps: firstly, carrying out abnormal data identification and deletion on a competitor drug flow direction data set based on a time sequence to obtain a processed competitor drug flow direction data set; secondly, using an improved k-means algorithm to cluster the processed competitor drug flow direction data set twice to obtain a profitable drug flow direction data set and a deficit drug flow direction data set; a competitor profit monitoring model is established, an optimal profit solution is obtained through an improved firefly optimization algorithm, and medicine flow profit monitoring is achieved; and finally, training to obtain a BP neural network early warning model, outputting an early warning result, and realizing drug flow loss monitoring. By processing and analyzing the competitor drug flow direction data, the purpose of intelligently monitoring the drug flow direction is achieved, and the method is objective and accurate.

The utility model discloses a lens assembly. The lens assembly comprises an outer frame skeleton, a bayonet support member, a bayonet and an appearance member. The bayonet supporting piece is mounted on the outer frame skeleton; the bayonet is mounted at one end, away from the outer frame skeleton, of the bayonet supporting piece and used for connecting a camera body; and the outer part covers the peripheries of the outer frame skeleton, the bayonet supporting part and the bayonet, and is detachably connected with the bayonet supporting part. According to the technical scheme, through the structural design, the appearance part is only connected with the bayonet supporting part during installation, the correlation amount during installation of the appearance part is reduced, and compared with a traditional appearance part installation mode, the camera body is connected with the lens through the bayonet, so that the installation efficiency is improved. And after the image quality of the lens is adjusted to be qualified, an outer part can be mounted without a bayonet dismounting step, so that the continuous stability of the image quality is ensured. The utility model also provides a lens and an image pickup device.

This invention discloses a compact MIMO antenna system based on a connecting line, comprising a ground plane, a first antenna, a second antenna, and a connecting line disposed outside the first and second antennas. The outer ends of the first antenna and the second antenna are respectively connected to the ground plane. One end of the connecting line is connected to a non-end portion of the first antenna, and the other end of the connecting line is connected to a non-end portion of the second antenna. The compact MIMO antenna system based on the connecting line of this invention features a compact antenna structure, small antenna element size, and close element spacing, achieving a highly integrated, highly compact, highly isolated, and compatible MIMO antenna system with multiple antenna types. This invention is applicable to various wireless communication devices, especially to the application of large-scale arrays in terminal devices.