17results about How to "Improve filtering accuracy" patented technology

The invention relates to the technical field of battery health state analysis, and discloses a battery health state comprehensive evaluation method, system and device and a storage medium, and the method comprises the steps: carrying out the preprocessing of data collected by a battery management system, and obtaining a standardized data sequence; calculating a first battery characteristic parameter according to the standardized data sequence; performing covariance matrix adaptive Kalman filtering processing on the first battery characteristic parameter to obtain a second battery characteristic parameter; constructing a capacity-internal resistance change rate index and a temperature-efficiency correlation index based on the second battery characteristic parameters to obtain a dual-threshold evaluation matrix; inputting the dual-threshold evaluation matrix into a deep reinforcement learning model, and performing forward propagation calculation and iterative training through an Actor network and a Critic network to obtain a feature fusion weight coefficient; and carrying out weighted calculation on the dual-threshold evaluation matrix to generate a battery health state value and a residual life prediction value, and realizing multi-dimensional evaluation of the battery health state.

The invention relates to the field of signal processing, and discloses a biological signal processing method and device, computer equipment and a storage medium, and the method comprises the steps: collecting a biological signal, and carrying out the preliminary filtering of the biological signal, and obtaining a preliminary filtering signal; performing noise characteristic analysis on the preliminary filtering signal, selecting a proper filtering mode according to an analysis result, and designing filter parameters through fuzzy logic processing; inputting the filter parameters into an adaptive filter to process the signal so as to obtain a filtered biological signal; in combination with a de-noising convolutional neural network model, noise identification and fuzzy logic processing are performed on the filtered biological signals, and filter parameters are dynamically adjusted; and performing multi-round iterative optimization until a target filtering effect is obtained, and outputting an optimized biological signal. According to the biological signal processing method, complex noise in the biological signal is effectively removed, the filtering precision is improved, meanwhile, the signal integrity is kept, and a reliable basis is provided for subsequent biological signal analysis and processing.

The invention discloses a dynamic harmonic filtering control method and system, and relates to the technical field of electric power, and the method comprises the steps: collecting a voltage signal and a current signal of a power grid in real time; performing spectral analysis on the collected signals, and extracting harmonic characteristics; based on the harmonic characteristics, generating a compensation signal and optimizing the compensation signal; dynamically adjusting filtering parameters, and injecting an optimized compensation signal into a power grid through a filter module; uploading a system operation state in real time, and detecting and marking an abnormal state; according to the invention, rapid response and efficient suppression of power grid harmonics can be realized, the adaptability and filtering precision of multi-band harmonics are significantly improved, and the total distortion rate of the harmonics is greatly reduced, so that the quality of electric energy is optimized. The system operation efficiency and stability are improved, the anti-interference capability is strong, and the system is suitable for a complex power grid environment; the system can monitor the operation state in real time, automatically give an alarm and mark abnormity, the reliability and safety of power grid operation are improved, and the maintenance cost is reduced.

The invention relates to the technical field of off-grid photovoltaic disturbance rejection, in particular to a method for enhancing off-grid photovoltaic disturbance rejection capability based on an intelligent filtering technology. Firstly, known disturbances such as illumination, temperature and load abrupt change in the system are analyzed, and a model is established. Designing an intelligent filter, monitoring and estimating unknown disturbance by using a self-adaptive algorithm, obtaining direct current and harmonic disturbance observation values, and fusing the direct current and harmonic disturbance observation values; and then an active disturbance rejection controller based on intelligent filtering is constructed, unknown disturbance and known disturbance are added to obtain a concentrated disturbance observation value, and parameters are designed to respond to suppression disturbance. And finally, obtaining a voltage preliminary control instruction, and combining with a filter observation value construction instruction to suppress current harmonic disturbance. Through the steps, the stability and reliability of the off-grid photovoltaic system are improved.

The invention discloses a fractional derivative positioning optimization method based on adaptive data driving, which belongs to the technical field of radio positioning, and comprises the following steps: S1, receiving a signal strength indication value at an actual position by using a sensor in a target area, and generating an RSSI data set; s2, according to the RSSI data set, generating and training a fractional derivative kernel adaptive filter; and S3, inputting the actual position and the RSSI data set into the trained fractional derivative kernel adaptive filter to obtain a target position. Compared with a traditional classical kernel adaptive filtering method, the method has the advantages that a fractional order derivative is adopted to construct a weight updating strategy, and higher and more stable filtering precision is achieved; according to the method, the fractional calculus is combined with the q Laplacian kernel technology to be introduced into the kernel adaptive filter to construct the positioning system for the first time, and experiments show that the algorithm can obtain better positioning precision compared with the current most advanced positioning algorithm.

The invention discloses a multichannel data acquisition device, a multichannel data acquisition method and a radar system. The multichannel data acquisition device comprises M analog acquisition channels, N AD chips, an FPGA, a clock module and a DDR. Each analog acquisition channel is used for converting a single-ended analog signal into a differential analog signal; each AD chip is used for converting the differential analog signal into a digital signal; the FPGA is used for synchronously capturing digital signals output by the plurality of AD chips; performing time-frequency conversion on each digital signal to obtain a corresponding frequency domain signal; carrying out filtering processing on each frequency domain signal; and storing and uploading the signal after the filtering processing. The clock module is used for providing a reference clock for the whole device; and the DDR is used for caching the data processed by the FPGA. The problems that a time domain filtering method is complex in calculation and frequency response is difficult to control are solved.

The invention relates to the technical field of filters, and discloses a design method and system for an ultrahigh-precision filter inductor, and the method comprises the steps: designing a filter structure, and generating a filter parameter corresponding to the filter structure; constructing a test current of the initial filter, and analyzing an output current change value on an initial inductor in the initial filter by using the test current and the output current; calculating the power loss of the initial filter under the change values of the test current and the output current, calculating the return loss of the initial filter, calculating the quality factor of the initial inductor, and analyzing the out-of-band rejection of the initial filter; constructing a target function of the initial filter, screening a standard structure and a standard parameter from the filter structure and the filter parameter, and constructing a standard filter by using the standard structure and the standard parameter; and extracting a demand distribution characteristic from the user current demand, identifying a current distribution characteristic of the standard filter, and comparing the demand distribution characteristic with the current distribution characteristic. The filtering precision of the filter inductor can be improved.

Embodiments of the present application provide a Bloom filter construction method applied to a distributed database, a data query method and related equipment. In the embodiments of the present application, a query execution engine receives an execution plan sent from a query optimizer, wherein target indication information contained in the execution plan explicitly specifies that a bit array size of a Bloom filter is dynamically determined according to a number of records of a right table distributed in a secondary node. In an execution process, the query execution engine triggers a construction operator in the execution plan to read all records of the right table distributed in the secondary node row by row in a full table scan manner, to determine the number of records of the right table distributed in the secondary node and a hash value of a connection key of each record, and to construct the Bloom filter according to the information. Thus, the bit array size of the Bloom filter is adaptively adjusted according to the actual number of records of the right table distributed in the secondary node, so that the Bloom filter can better adapt to data distribution characteristics, and the screening effect of the Bloom filter is improved.

The invention relates to a port illumination system life evaluation and health index model training method, which specifically comprises the following steps: firstly, acquiring operation data such as brightness, temperature and voltage of an illumination system and salt mist concentration data in real time through multiple sensors, and constructing a time sequence data set with a real residual life label; mapping the operation data to a quantum space to construct a quantum state and Hamiltonian, and solving an optimal filtering path through annealing optimization to obtain filtering data; then mapping the filtering data into a Riemannian manifold, calculating a Chen-Simons integral value in combination with a Laplacian operator, external differential and the like, and generating a feature deconstruction operator; constructing a fractional order space-time convolution network to extract features, and constructing and optimizing a residual life prediction model in combination with salt mist concentration and Weber distribution shape parameters; and finally, inputting new preprocessed data to obtain a prediction result, and calculating a health index. According to the invention, the data quality and the characteristic representation capability are improved, and the residual life prediction precision of the port lighting system can be effectively improved.

The invention provides a log analysis method and device, equipment and a storage medium, which are used for solving the problem of low analysis accuracy and efficiency in a log analysis process. The method at least comprises the following steps: extracting a plurality of universal fields contained in each collected log to be analyzed; aiming at a plurality of preset time intervals, respectively executing the following steps: respectively counting respective field values of a plurality of universal fields and respective occurrence times of each field value in each to-be-analyzed log in the time intervals to obtain a corresponding to-be-processed item set; based on the respective field values and the occurrence times of the plurality of universal fields, respectively determining respective value uniformity of the plurality of universal fields; when it is determined that a target field with the value uniformity smaller than a uniformity threshold exists in the multiple universal fields, data filtering is conducted on the to-be-processed item set corresponding to the target field, and a target item set is obtained; and on the basis of each obtained target item set, performing association behavior analysis on each to-be-analyzed log to obtain a log analysis result.

A highly maneuverable target tracking method, the implementation steps of which are: (1) algorithm initialization, determining the algorithm initial conditions according to system prior information; (2) calculating the mixing probability; (3) model interaction; (4) parallel filtering of r filters; (5) updating the model posterior probability; (6) adaptive updating of the model transition probability; (7) state fusion output; (8) taking n=n+1, repeating steps (2) to (6) to achieve position tracking of the highly maneuverable target. The present invention can achieve position tracking of the highly maneuverable target when the prior information is insufficient or inaccurate. The present invention can be applied to fields such as communication, radar, measurement and control that have a demand for maneuverable target tracking.

This invention discloses a cavity electrically tunable preselection filter, relating to the field of wireless communication equipment. It includes a mounting cavity and a PCB board. The PCB board is fixed to the mounting cavity via connectors and houses an integrated circuit and a coil assembly. The mounting cavity contains four resonant cavities and two circuit-shielded resonant cavities. The four resonant cavities are arranged side-by-side along the length of the mounting cavity. The two circuit-shielded resonant cavities are arranged side-by-side along the width of the mounting cavity and located in the middle of the four resonant cavities, serving to shield the integrated circuit and provide resonance for the resonant cavities on either side. The coil assembly includes four coils for specific frequency bands. The four coils are electrically connected to the PCB board and located within the four resonant cavities. It features a compact structure, excellent filtering performance, and adaptability to complex electromagnetic environments.

The invention discloses a rectangular coaxial adapter with a filtering function, and relates to the technical field of coaxial adapters, the rectangular coaxial adapter comprises a shell, a conductor is slidably mounted in the shell, bushes are arranged at the two ends of the conductor, a via hole is formed in the side wall of each bush, a contact sleeve is arranged on one side of each via hole, and the contact sleeves are arranged in the via holes. Firstly, the conductor is installed in the shell, then the insulation sleeve is arranged on the output conductor rod and the input conductor rod in a sleeving mode, then the bush and the contact head sleeve are sequentially arranged on the outer sides of the output conductor rod and the input conductor rod in a sleeving mode, and at the moment, the bolt rod on the side wall of the bush is inserted into the bolt groove; limiting and supporting work of the conductor in the shell can be completed through cooperation of the bolt groove and the bolt rod, then equipment connection work can be conducted, the insulation sleeve can conduct limiting and supporting on the output conductor rod and the input conductor rod, it is guaranteed that the shell and the input conductor rod are located at the axis of the connecting hole, the structure is light and handy, dismounting and mounting are convenient, welding is not needed, and the cost is low. Installation and debugging are convenient.

This invention provides a signal correction method and system for a strain gauge earth pressure sensor under centrifugal conditions with hypergravity. The method acquires the original signal from the earth pressure sensor, piezoelectric vibration signals from the bottom of the basket and the top of the model box, and a strain gauge resultant force signal coplanar with the earth pressure sensor. The vibration signals are then subjected to FFT transformation and frequency domain difference to obtain the vibration frequency domain of the earth pressure sensor plane, and filtered. The actual hypergravity is decoupled by subtracting the bottom surface vibration signal from the strain gauge resultant force signal, and then converted back to the actual hypergravity at the sensor location. Pre-calibrated centrifugal acceleration-zero drift and centrifugal acceleration-sensitivity relationships are used for dual compensation, and the corrected pressure data is output. This invention systematically eliminates zero-point drift, sensitivity changes, and vibration interference caused by hypergravity, significantly improving measurement accuracy.

This invention provides a vegetation filtering method for dangerous rock slopes based on multi-dimensional features, comprising: Step 1, obtaining an initial three-dimensional model of the dangerous rock slope; Step 2, generating a reference surface based on the initial three-dimensional model; Step 3, initially filtering the vegetation of the dangerous rock slope; Step 4, further filtering the vegetation of the dangerous rock slope. This invention employs a layered filtering strategy of "initial filtering and secondary filtering" to filter vegetation on dangerous rock slopes, significantly improving filtering accuracy and effectively avoiding the problems of "over-stripping" and "under-stripping" in traditional methods.

This invention discloses a data quality control method for millimeter-wave cloud measuring instruments based on the XGBoost model, belonging to the field of millimeter-wave cloud measuring instrument data quality control technology. The method includes the following steps: S1, cloud measuring instrument data input; S2, multi-dimensional feature extraction: multi-dimensional feature extraction is performed on the data input in step S1 by calculating the time-domain coefficient of variation, extracting multivariate texture features, and considering echo energy distribution skewness; S3, preliminary quality control processing; S4, secondary verification based on power spectrum data; S5, outputting the filtered result. This invention extracts specific features for two types of interference echoes using multi-dimensional features—time-domain coefficient of variation, texture features, and energy distribution skewness—and utilizes the model to adaptively learn the interference patterns under different scenarios, achieving simultaneous identification and filtering of the two types of interference. Simultaneously, power spectrum recognition is introduced to further improve the filtering accuracy. This method eliminates the need for separate processing logic, simplifying the process while improving filtering accuracy and scenario adaptability.

This invention discloses a high-Q tuned filter, belonging to the field of tuned filter technology. The tuned filter includes two symmetrically arranged resonant cavities, each connected in parallel with an inductor. This invention reduces the current density of a single coil, suppresses high-frequency skin and proximity effects, significantly reduces the high-frequency loss resistance of the inductor, and significantly improves the Q value at the same operating frequency. It directly reduces in-band insertion loss and improves out-of-band rejection depth, resulting in a significant improvement in filtering accuracy and anti-interference capability.