163results about How to "Improve filtering performance" patented technology

In one embodiment, an audio-visual content delivery system, such as a set-top box / personal video recorder system, is configured to interface with a local area network (LAN). A packet processing circuit comprised in the system may be configured to filter and route Ethernet packet data received from the LAN to specific ports and / or queues without host processor intervention. The packet processing circuit may utilize a set of filter and routing mechanisms configurable in hardware to interpret various Internet Engineering Task Force (IETF) networking transport protocols, and may transfer the packet data in a format recognized by a variety of consumer subsystems, each of which may be coupled to the decoder. The packet processing circuit may be implemented as a semiconductor device, and may allow encapsulated application data, (encapsulated using standardized encapsulation techniques), to be routed to a plurality of different types of application sinks or processors, forming a point-to-point or multi-point serial or parallel data stream over a standard transport covering numerous levels of the ISO data communications stack.

The invention discloses a novel radio frequency reconfigurable band-pass filter with a wide frequency and a wide adjustment range of broadband. The reconfigurable filter comprises an upper-layer microstrip structure, a middle-layer medium substrate and the lower-layer grounded metal. The upper-layer microstrip structure is composed of an input port, an input feed network, a cross adjustable multimode resonator, an output feed network and an output port which are connected in succession. The input feed network and the output feed network are both composed of a blocking condenser and a 50-ohm microstrip line. The cross adjustable multimode resonator is composed of two microstrip lines which are perpendicular to each other and six varactors with two ends of the varactors connected. The input feed network and the output feed network tap feed electricity respectively from the two varactors with two ends of the varactors in series connection of the left side and the right side of the multimode resonator. The reconfigurable filter has the advantages of being wide in central frequency and adjustment range of bandwidth and good in passband stability, and being capable of being reconfigured the broadbands with different central frequencies or a narrow-band band-pass filter and can be used for the front end of the reconfigurable radio frequency.

An embodiment of the invention provides a filtering device and a power supply system. The filtering device is connected between an external power supply and equipment and comprises a feedback-type active common-mode filter and a feedforward-type active common-mode filter. The feedback-type active common-mode filter comprises a common-mode noise detection device and a first filtering circuit. The feedforward-type active common-mode filter comprises the common-mode noise detection device and a second filtering circuit. The first filtering circuit is connected between the common-mode noise detection device and the equipment and is used for carrying out feedback-type filtering on a first common-mode noise signal to obtain a second common-mode noise signal; the common-mode noise detection device is connected between the first filtering circuit and the second filtering circuit and is used for detecting the second common-mode noise signal and providing the second common-mode noise signal to the second filtering circuit; and the second filtering circuit is connected between the external power supply and the common-mode noise detection device and is used for carrying out feedforward-type filtering on the second common-mode noise signal. The filtering device and power supply system can effectively improve filtering efficiency of the common-mode noises to some extent.

The invention discloses an optimized permanent magnet synchronous motor sensorless control method, which comprises the following steps: Clark coordinate transformation is performed on a three-phase current and a three-phase voltage acquired by a sensor to obtain a voltage and a current under a two-phase static coordinate system, and the voltage under the two-phase static coordinate system is inputto a sliding mode current observer to obtain an observed current; the difference between the observed current and an actual current is input to a back electromotive force observer based on a saturation function to obtain an estimated initial value of back electromotive force, and the sliding mode gain ksmo in the back electromotive force observer based on a saturation function is adjusted by adopting a variable universe fuzzy control algorithm; the estimated initial value of back electromotive force is filtered through a variable cut-off frequency filter to obtain a smooth estimated value ofback electromotive force; and variable lag compensation design is performed on the rotor position after filtering, and the rotor speed and the rotor position value of the motor are calculated. The buffeting phenomenon can be weakened, the rotor speed and position information estimation precision can be improved, and good dynamic characteristic is achieved.

The invention discloses a channel estimation method for a multi-input multi-output spacial multiplexing system. The method comprises the following steps: acquiring a primary channel coefficient in a frequency domain by a least square estimation method, and transforming the primary channel coefficient to a time domain through inverse Fourier transform; introducing a concept of an effective tap of a channel impulse response, and taking the maximum of the channel impulse response outside a circular preamble range as a judgment condition; searching an effective tap range, and filtering noise interference in the part outside the effective tap range by a zero padding method in the time domain; and converting to the frequency domain, replacing by a least square estimation channel coefficient of a data carrier, and restoring a channel coefficient of a null subcarrier by an iterated interpolation method; and finally acquiring channel estimation approximate to an actual channel coefficient. The time domain characteristics of the channel response of a wireless wide-band communication system are fully utilized, the traditional channel estimation method is improved through noise suppression, and the influence of the noise on a channel estimator is eliminated.

A fuel pump module, in which an in-tank filter and a fuel pump are installed at respective locations and held by a pump retainer, thereby allowing fuel to easily flow in the fuel pump module and realizing smallness and improved filtering performance of the fuel pump module.

A respiratory information detecting device comprises a respiratory signal collection unit, a pre-amplification unit, a demodulation unit, a gain and offset unit, an analog-to-digital (A / D) conversion unit, a signal processing unit and a display. The A / D conversion unit is a multi-bit conversion unit of at least 16 bits. The inventive method and device is characterized by adopting a multi-bit A / D (more than or equal to 16 bit) conversion scheme, pre-amplifying a respiratory signal by low-power gain, directly acquiring the original respiratory signal by using multi-bit A / D conversion, analyzing and identifying minutiae features of the original respiratory signal by using software in MCU to obtain respiratory waveform approximate to a real signal as well as base resistance information of human body, and setting amplitude threshold of respiratory rate calculation by using the base resistance information of human body. Due to adopting the multi-bit A / D conversion, a hardware circuit has a relatively wide variable resistance detection range, the respiration measurement accuracy is improved, and the hardware circuit is simplified.

A high pass digital filtering method of nutation frequency automatic track includes testing K phi phi without cross decoupling control, using Hall transducer to detect rotary speed of rotor and obtaining said rotary speed by microprocessor after regulation, substituting obtained rotary speed into nutation frequency calculation formula and on-line calculating out nutation frequency, using calculated out nutation frequency as cut-off frequency to design digital filter and inputting rotor displacement signal to carry out on-line filtering calculation.

The present invention discloses distributed prague reflection-type single frequency optical fiber laser based on phase displacement optical fiber grating, characterized in that filtering frequency selection is realized by utilizing narrowband filter effect of reflection peak of equivalent Fabry-Perot cavity that composed by the phase displacement optical fiber grating and optical fiber grating. The whole device comprises a pumping source, a wavelength division multiplexer, optical fiber grating, phase displacement optical fiber grating, high concentration adulterate optical fiber and optical fiber grating, wherein the optical fiber grating, the phase displacement optical fiber grating, the high concentration adulterate optical fiber and the optical fiber grating constitutes the laser resonant cavity. The present invention is capable of outputting high power single frequency laser.

The invention discloses an indoor pedestrian navigation method based on a self-backtracking algorithm. A pedestrian being navigated wears an MEMS sensor. The navigation method comprises the followingsteps: carrying out pedestrian dead reckoning according to an MEMS gyroscope and an MEMS accelerometer to obtain pedestrian walking information during a period of time; carrying out straight line detection on the walking information in the period of time to judge whether the walking route of the pedestrian in the period of time is a straight line; if determining that the pedestrian walks in a straight line, constructing pseudo observed quantity according to heading angle information; carrying out navigation error detection and correction compensation through a forward and backward adaptive kalman filtering algorithm; and outputting a pedestrian indoor navigation and positioning result. By fully exploring the gait characteristics of the pedestrian when walking, the method carries out navigation error estimation and compensation on pedestrian position and attitude information by utilizing walking constraint conditions, thereby realizing a purpose of improving position and posture determination accuracy of the indoor pedestrian navigation scheme based on the motion sensor.

The invention discloses a double-frequency substrate-integrated waveguide band-pass filter with a double-layer structure. The double-frequency substrate-integrated waveguide band-pass filter comprises a lower-layer structure and an upper-layer structure, wherein a substrate-integrated waveguide is formed by the lower-layer structure and the upper-layer structure, and the lower-layer structure comprises a first metal sheet, a first dielectric base plate, a second metal sheet and a plurality of lower-layer metal through holes; the second metal sheet is encircled into a first resonator and a second resonator through the lower-layer metal through holes, a first passing hole is loaded in the first resonator, a first gap is etched, a second passing hole is loaded in the second resonator, and a second gap is etched; a third metal sheet is encircled into a third resonator and a fourth resonator through upper-layer metal through holes, a third passing hole is loaded in the third resonator, and a fourth passing hole is loaded in the fourth resonator; the first resonator is coupled to the third resonator through the first gap, and the second resonator is coupled to the fourth resonator through the second gap. The filter disclosed by the invention has the advantages of high quality factor, small size, simple design and good performance.

The invention discloses an EMI noise source impedance equivalent parameter extraction method based on scattering parameters and an intelligent algorithm. The EMI noise source impedance equivalent parameter extraction method is based on a scattering parameter principle, adopts a vector network analyzer and two current probes, respectively measures transmission parameters and reflection parameters when in short circuit, loading standard resistance and loading noise source impedance, calculates to obtain the impedance of a measured noise source, carries out optimizing processing on amplitude-frequency and phase-frequency characteristics of the measured impedance by utilizing a GA algorithm in the intelligent algorithm to obtain equivalent RLC parameters of the impedance, and provides general expression of amplitude value and phase position of the impedance. The EMI noise source impedance equivalent parameter extraction method is simple in instrument, does not need to use an auxiliary module, reduces hardware error caused by auxiliary modules such as a signal source and a frequency spectrograph, estimates parameters by using the GA algorithm, avoids the problem that initial values of parameters need to be given when methods such as the Newton-Gauss method and Marquardt method are adopted, increases measuring precision, and provides an accurate reference basis for the design of an EMI filter.

The invention relates to a yellow filter quartz tube which is prepared by using high purity quartz sand as basic raw materials in parts by weight, and compounding with 0.7-1.35 parts by weight of colorant, wherein the colorant comprises diboron trioxide (B2O3), zinc oxide (ZnO), potassium oxide (K2O), sodium oxide (Na2O), europium sesquioxide (Eu2O3) and titanium dioxide (TiO2). The yellow filter quartz tube molded by drawing through adopting a tube wall integral coloring preparation process has good uniformity, high temperature resistance, stable filter performance and large light absorption range, and the absorbed cutoff wavelength can reach 420 nanometers. A tube material is not deformed at the temperature of 1600 DEG C, and when the thickness of the quartz tube is 1+ / -0.5 millimeters, the spectral transmission rate is about 63 percent and the light wavelength is 540-620 nanometers.

The invention provides a sludge storing pipe bag for dehydration and a dehydration method. The sludge storing pipe bag comprises a body (1) and is characterized in that an anode (2) and a cathode (3) connected with a power supply are installed inside the body (1). The invention effectively improves dehydration efficiency of the sludge storing pipe bag, and the dehydration and consolidation are carried out on the site and are clean and efficient, thereby greatly shortening time for dehydration and consolidation; water filtered and discharged from the wall of the pipe bag substantially meets the standards, which avoids the secondary pollution caused by harmful elements seeping into the ground during dehydration in a settling tank; and consolidated sludge can be conveniently applied to various works with soil. The invention has obvious economic benefit and social ecological benefit.

The invention discloses a broadband balance power amplifier based on GaN. The broadband balance power amplifier based on the GaN comprises a Lange coupler, a microstrip transmission line, a Banding metal wire, a filter inductor, a stabilizing resistor and a wide band gap GaN-based HEMT. A biasing circuit of the power amplifier comprises a T-shaped connector, the microstrip transmission line, the filter inductor, the stabilizing resistor and the GaN-based HEMT, the GaN-based HEMT (H) is an HEMT pipe core with the type of NRF01-02a, and a four-line 3dB coupling Lange coupler with the characteristic impedance of 50 ohms is selected and used as the Lange coupler. The broadband balance power amplifier based on the GaN is simple in structure and low in cost, the Lange coupler is adopted for constructing the structure of the balance power amplifier, a multi-section impedance matching technology is adopted for designing an input / output matching network, the power amplifier broadband characteristics are achieved, and within the 115-315 GHz frequency band, the power amplifier linear gain is larger than 12 dB, and the saturation output power is larger than 8 W.

The present invention discloses a transformerless hybrid power filter. The transformerless hybrid power filter comprises a power grid-side voltage source, a system impedance, an inverter-side DC source, an active voltage source, a passive filter branch, a passive series fundamental wave resonance branch and an output filter. The passive filter branch and the passive series fundamental wave resonance branch are both based on LC filters, the output filter is based on an LCL filter circuit with the damping resonance, and one end of the power grid-side voltage source is connected with the other end of the voltage source via the system impedance, and via the passive filter branch and the passive series fundamental wave resonance branch which are in series connection with each other. One end of the voltage source is connected with the other end of the voltage source via the system impedance and via the inverter-side DC source, the connection point of the passive filter branch and the passive series fundamental wave resonance branch is connected with the output filter, and the output filter is connected with the two ends of the active voltage source. According to the present invention, the filter characteristic and the system stability are guaranteed on the condition of not having a coupling transformer.

The invention relates to the technical field of medical care, in particular to a method and a device for improving accuracy of non-invasive blood pressure (NIBP) measurement. The method comprises the following steps: converting original pressure signals to digital pressure signals, respectively carrying out high-pass digital filtering, low-pass digital filtering and storage on the digital pressure signals, analyzing pulse signals obtained by high-pass digital filtering and cuff pressure signals obtained by low-pass digital filtering and judging whether the signals are abnormal or not; if abnormal, extracting the stored corresponding digital pressure signal and judging whether the digital pressure signal is abnormal or normal; if abnormal, outputting abnormal information; if normal, extracting the signal and carrying out comprehensive computation and analysis to obtain data results. In the method and the device provided by the invention, since digital signal filtering is adopted, the introduction of electrical noise is reduce and filtering performance is optimized; and the original digital pressure signals are stored. thereby facilitating backtrack and analysis and ensuring the accuracy and stability of the NIBP measurement.

The invention relates to a voltage stabilizing system, in particular to a boost chopping technology-based variable-frequency voltage stabilizing control system and a method, which belong to the electric automation field. The booster chopping technology-based variable-frequency voltage stabilizing control system mainly comprises a host computer, a program logic controller (PLC) 1, a variable-frequency system 2, a CC-LINK on-site bus 6, a remote monitor 5, a speed regulator 10 and a boost voltage frequency control conversion circuit; the PLC 1 is connected with the variable-frequency system 2 in a communication way through the CC-LINK on-site bus 6; the signal input end of the PLC 1 is respectively connected with an energy storage signal, an energy release signal and an alarming signal, and then the PLC 1 operates and processes the inputted data through a central processing unit CPU 13; and the output of the stabilized voltage of the system can be realized through a voltage / frequency (V / F) conversion circuit 24. The voltage stabilizing control system has good filter performance, a good energy-saving effect, good static performance and good dynamic performance, realizes the real-time monitoring, and improves the transmission accuracy and reliability of the control signal.

The invention relates to a wideband and low-noise differential application circuit for measuring a weak signal. The amplification circuit comprises a first-stage amplification circuit, a second-stage amplification circuit and a third-stage amplification circuit which are sequentially coupled and connected; both the first-stage amplification circuit and the third-stage amplification circuit are proportional differential amplification circuits; the second-stage amplification circuit is a high-pass filter differential amplification circuit; the first-stage amplification circuit comprises a first resistor, a second resistor, a third resistor, a first integrated operation amplifier and a second integrated operation amplifier; the second-stage amplification circuit comprises a fourth resistor, a fifth resistor, a sixth resistor, a seventh resistor, a first capacitor, a second capacitor and a third integrated operation amplifier; the third-stage amplification circuit comprises an eighth resistor, a ninth resistor and a fourth integrated operation amplifier. The wideband and low-noise differential application circuit for measuring the weak signal is a wideband and low-noise amplification circuit provided for the shortages of a weak signal output detecting circuit of a marine sensor; the amplification circuit is capable of effectively amplifying nanometer-stage weak signals within the wideband range of 0.1Hz to 10MHz.

The invention relates to a plane magnetic integrated electro magnetic interference (EMI) filter. The plane magnetic integrated EMI filter comprises a plane magnetic integrated common mode inductor, a printed circuit board (PCB) which is welded with a differential mode capacitor or a common mode capacitor, a non-magnetic metal shell, a grounding leading-out terminal and an inductor leading-out terminal, wherein the plane magnetic integrated common mode inductor comprises a pair of U-shaped magnetic cores and cylindrical or cubic magnetic cores between the pair of U-shaped magnetic cores; the U-shaped magnetic cores are covered with single-layer or multi-layer windings; the cylindrical or cubic magnetic cores are accommodated in the non-magnetic metal shell and are insulated; the PCB is connected with the plane magnetic integrated common mode inductor; and one end of the inductor leading-out terminal is connected with the windings, and the other end of the inductor leading-out terminal is connected with the PCB. The plane magnetic integrated EMI filter has the advantages of flat shape, smaller volume, high power density, high filtering performance, standardized and modularized production and the like.

The invention discloses a method for realizing guided-mode resonance filtering through a single gradient-material grating structure, and belongs to the field of an optical communication and MOEM (micro optical electro mechanical) system. The method is characterized in that an optical thin film, the refractive index of which increases progressively with the thickness, is prepared on a substrate, and by etching the refractive-index-gradient thin film, a guided-mode resonant grating structure is obtained, and furthermore, guided-mode resonance filtering can be realized; on the basis of the above, by selecting different etching depths, channel position of a filter can be adjusted; and with the etching depth being kept unchanged, by reducing gradient coefficient, multichannel filtering can be realized. The filtering performance of the guided-mode resonant grating structure is highly insensitive to change of substrate refractive index, so that even if the substrate refractive index is higher than the maximum value of the refractive index of the gradient thin film, the guided-mode resonance filtering performance keeps excellent; and the method breaks away from the limit that refractive index of a waveguide layer in a conventional guided-mode resonance filter needs to be higher than the substrate refractive index, and is more advantageous in practical application.

The invention provides a composite film with a heat conducting, wave absorbing and shielding three-in-one function. The composite film comprises a metal layer, a wave absorbing layer and carbon coating layers; and the carbon coating layers are arranged on the two sides of the metal layer correspondingly, and the wave absorbing layer is further arranged on the other side of one of the carbon coating layers. A film material prepared from the carbon coating layers, the metal layer and the wave absorbing layer in a compositely designed mode has the multiple functions of heat conducting and dissipation, electromagnetic wave absorbing and electromagnetic shielding, can be applied to various working environments, and the problems that in the prior art, the electromagnetic shielding effect is poor, and heat conducting and dissipation performance is poor are solved.

The invention discloses a terahertz dual-band elimination filter. The terahertz dual-frequency band elimination filter comprises N * N identical square units which are periodically arranged on a planeperpendicular to a wave incident direction, and no interval exists between the units; each unit comprises a metal resonance layer and a substrate; and each metal resonance layer comprises a first semi-metal ring, a second semi-metal ring, a metal ring, a first metal strip, a second metal strip and a metal circle; the metal circle is arranged in the center of the metal ring, the first metal stripand the second metal strip are respectively arranged right above and right below the metal ring, the first semi-metal ring and the second semi-metal ring are respectively arranged right above and right below the first metal strip and the second metal strip, and the first semi-metal ring and the second semi-metal ring form a circular ring with two gaps. The dual-stop-band characteristic of the filter in the terahertz wave band is achieved, and the filter has the advantages of being low in loss, simple in structure and low in material and process cost.

The invention discloses a multi-channel terahertz wave filter. The multi-channel terahertz wave filter comprises dielectric cylinders and a signal input end, a first signal output end, a second signal output end, a third signal output end, a fourth signal output end, a fifth signal output end, a sixth signal output end, a seventh signal output end, a first photonic crystal waveguide, a second photonic crystal waveguide, a third photonic crystal waveguide, a fourth photonic crystal waveguide, a fifth photonic crystal waveguide, a sixth photonic crystal waveguide, a seventh photonic crystal waveguide, a first resonant cavity, a second resonant cavity, a third resonant cavity, a fourth resonant cavity, a fifth resonant cavity and a sixth resonant cavity which are arranged among the dielectric cylinders, wherein the dielectric cylinders are arranged in a two-dimensional period, and terahertz waves with seven different frequencies are input and are resonant to different output ports to be separately output respectively by different structured photonic crystal resonant cavities, so that a filtering function is achieved. The multi-channel terahertz wave filter has the advantages of simplicity in structure, high filtering performance, small size, low cost and the like, and is easy to integrate.

The invention relates to a weak signal detection technology, aims to overcome the defects of narrow frequency band and high noise of a detection circuit for outputting weak signals by a marine sensor,and aims to provide a broadband and low-noise amplification circuit whichCan effectively amplify Nanoscale (nV) weak signals in a 0.1Hz-10MHz broadband range, and is low in the self noise level. Therefore, the technical scheme adopted by the invention is as follows: the broadband and low-noise differential amplification circuit for weak signal measurement comprises three stages of amplification circuits, the first stage amplification circuit adopts a proportional differential amplification circuit, the second stage amplification circuit adopts a high-pass filtering amplification circuit, andthe third stage amplification circuit adopts a proportional amplification circuit; And two paths of differential output of the first-stage amplifier respectively enter the in-phase input end and the inverted input end of the second-stage amplifier. The method is mainly applied to weak signal detection occasions.

The invention discloses a method for designing a one-dimensional photonic crystal filter based on digital signal processing. The method for designing the one-dimensional photonic crystal filter comprises the following steps that (1), the target filtering reflection spectrum of the one-dimensional photonic crystal filter is designed by utilizing the digital signal processing, the optimal expression coefficient of the target filtering reflection spectrum in a numeric field is calculated, and the optimal expression of the target filtering reflection spectrum in the numeric field is obtained; (2), a layer peeling algorithm is adopted to calculate the interface reflection coefficients of the dielectric layers of the one-dimensional photonic crystal filter; (3) the refractive indexes of the dielectric layers of the one-dimensional photonic crystal filter are calculated according to the interface reflection coefficients, obtained in the second step, of the dielectric layers of the one-dimensional photonic crystal filter. The method for designing the one-dimensional photonic crystal filter directly calculates and obtains the refractive indexes of the dielectric layers of the needed one-dimensional photonic crystal filter according to the needed filtering spectrum based on the filtering spectrum, and has the advantage of being simple and direct in calculation, the optimal expression of the target filtering reflection spectrum in the numeric field can be conveniently obtained, and therefore the one-dimensional photonic crystal filter obtained through designing is excellent in filtering performance.

The invention discloses an active power filter device for specific sub-harmonics. The device is applicable to a circuit comprising a system power supply and a harmonic load, and comprises a transformer, an inverter and a modulation signal generation module; the primary winding of the transformer is connected in series with the harmonic load, and connected to a circuit formed by the system power supply and the harmonic load; the modulation signal generation module is used for generating modulation signals for the inverter, and the modulation signals are generated by superposing primary fundamental wave current amplified reversely and primary specific n sub-harmonic current amplified positively; the inverter is used for converting direct current signals into alternating current signals having the same frequency as the modulation signals and injecting the alternating current signals into the secondary winding of the transformer, so that the primary fundamental wave equivalent impedance of the transformer is Z1+(1-a)Zm, and the primary n sub-harmonic equivalent impedance of the transformer is nZ1+(1+beta)nZm. Based on fundamental wave and specific sub-harmonic flux hybrid control, the primary side of the transformer is short-circuited relative to fundamental waves and open-circuited relative to specific sub-harmonics.

A design method for a reverse class-F power amplifier based on 3 / 4 spiral virtual ground structure comprises the steps as follows: confirming the work frequency of the power amplifier and obtaining the harmonic frequency of a harmonic control circuit; changing the dimension of the 3 / 4 spiral structure and designing corresponding harmonic control circuit according to the needed harmonic frequency in the design; inserting the harmonic control circuit into the output end of a power amplifier tube and designing the input-output matching circuit of a power amplifier circuit; and designing the input-output parameters of the power amplifier through measurement, and adjusting the length of a magic line to reach the maximum work efficiency. The amplifier designed by adopting the method comprises an input matching circuit, a power amplifier circuit, a harmonic control circuit and an output matching circuit which are connected onto a microstrip line in sequence; the harmonic control circuit comprises a 3 / 4 spiral virtual ground structure arranged below the microstrip line, and the virtual ground structure is formed by a complete spiral virtual ground structure cut by 1 / 4; and the 3 / 4 spiral virtual ground structure comprises two spiral structure parts connected with each other and in axial symmetry along the microstrip line.

The invention discloses a method for filtering parallel Gauss particles based on quasi-Monte Carlo sampling, belongs to the technical field of signal processing, and mainly solves the problem that the prior filtering method based on the quasi-Monte Carlo sampling reduces performance for estimating state of a nonlinear dynamic system when the sample number is less. The filtering method comprises: a generating step of a parallel quasi-Monte Carlo sequence and a filtering step of the parallel Gauss particles, namely adopting a generating method for jumping the quasi-Monte Carlo sequence first, and generating a P quasi-Monte Carlo random sample sequences required by subsequent filtering in parallel; then converting each quasi-Monte Carlo random sample sequence into a quasi-Gauss sample sequence obeying designated Gaussian distribution through P execution units; at the same time, updating the time and the state for the state of the nonlinear dynamic system; and finally carrying out comprehensive treatment on all the execution units to finish filtering the nonlinear dynamic system. The method has the advantages of high filtering performance and good real-time property, and can be used in the field of signal processing, automation and artificial intelligence.