27 results about "Feedforward filter" patented technology

The present invention discloses a hybrid active noise cancellation superposition method based on a neural network, including: W f W(n) or / and W b (n) is implemented by a linear filter FIR or IIR, and the filter parameters are iterated according to the LMS or RLMS algorithm. Since S(z) is non-linear, S’(z) can still be estimated by a neural network, and the parameters of the neural network are iteratively solved according to the BP algorithm. For the non-linear link existing in the path of the original noise propagation from the A-bit noise source to the B-bit speaker, such as the case where the original noise is too large, resulting in non-linearity of the device, if the feedforward filter can use a non-linear filter, it can better control the noise; for the non-linear link existing in the control channel from the C-bit error microphone to the B-bit speaker, such as the non-linearity of the error microphone or the non-linearity of the amplifier, if the feedback filter can use a non-linear filter, it can better control the noise.

The invention belongs to the field of aircraft noise reduction, and discloses an aircraft cabin active noise reduction control method and related equipment, and the method comprises the steps: collecting the vibration, rotating speed, cabin sound pressure and other multi-modal data of an aircraft in real time, and carrying out the flight state classification through a pre-trained multilayer neural network model; and the current accurate working condition is determined by combining state machine conversion logic, and finally, the parameters of the feed-forward filter are dynamically adjusted through a minimum mean square algorithm to realize noise reduction. According to the method, the high-frequency component suppression capability in a multi-source coupled noise scene is effectively improved, the noise reduction efficiency is kept stable in the whole flight period, the convergence speed is improved compared with a traditional scheme, and the local optimum problem is avoided through a state self-adaptive feed-forward mechanism.

The present invention relates to an active damping control system for suppressing vehicle speed oscillation, comprising a signal acquisition module, a vehicle drive-brake demand torque calculation module, a feedforward control module, a suppression mode judgment module, a feedback control module, an output torque coordination control module, and a motor torque control module. The present invention also relates to an active damping control method for suppressing vehicle speed oscillation, comprising the steps of: collecting accelerator pedal opening information, brake pedal opening information, wheel speed information, rotation speed information, gear information, vehicle mass information, vehicle speed information, and vehicle control mode; obtaining the vehicle drive-brake demand torque; obtaining the feedforward filtering torque; feedforward regulation control; judging and outputting the vehicle speed suppression mode; rotation speed feedback loop control, vehicle speed feedback loop control; output coordination control; motor target torque. The present invention effectively improves the vehicle speed oscillation suppression control effect and driving experience; and effectively improves oscillation in the case of drastic load changes in commercial use.

The invention discloses a Bluetooth headset noise reduction coefficient generation method, which is applied to an active noise reduction path, and comprises the following steps: acquiring multiple groups of preset filter coefficients of a feed-forward controller through offline modeling, and acquiring a feed-forward filter coefficient of the feed-forward controller corresponding to an initial wearing state; calculating and obtaining a suboptimal transfer function of a feed-forward controller corresponding to the current wearing state by counting parameter data of a preset frame number corresponding to the feed-forward filter coefficient; obtaining the optimal transfer function of the feedforward controller coefficient according to the suboptimal transfer function and a plurality of groups of preset filter coefficients; and finally obtaining a second-order section filter coefficient corresponding to the optimal transfer function. The problem of poor noise reduction effect caused by the fact that the coefficient is fixed by the current wearing state of the current Bluetooth earphone can be avoided.

The invention provides a sequence detection device which comprises a feed-forward filter and a sequence detection circuit. The feed-forward filter is used for processing a received signal to generate an equalized signal. The sequence detection circuit is used for carrying out sequence detection on the equalization signal to generate and output a symbol sequence. The sequence detection circuit comprises a region estimation circuit and a grid selection circuit. The region estimation circuit is configured to classify each of a plurality of samples included in the equalized signal into one of a plurality of regions. The grid selection circuit is used for selecting one of a plurality of grid schemes to perform branch metric calculation according to the area estimation results of two of the plurality of samples output by the area estimation circuit.

The embodiment of the invention provides an active noise reduction method, electronic equipment, a storage medium and a program product. The method comprises the following steps: acquiring environment noise time domain data acquired by a feedforward microphone, ear canal noise time domain data acquired by a feedback microphone and loudspeaker time domain data output by a loudspeaker; determining acoustic coupling data based on the environmental noise time domain data, the ear canal noise time domain data and the loudspeaker time domain data; the acoustic coupling data is used for representing the coupling condition of the earphone and the auditory meatus; determining a compensation transfer function according to the acoustic coupling data; injecting the compensation transfer function into a feed-forward filter and a feedback filter to obtain a compensated feed-forward filter and a compensated feedback filter; based on the compensated feedforward filter and the compensated feedback filter, processing the environment noise time domain data and the ear canal noise time domain data to obtain reverse-phase sound waves; the anti-phase sound waves are used for active noise reduction. The method is used for improving the active noise reduction effect.

A sequence detection device includes a feed-forward filter, a feedback filter, a combining circuit, a decision circuit, and a sequence detection circuit. The feed-forward filter processes a received signal to generate a first equalized signal. The feedback filter processes a symbol decision signal to generate a second equalized signal. The combining circuit combines the first and second equalized signals to generate a sample signal. The decision circuit performs hard decision upon the sample signal to generate the symbol decision signal. The sequence detection circuit performs sequence detection upon the first equalized signal, and includes a region estimation circuit and a trellis selection circuit. The region estimation circuit is independent of the decision circuit, and categorizes each sample of the sample signal into one region. The trellis selection circuit selects one trellis scheme for branch metric calculation according to region estimation results output from the region estimation circuit.

The present invention discloses a signal processing optimization method and system for training BCJR using the least squares method, which belongs to the technical field of short-distance high-speed optical communication signal processing. At the receiving end, the signal is processed by a feedforward filter FFE and then processed by a feedback filter PF; the output of the feedback filter PF and the training sequence are used to perform channel estimation using the least squares LS algorithm; based on the channel response estimated by the least squares LS algorithm, the BCJR algorithm is used for soft decision decoding to calculate the maximum a posteriori probability MAP of each symbol; based on the maximum a posteriori probability MAP of each symbol, the bit-by-bit log-likelihood ratio is calculated; and the signal is restored based on the obtained log-likelihood ratio. The present invention can solve the problem that the performance of the traditional BCJR algorithm is significantly reduced when facing severe inter-symbol interference. By optimizing the coefficients of the BCJR algorithm, it is made closer to the original channel response, thereby effectively reducing the bit error rate and improving system performance.

Aspects relate to mechanisms for implementing a frequency domain phase constraint at a decision feedback equalizer (DFE) of a wireless device. The DFE includes a feedforward filter (FFE) operating in a frequency domain. The FFE further includes an accumulator having a main-tap constraint configured to maintain a constant value of a main-tap of the DFE and a phase constraint module configured to maintain a constant phase of a frequency bin associated with the main-tap to produce a timing error used for timing recovery of a receiver of the wireless device.

The present invention discloses a coefficient initialization system and method for a decision feedback equalizer, comprising: a feedforward filter and a feedback filter, wherein an input signal is converted into a digital signal through the feedforward filter; the digital signal is subjected to coefficient convergence operation through an update component; the update component transmits the digital signal after operation to a decision module to judge the signal; an initialization control module can receive the coefficient signal of the feedforward filter, the initialization control module is provided with a timer, the coefficient convergence time can be limited, and the coefficient initialization operation can be performed, and the initialized coefficient signal is transmitted to the feedforward filter and the feedback filter, and at the time interval when the timer expires, the feedforward filter and the feedback filter are reinitialized in an interlaced manner, and the device is provided with a timer in the initialization control module, and the coefficient convergence cycle period is set by the timer, so as to realize the adaptive convergence of the decision feedback equalizer.

An audio device comprising an interface with a receiver with an output sound based on an audio output signal; a feed-forward microphone; a distributor for provision of a first feed-forward input signal and a second feed-forward input signal; a feed-back microphone; a feed-forward filter for provision of a feed-forward output signal; an intermediate filter module, wherein the intermediate filter module comprises a hear-through filter based on the second feed-forward input signal and wherein the intermediate filter module is configured to provide the intermediate signal based on the hear-through signal and the audio input signal; a first mixer for provision of an error signal based on the feed-back input signal and the intermediate signal; a feed-back filter signal based on the error signal; and a second mixer for provision of the audio output signal based on the feed-forward output signal and the feed-back output signal.

The invention relates to an active noise reduction method and hearing aid equipment. The method is applied to a hearing aid device, and specifically comprises the following steps: in an active noise reduction mode, obtaining current frequency response information and original target frequency response information of a feed-forward filter; determining the current noise reduction depth of the hearing aid device according to the current frequency response information and the original target frequency response information; and if the current noise reduction depth does not reach the target noise reduction depth, updating the active noise reduction parameter of the hearing aid device, re-determining the current frequency response information of the feed-forward filter according to the updated active noise reduction parameter, and returning to execute the step of determining the current noise reduction depth of the hearing aid device according to the current frequency response information and the original target frequency response information. Wherein the target noise reduction depth is determined based on the hearing loss information of the target object; and if the current noise reduction depth reaches the target noise reduction depth, performing active noise reduction in the hearing aid device according to the active noise reduction parameter of the hearing aid device. By adopting the method, the active noise reduction effect can be improved.

The invention discloses a signal processing optimization method and system for training a BCJR by using a least square method, and belongs to the technical field of short-distance high-speed optical communication signal processing. At a receiving end, the signal is processed by a feed-forward filter (FFE) and then processed by a feed-back filter (PF); channel estimation is carried out through a least square (LS) algorithm by using the output of a feedback filter (PF) and a training sequence; based on channel response estimated by a least square (LS) algorithm, soft decision decoding is carried out by using a BCJR algorithm, and the maximum posterior probability MAP of each symbol is calculated; according to the maximum posterior probability MAP of each symbol, calculating a bit-by-bit log-likelihood ratio; and recovering the signal according to the obtained log-likelihood ratio. The method can solve the problem that the performance of a traditional BCJR algorithm is remarkably reduced when the traditional BCJR algorithm faces serious inter-symbol interference. By optimizing the coefficient of the BCJR algorithm, the BCJR algorithm is closer to the original channel response, so that the bit error rate is effectively reduced, and the system performance is improved.

The invention discloses a recording method and related equipment, and the method comprises the steps: collecting a first audio in an environment through a feed-forward microphone of earphone equipment, and collecting a second audio in the environment through a feedback microphone of the earphone equipment; performing high-frequency enhancement on the first audio through a feed-forward filter to obtain an enhanced audio; performing audio playing based on the enhanced audio; and obtaining a first recording audio based on the second audio and the audio played by the earphone device and collected by the feedback microphone. According to the invention, the recorded audio can restore the actual listening feeling of the user.

The application discloses an earphone noise reduction method and device, electronic equipment and a computer readable storage medium, and is applied to an earphone. The earphone noise reduction method comprises the following steps: detecting a noise signal direction corresponding to an external noise signal; determining corresponding feedforward filter coefficients according to the noise signal direction; and performing earphone noise reduction according to the feedforward filter coefficients. The application solves the technical problem of poor active noise reduction effect of the earphone.

The invention provides a sequence detection device and a sequence detection method. The sequence detection device comprises a feed-forward filter, a feedback filter, a combining circuit, a decision circuit and a sequence detection circuit. The feed-forward filter processes a received signal to generate a first equalized signal. The feedback filter processes the symbol decision signal to generate a second equalized signal. The combining circuit combines the first and second equalized signals to generate a sample signal. The decision circuit performs a hard decision on the sample signal to generate the symbol decision signal. The sequence detection circuit performs sequence detection on the first equalization signal and comprises a region estimation circuit and a grid selection circuit. The region estimation circuit classifies each sample in the sample signal into a region. The grid selection circuit selects a grid scheme for branch metric calculation according to the area estimation result output by the area estimation circuit.

The embodiments of the present invention disclose a demodulation method, device, system, and electronic device for Bluetooth GFSK signals. The method includes: performing differential phase detection on the signal phase and the noise phase to obtain a plurality of differential phase signals; selecting a plurality of differential phase signals and combining them with a noise differential combined signal to obtain a combined differential signal; normalizing the coefficients of a Gaussian filter function to obtain a Gaussian discrete function; obtaining a feedforward filter function and a feedback filter function based on the Gaussian discrete function; performing a convolution operation on the feedforward filter function and the combined differential signal to obtain a first convolution value, and performing a convolution operation on the feedback filter function and the signal phase at a previous moment to obtain a second convolution value; subtracting the second convolution value from the first convolution value to obtain a decision signal; and determining the phase sign of the GFSK signal based on the sign of the decision signal. Through the above-mentioned method, the embodiments of the present invention can improve the demodulation sensitivity of the GFSK signal without increasing the computational complexity.

A sequence detection device includes a feed-forward filter and a sequence detection circuit. The feed-forward filter processes a received signal to generate an equalized signal. The sequence detection circuit performs sequence detection upon the equalized signal to generate and output a symbol sequence. The sequence detection circuit includes a region estimation circuit and a trellis selection circuit. The region estimation circuit categorizes each of samples included in the equalized signal into one of regions. The trellis selection circuit selects one of trellis schemes for branch metric calculation according to region estimation results of two of the samples output from the region estimation circuit.

A sequence detection device includes a feed-forward filter and a sequence detection circuit. The feed-forward filter processes a received signal to generate an equalized signal. The sequence detection circuit performs sequence detection upon the equalized signal to generate and output a symbol sequence. The sequence detection circuit includes a region estimation circuit and a trellis selection circuit. The region estimation circuit categorizes each of samples included in the equalized signal into one of regions. The trellis selection circuit selects one of trellis schemes for branch metric calculation according to region estimation results of two of the samples output from the region estimation circuit.

The invention discloses a feed-forward voltage sampling double-compensation active electromagnetic interference filter based on impedance mismatch and a design method thereof, and belongs to the field of power electronics, and the method comprises the steps: constructing an equivalent circuit according to an impedance mismatch principle, and determining the circuit topology of the filter according to the equivalent circuit; determining a resistance value and a capacitance value according to the cut-off frequency and impedance of the RC circuit; determining an amplification factor according to the unity gain bandwidth of the operational amplifier; determining the turn ratio of the transformer according to the impedance mismatch principle and the amplification factor; determining a resistance value of the voltage injection module according to the damping effect; and determining a capacitance value of the current injection module according to the electric leakage constraint. According to the method, the interference suppression frequency band can be widened, the insertion loss can be improved, impedance matching can be achieved, meanwhile, the number of turns on the main power side of the transformer can be reduced due to the parameter design characteristic of the feed-forward filter, the size of the filter is greatly reduced, and the requirements of a new generation of power electronic equipment for the small size and the high suppression effect are met.

The application discloses a feedforward voltage sampling double-compensation active electromagnetic interference filter based on impedance mismatch and a design method thereof, and belongs to the field of power electronics, which comprises the following steps: constructing an equivalent circuit according to an impedance mismatch principle, determining a circuit topology of the filter according to the equivalent circuit, determining resistance and capacitance values according to a cutoff frequency and impedance of an RC circuit, determining an amplification multiple according to a unit gain bandwidth of an operational amplifier, determining a turns ratio of a transformer according to the impedance mismatch principle and the amplification multiple, determining a resistance value of a voltage injection module according to a damping effect, and determining a capacitance value of a current injection module according to a leakage constraint. The method disclosed by the application can widen an interference suppression frequency band, improve an insertion loss, realize impedance matching, and greatly reduce the volume of the filter due to the parameter design characteristics of the feedforward filter, thereby adapting to the demand of a new generation of power electronic equipment for small volume and high suppression effect.

The present invention discloses a method for quickly updating a feedforward filter of a noise reduction headset and its related components. The method includes: respectively obtaining a first electrical signal and a second electrical signal of the noise received by a feedforward microphone and an in-ear microphone and respectively calculating the complex spectrum of the Fourier transform to calculate a transfer function and obtain a first transfer function; adjusting the gain coefficient or frequency domain response function of the feedforward filter and recalculating the transfer function to obtain a second transfer function, and calculating a theoretical optimal filter; comparing a built-in filter with the theoretical optimal filter, taking the built-in filter closest to the theoretical optimal filter as a preferred filter and calculating the corresponding gain. The present invention obtains the transfer function in the corresponding state by changing the state of the feedforward filter, thereby calculating the theoretical optimal filter, screening out the preferred filter closest to the theoretical optimal filter from the built-in filter and calculating the gain, the whole process has a short update time, and the preferred filter is obtained quickly.

A high-magnitude rotational inertia simulation method for a full-power test bench of a speed-sensorless wind turbine, comprising the following steps: determining the rotational inertia J of the low-speed side of the wind turbine based on the actual wind turbine model to be simulated by the full-power test bench t ; measuring the rotational inertia J of the full-power test bench itself s ; measuring the communication time delay τ of the full-power test bench, determining the time-delay order k0 based on the communication time delay τ; determining the k0-order filter parameter α of the torque compensation loop according to the time-delay order k0 d ; constructing a feedforward filter transfer function according to the above parameters. This method corrects the aerodynamic torque of the test bench by adding a feedforward filter to the aerodynamic torque calculation loop, suppressing the acceleration deviation of the existing dynamic simulation experiment system, thereby eliminating the high-frequency oscillation of the rotational speed.

A computer-implemented method for automatically designing a feedforward filter that is optimized for an audio transparency mode of operation of an audio device, the audio device comprising the feedforward filter and a feedback filter, the method comprising optimizing the feedforward filter for the audio transparency mode of operation, wherein optimization of the feedforward filter is dependent on the feedback filter.

A system configured to perform adaptive active noise cancellation (ANC) processing. Specifically, the system may adaptively determine a feed-forward ANC filter by maximizing a ratio of A:B, where A corresponds to microphone-ear coherence and B corresponds to microphone-microphone coherence between the feed-forward microphones. By maximizing this ratio, the system may determine weighted gain values used to combine the feed-forward microphone signals. In addition, the system may (i) apply a fixed feed-forward ANC filter profile selected based on a geometry of the device and a generalized ear response, (ii) monitor a secondary path to select from a plurality of feed-forward ANC filter profiles based on an individual user's ear response, or (iii) adaptively update the feed-forward ANC filter based on a feedback microphone signal.

The present disclosure relates to an active noise reduction method for headphones, an active noise reduction system and noise reduction headphones. The headphones include a speaker, a feedforward noise reduction path including a feedforward microphone and a feedforward IIR filter group in sequence, and an in-ear microphone. The active noise reduction method includes connecting a feedforward adaptive FIR filter downstream of the feedforward IIR filter group; turning on the feedforward adaptive FIR filter when the external environmental noise is large, calibrating the echo filter parameters downstream of the feedforward adaptive FIR filter, and determining the first filter coefficient for configuring the feedforward adaptive FIR filter based on the first noise signal obtained through the feedforward noise reduction path and the residual noise signal obtained by the in-ear microphone, thereby optimizing the active noise reduction effect. The active noise reduction method disclosed in the present disclosure can adapt to the ear canal shape of each wearer, provide personalized ANC parameters, improve the problem of poor consistency of ANC effect caused by differences in earphone cavities, and skip the production line calibration step to save earphone production costs.

This utility model discloses a dual-feed over-ear noise-canceling headphone, relating to the field of over-ear noise-canceling headphone technology. It includes an earphone shell, with a speaker mounting shell internally connected to it. A speaker is mounted on the right side of the speaker mounting shell, and a silicone pad is attached to the outer side of the speaker mounting shell. A bracket is rotatably connected to the outside of the earphone shell, and an extension bracket is slidably connected to the inner side of the bracket. A noise-canceling component is mounted on the other side of the speaker mounting shell. This dual-feed over-ear noise-canceling headphone uses a feedforward microphone and a feedback microphone to collect external ambient noise and residual noise within the earcups, respectively. The earphone shell also houses a speaker for playing audio signals and processed noise-canceling signals. The headphone includes a noise-canceling control circuit, including a feedforward filter and a feedback filter, which process the noise signals collected by the feedforward and feedback microphones, respectively.