In one illustrative embodiment, an equalizer includes: a shift register, an array of multipliers, an array of multiplexers, and a summer. The shift register provides receive signal samples at each tap. Each multiplier in the array multiplies one of said receive signal samples by a respective coefficient to produce a product, with at least one of said multipliers coupled to a fixed tap. Each multiplexer in the array supplies an associated one of said multipliers with a receive signal sample from a selectable tap. The summer sums the products to produce a filtered output signal. To reduce hardware requirements, coefficient multipliers may be multiplexed to a reduced set of taps, and the dynamic range of the coefficients may be increased by overlapping the sets for different multipliers. Methods of tap selection and coefficient adaptation are disclosed.

The configuring of a micromirror to suppress a resonance of the micromirror. As part of the configuring process, the micromirror is subjected to multiple actuation frequencies, and the micromirror response is measured in response to at least some of these actuation frequencies. A resonant frequency of the micromirror is then determined using at least some of the measured mechanical responses. Then, depending on this determined resonant frequency of the micromirror, notch filter parameters are selected. There is more than one possibility for notch filter parameters, where the selected possibility depends on the determined resonant frequency. The notch filter is then configured with the selected notch filter parameters.

Disclosed are implementations, including a method that includes obtaining measurement samples relating to electrical operation of an electric motor drive providing power to an electric motor, deriving, based on the samples, instantaneous estimates for parameters characterizing speed and/or position of the motor according to an optimization process based on a cost function defined for the samples, and applying a filtering operation to the instantaneous estimates to generate filtered values of the motor's speed and/or position. The filtering operation includes computing the filtered values using the derived instantaneous estimates in response to a determination that a computed convexity of the cost function is greater than or equal to a convexity threshold value, and/or applying a least-squares filtering operation to the derived instantaneous estimates and using at least one set of previous estimates derived according to the optimization process applied to previous measurement samples.

A method and a device for updating a coefficient vector of a finite impulse response filter are provided. The update method includes: obtaining an updated step-size diagonal matrix for a coefficient vector of the FIR filter; and obtaining an updated coefficient vector of the FIR filter based on the updated step-size diagonal matrix.

Disclosed are implementations, including a method that includes obtaining measurement samples relating to electrical operation of an electric motor drive providing power to an electric motor, deriving, based on the samples, instantaneous estimates for parameters characterizing speed and/or position of the motor according to an optimization process based on a cost function defined for the samples, and applying a filtering operation to the instantaneous estimates to generate filtered values of the motor's speed and/or position. The filtering operation includes computing the filtered values using the derived instantaneous estimates in response to a determination that a computed convexity of the cost function is greater than or equal to a convexity threshold value, and/or applying a least-squares filtering operation to the derived instantaneous estimates and using at least one set of previous estimates derived according to the optimization process applied to previous measurement samples.

An inference processing apparatus includes an input data storage unit that stores pieces X of input data, a learned NN storage unit that stores a piece W of weight data of a neural network, a batch processing control unit that sets a batch size on the basis of information on the pieces X of input data, a memory control unit that reads out, from the input data storage unit, the pieces X of input data corresponding to the set batch size, and an inference operation unit that batch-processes operation in the neural network using, as input, the pieces X of input data corresponding to the batch size and the piece W of weight data and infers a feature of the pieces X of input data.

According to one embodiment, a sensor module includes at least one sensor and at least one switch. The sensor includes a first piezoelectric element. The first piezoelectric element includes a first electrode. The first piezoelectric element is set with a resonance frequency to resonate at a vibration frequency of a detection target. The switch includes a second piezoelectric element. The second piezoelectric element includes a second electrode connected to the first electrode and a third electrode electrically separated from the second electrode.

The present invention is a computationally-efficient compensator for removing nonlinear distortion. The compensator operates in a digital post-compensation configuration for linearization of devices or systems such as analog-to-digital converters and RF receiver electronics. The compensator also operates in a digital pre-compensation configuration for linearization of devices or systems such as digital-to-analog converters, RF power amplifiers, and RF transmitter electronics. The adaptive Volterra compensator effectively removes nonlinear distortion in these systems by implementing an adaptive background algorithm to periodically update actual filter coefficients to maintain optimal performance in operating conditions varying over time (e.g., temperature, frequency, signal level, and drift); or both. The xadaptive background algorithm calculates the optimal nonlinear filter coefficients to reduce nonlinear distortion.

The present invention relates to a method and an apparatus for processing a signal, which are used for effectively reproducing a multimedia signal, and more particularly, to a method and an apparatus for processing a signal, which are used for implementing filtering for multimedia signal having a plurality of subbands with a low calculation amount.

To this end, provided are a method for processing a multimedia signal including: receiving a multimedia signal having a plurality of subbands; receiving at least one proto-type filter coefficients for filtering each subband signal of the multimedia signal; converting the proto-type filter coefficients into a plurality of subband filter coefficients; truncating each subband filter coefficients based on filter order information obtained by at least partially using characteristic information extracted from the corresponding subband filter coefficients, the length of at least one truncated subband filter coefficients being different from the length of truncated subband filter coefficients of another subband; and filtering the multimedia signal by using the truncated subband filter coefficients corresponding to each subband signal and an apparatus for processing a multimedia signal using the same.

The invention provides a method and apparatus for filtering a temporal signal. A target magnitude frequency response H.sub.T(f) is specified (101,201) of frequency f in terms of a column vector l of K weights l.sub.k where log H.sub.T(f)=l.sup.TW(f) and W(f) is a column vector of K magnitude basis functions W.sub.k(f). A constrained frequency response H.sub.c(f) is computed (102,214) defined by log H.sub.c(f)=g.sup.TV(f) , where V(f) is a column vector of N constrained basis functions V.sub.n(f) for which each exp g.sub.nV.sub.n(f) satisfies a constraint preserved by concatenation, and g is a column vector of N coefficients satisfying a matching criterion between l.sup.TW(f) and g.sup.TV(f). An input temporal signal is received (103,212) and filtered (104,210) with the constrained frequency response H.sub.c(f) to form a filtered temporal signal; and the filtered temporal signal is output (105,211).