8 results about "Infinite impulse response" patented technology

The application discloses a decision feedback equalizer for PAM4 encoding band infinite impulse response filter, belongs to the field of signal processing, and comprises a high-pass filter and a linear equalizer, a decision feedback adder, and an infinite impulse response filter coefficient feedback control module; noise filtering is performed on a high-speed signal, low-frequency attenuation of the high-speed signal is compensated, a tailing effect in a frequency response is eliminated, high-speed inter-symbol interference is suppressed and eliminated, and an offline post-cursor in the frequency response is fed back and compensated; meanwhile, aiming at PAM4 encoded data, an RLM (Level separation mismatch ratio) calibration module and a clock data recovery circuit are added, which are used for improving the speed of clock receiving and recovery, and due to the addition of more high-speed modules, the load of the equalization adder is increased; the application uses a layout planning of a double-difference six-port mutual inductance bridge type (T-coil) inductive network, and bandwidth is expanded by using inductive peaking technology.

Technologies of a device-based Fast Presence Detection (FPD) for a contactless sleep-tracking device are described. One method of a sleep-monitoring device includes receiving radar data from a radar unit. The radar data includes i) first data representing a breathing waveform associated with a user, ii) a first set of range values, and iii) a first set of confidence values associated with the first data. The method determines absolute magnitude values, first infinite impulse response (IIR) values using the first set of range values, and second IIR values using the first set of confidence values. The method determines a first event representing the user located in a first region using the absolute magnitude values and the first and second IIR values. The method sends an indication of the first event to a cloud service that causes one or more devices in the environment to perform one or more actions.

This invention discloses a disturbance compensation method based on a cyclic adaptive identification radial basis function neural network. The method includes: first, introducing an error transformation function and calculating the instantaneous quantity, then averaging it to obtain the average instantaneous quantity. Next, acquiring the q-axis current, first filtering it using an infinite impulse response filter, then filtering it a second time using a moving average filter, and constructing a trend curve of the q-axis current. Differentiating this trend curve using a backward differential, and then performing phase response compensation. Combining this compensation signal with the identification rate to generate the operating condition identification quantity. Finally, constructing a cyclic adaptive identification radial basis function neural network, using the average instantaneous quantity and the operating condition identification quantity as inputs, injecting the compensation quantity generated by the network into the current loop to achieve closed-loop compensation control of the system. By using this invention, efficient feedforward compensation for periodic disturbances is achieved. This invention can be widely applied in the field of permanent magnet synchronous motor control technology.

An infinite impulse response filter includes at least one N x N multimode optical coupler and one or more mode scrambling loops. The at least one N x N multimode optical coupler has N input ports and N output ports. One of the N input ports is coupled to a laser source that generates a light beam. The at least one N x N multimode optical coupler splits the light beam into a plurality of sub-beams that are partitioned in an amplitude domain across the N output ports. One of the N output ports is coupled to a detection system. The one or more mode scrambling loops include a mode scrambler and couple one of the N output ports to one of the N input ports. The one or more mode scrambling loops produce temporal incoherence and spatial incoherence that reduce peak power and speckle contrast of the light beam.

This invention discloses a method, apparatus, device, and medium for PPG spectrum enhancement based on adaptive weighting. The method includes: performing a Fast Fourier Transform on the PPG signal after removing motion artifacts to generate continuous spectral frame data, and calculating the signal-to-noise ratio (SNR) of the spectral frame data; quantizing the user's current motion level based on the motion acceleration component signal, and dynamically determining the fusion weight between the current spectral frame data and historical spectral frame data based on the calculated SNR of the current spectral frame data; fusing the current spectral frame data and historical spectral frame data using an infinite impulse response recursive weighted accumulation algorithm based on the determined fusion weight between the current spectral frame data and historical spectral frame data to obtain enhanced current spectral frame data; and detecting the user's current heart rate based on the enhanced current spectral frame data. This invention denoises the PPG signal with low computational overhead, improving the accuracy of user heart rate detection.

The invention relates to the technical field of deep learning and neural network model compression, and discloses a model filter compression method based on gradient guidance and terminal equipment in order to solve the technical problems of high reasoning cost and difficult deployment of an FSMN model caused by a high-order FIR filter. A trained FSMN model is obtained, the model comprises at least one high-order finite impulse response filter layer, the weight of the high-order finite impulse response filter layer is defined as the step b, and the gradient of a final loss function of the FSMN model relative to the gradient is determined and calculated; and step c, searching and determining an infinite impulse response filter of which the order is lower than that of the FIR filter based on the guidance of the gradient, and defining the weight of the infinite impulse response filter as step d, and replacing the generated compressed FSMN model with the infinite impulse response filter. The problem that the FSMN model is high in reasoning cost is solved, and the parameter quantity and the calculation complexity of the model are remarkably reduced.