63 results about "Quantization (signal processing)" patented technology

A coherent receiver 1 assigns a first transmission signal to first transmission polarization, assigns a second transmission signal to second transmission polarization, and receives a quadrature multiplexed signal formed by applying quadrature multiplexing to the first transmission polarization and the second transmission polarization. The coherent receiver includes a detection means 10 for detecting the first transmission polarization and the second transmission polarization according to prescribed first reception polarization and second reception polarization and obtaining a first detected signal and a second detected signal; a quantization means 20 for quantizing the first detected signal and the second detected signal and obtaining a first quantized signal and a second quantized signal; and a signal processing means 30 for, when filtering the first quantized signal and the second quantized signal using a prescribed filtering control algorithm to form a first demodulated signal and a second demodulated signal respectively, adjusting filter coefficients of the filtering control algorithm according to the first quantized signal and the second quantized signal and the first demodulated signal and the second demodulated signal, and outputting the first demodulated signal and the second demodulated signal to a first output terminal and a second output terminal, respectively.

The invention relates to the field of digital signal processing, in particular to a quantization method and system based on neural network difference. Specifically include: training the network model, initializing the network model; using the training parameter expression form of the network model as the change amount based on the initialization model parameters; clustering and quantifying the change amount of the training parameter expression form of the network model, The corresponding compact representation of the network model is obtained. The invention solves how to improve the performance of the neural network model as much as possible under the condition of low-bit quantization under the condition that the pre-training model exists.

The invention provides a signal processing device, method and analog-to-digital converter. Wherein, the signal processing device includes: an input terminal for receiving an analog signal; a quantization module for converting the analog signal into a digital signal, and the digital signal includes a multi-digit digital output code; a randomization module for receiving the h-bit digital output code randomly flipped by thermal noise in the multi-bit digital output code, and using the h-bit digital output code to randomize the quantization path, where h is a positive integer.

The invention discloses a digital receiver system based on FPGA and a radio frequency analog-to-digital conversion method. The system comprises a power distribution module, a reference signal generation module, a radio frequency analog-to-digital conversion module and a digital signal processing module. The radio frequency analog-to-digital conversion method comprises the following steps of: dividing a radio frequency signal in any frequency band into (n-1) paths; the digital reference signal generated by FPGA is filtered to obtain the analog reference signal. A digital signal is obtained by analog-to-digital conversion of a radio frequency signal in any frequency band. The direct RF analog-to-digital conversion of the analog signal can be realized by the RF analog-to-digital method. For RF signals of different frequency bands and different standards, the whole signal processing process from RF quantization to baseband demodulation is carried out in the digital domain of FPGA, which eliminates the analog local oscillator, mixer and IF filter, and effectively avoids the loss and distortion of analog devices.

The invention discloses an accurate and efficient realization method and device for linear frequency modulation signal autocorrelation operation. The method includes: quantization step, buffer step, filter step and output step. Each unit of the device for accurately and efficiently realizing the autocorrelation operation of the linear frequency modulation signal corresponds to the method. The present invention provides an accurate and efficient implementation method and device for the autocorrelation operation of the chirp signal. The biggest feature of the method is to use the FIR filter bank to realize the autocorrelation operation of the chirp signal. The overall effect is parallel processing through the FIR filter The architecture can realize simultaneous output of autocorrelation results with different delays. The FIR filter can be realized by using the FFT algorithm, and the operation speed is relatively fast. The method also has the advantages that the processing signal is a digital quantity, the calculation accuracy is high and there is no drift; it is suitable for high-speed parallel multi-channel digital signal processing; the memory can be used efficiently, and in some application scenarios, the FIR filter bank can be stored in a memory such as ROM. coefficient, the implementation cost is small.