32results about How to "Reduce estimation complexity" patented technology

The invention provides a data-guiding channel estimation method for OFDM systems, and an implementation device thereof, which belong to the technical field of digital information transmission. The method demodulates and recovers frequency-domain data symbols at the receiving end of an OFDM system and performs channel estimation. The method comprises the steps of utilizing a frame structure to obtain initial channel estimation result, building a time-domain signal frame and the cyclic convolution of channel impulse response, demodulating frequency-domain data symbols of a current frame, and performing data-guiding channel estimation and post processing. The device comprises an initial channel estimation module, a channel estimation gating module, a cyclic convolution building block, a frequency-domain data block demodulation module, a data-guiding channel estimation module, a channel estimation post-processing module and a control module. The method and the device use hidden informationcarried by frequency-domain data sent by the OFDM system in symbol modulation, as well as initial channel estimation provided by training information, completes receiving the channel estimation of all signal frames, and can reduce complexity, improve precision and increase the utilization ratio of frequency spectrum, mobile performance and other effects.

The invention discloses a millimeter wave sparse array plane channel estimation method based on a deep learning network, which takes millimeter wave channel sparse characteristics as priori information to train a designed full-connection deep neural network for millimeter wave array plane communication channel estimation. Firstly, a full-connection phase shifter network is adopted, and isotropic analog transceivers are designed by configuring phase uniform distribution of phase shifters; and the obtained channel sparse information and the designed optimal digital estimator are taken as training data of the full-connection deep learning network. For the sparse channel under each signal-to-noise ratio, the sparse information of the channel is input into a network to obtain a corresponding digital estimator, thereby obtaining a channel estimation result. The sparse channel estimator provided by the invention can reduce errors caused by nonlinear quantization of a low-precision analog-to-digital converter, and is realized by using a deep learning network, so that the channel estimation complexity is reduced, and the performance of the sparse channel estimator can approach to a theoretically optimal channel estimation method.

The invention discloses a hybrid network station frequency hopping parameter blind estimation method based on short-time Fourier transform (STFT) and smooth pseudo Wigner distribution (SPWVD), which is used for solving the problems of low frequency hopping parameter estimation precision and high complexity under the condition of low signal-to-noise ratio in the prior art. According to the implementation scheme, the method comprises the following steps of firstly, transforming an antenna receiving signal into a time-frequency domain by using the short-time Fourier transform, and carrying out the adaptive noise reduction processing on the time-frequency signal so as to improve the anti-noise performance of a system; secondly, carrying out fine estimation on the frequency through a K-means mean value clustering algorithm; then extracting the time-frequency information according to the finely estimated frequency to obtain the time-hopping coarse estimation; and finally, carrying out fine estimation on a hopping moment by adopting the smooth pseudo Wigner transformation and a corrected truncation threshold. According to the method, the complexity is reduced, the frequency resolution isincreased, the estimation precision of the frequency hopping moment is improved, and the method can be used for the parameter estimation of the frequency hopping signal in a complex electromagnetic environment.

The invention belongs to the precise manufacturing technology field and discloses a contour error evaluation method for a triaxial numerical control machine tool. The method comprises steps that (a),an ideal processing reference track is planned, a three-dimensional coordinate of each processing point of the reference track after interpolation is stored, and a three-dimensional coordinate of eachprocessing point of an actual processing track within each sampling period is monitored and stored; (b), reference position points of the reference track closest to the actual processing points are acquired; and (c), according to different position relationships between the closest reference position points and adjacent points of the closest reference position points, a contour error is evaluatedthrough employing different methods according to different situations. The method is advantaged in that the evaluation process needs no parameter information and no curvature information of the idealprocessing reference track, the method is suitable for occasions with relatively large curvature in the processing process, realization difficulty is low, computational complexity is small, the consumption time is short, timeliness is realized, and the application scope is wide.

The invention discloses a frequency domain channel estimation method of an asynchronous multicarrier system. The frequency domain channel estimation method of the asynchronous multicarrier system comprises the steps of S1, mixing a time domain pilot frequency sequence and a data sequence into a frame by each sub-channel of a sending terminal, and carrying out multicarrier signal synthesis for the framed signals through a piecewise rapid convolution multi-carrier synthetic filter set, wherein the pilot frequency sequences of different sub-channels only occupy the frequency bands occupied by the corresponding sub-channels; S2, extracting and receiving time domain pilot frequency sequences based on a pilot frequency insertion scheme of the sub-channels by a receiving terminal, and obtaining time domain tap coefficients of the sub-channels through channel estimation; S3, after carrying out time domain de-noising processing for the estimated time domain tap coefficients, zero-filling at the tails and transforming the coefficients to a DFT domain to obtain a frequency domain channel estimation value in each sub-channel at the current moment; and S4, performing interpolation on the frequency domain channels estimated by different pilot frequency sequences on the time dimension to obtain frequency domain channel information of the data sequence position for frequency domain balancing. The frequency domain channel estimation method of the asynchronous multicarrier system greatly improves the estimation precision and reduces the channel estimation complexity in the presence of the minimum mean square error criterion.

The invention discloses a single signal arrival angle estimating method based on a large-scale multi-antenna system and belongs to the technical field of processing of signals in large-scale multi-antenna system channel estimation. The method comprises the following steps: (1) pre-estimating a channel matrix; (2) constructing a sparse transformation matrix by utilizing the correlation among receiving antennas and performing sparse transformation on a channel estimated value matrix to obtain a matrix with sparse properties; (3) accurately estimating nonzero elements in the sparse matrix according to a compressive perception principle; (4) estimating a signal arrival angle by adopting step-by-step searching. According to the single signal arrival angle estimating method, a signal arrival angle estimated value can be obtained only through a signal sample, the precision is high, and the spectrum utilization ratio can be improved. Through the step-by-step searching, the searching precision is improved, the signal arrival angle estimation precision is improved, and the estimation complexity is reduced. Through higher practicability, the single signal arrival angle estimating method can be applied to the estimation of various system parameters with sparse properties.

The invention relates to a multi-user large-scale MIMO channel estimation method and a multi-user large-scale MIMO channel estimation device, which aim at a Massive MIMO multi-user communication scene. The device comprises a user receiving signal module, a receiving signal covariance matrix smoothing module, a user arrival angle rapid estimation module, a beam forming technology estimation channelgain module and a channel state information matrix reconstruction module. According to the method, the angle of each scattering path of each user is accurately estimated by using a rapid channel covariance matrix smoothing technology and a rapid MUSIC spectrum search method, extremely low complexity is ensured, then channel gain on each scattering path is estimated by using a beam forming technology, and finally, a state information matrix is reconstructed. The MMSE channel estimator has very excellent channel estimation performance under the condition of full rank or low rank of the channel,exceeds the current optimal MMSE channel estimator, and is particularly suitable for the current 5G communication scene with very high requirements on time delay and transmission rate.

The invention discloses a signal detection method for an underwater acoustic FBMC communication system. The method is a multi-carrier (FBMC) underwater acoustic communication technology based on a filter bank. The underwater acoustic FBMC communication signal detection method provided by the invention is a method for an underwater acoustic time delay-Doppler double extension channel model, and comprises the following steps of: realizing signal synchronization and Doppler estimation by using a group of positive and negative frequency modulation HFM signals; realizing carrier frequency offset (CFO) estimation by using a known pilot symbol; realizing channel estimation by using the pilot frequency symbol subjected to CFO compensation; constructing a joint transmission matrix according to the estimated channel; and segmenting the joint transmission matrix into a plurality of sub-matrixes, and sequentially carrying out channel equalization on each sub-matrix to obtain a symbol to be detected. The method has the advantages that the system overhead is reduced by multiplexing the HFM and the preamble; the constructed joint transmission matrix comprises all non-additive interference suffered by each receiving symbol, so that the equalization performance is improved; a balancing method of a sliding window effectively reduces the calculation complexity of the system.

The invention discloses a non-continuous phase 2FSK signal symbol rate estimation method based on an FPGA, and the method comprises the steps: obtaining non-continuous phase 2FSK baseband signal data, selecting a carrier frequency, and extracting a carrier frequency parameter; respectively carrying out secondary down-conversion on an I path and a Q path of discontinuous phase 2FSK baseband signal data through carrier frequency parameters by utilizing a DDS, an adder and a multiplier IP core of the FPGA to obtain corresponding output results; respectively performing low-pass filtering on the I-path output result and the Q-path output result by using a filter IP core of the FPGA; obtaining an envelope power spectrum of the discontinuous phase 2FSK baseband signal through calculation according to the I-path output result and the Q-path output result after low-pass filtering by means of an FFT IP core of the FPGA; performing spectral line highlighting processing on the envelope power spectrum of the discontinuous phase 2FSK baseband signal, extracting a symbol rate spectral line from a processing result according to a threshold value, and calculating the symbol rate of the symbol rate spectral line according to the current sampling rate and the current FFT point number. According to the invention, the complexity of symbol rate estimation is greatly reduced, and hardware resources of the FPGA can be saved.