37 results about "Subspace algorithms" patented technology

The invention discloses a transmission tower modal parameter identification method based on an improved subspace algorithm and used for completing the modal parameter identification of a large span transmission tower structure through individually utilizing response data under the excitation of a field environment. According to the identification process, multi-group packet response data are integrated into synchronous pulse response data based on a random subspace identification principle; the integrated data avoid fitting errors caused by the packet identification, so the synchronous integrated identification of the data is realized; and a stable graph is constructed by utilizing modal parameter identification results under different system orders, so a modal parameter identification order determination problem under the environment excitation is solved, modal omission and repetition phenomena of traditional methods of a peak value method and the like are avoided, the obtained identification result is accurate and stable, and bases are established for subsequent structure damage detection and life estimation.

The invention discloses a single-snapshot data-based coherent signal DOA estimating method and relates to the technical field of array antenna systems or methods capable of indicating different signal directions. The method comprises rearranging the single-snapshot received data of an array to obtain two pseudo-covariance matrixes, and then through two pseudo-covariance matrixes, expanding the covariance matrix of a subspace restructuring algorithm; performing singular value decomposition on the new pseudo-covariance matrixes to obtain a signal subspace and a noise subspace, and performing DOA estimation on incoming coherent wave signals through a MUSIC spectrum estimating method. The single-snapshot data-based coherent signal DOA estimating method can eliminate coherence among source signals under single-snapshot conditions and meanwhile further improve the DOA estimating precision. The single-snapshot data-based coherent signal DOA estimating method is mainly applied to rapid estimation of incoming directions of coherent signals under the single-snapshot conditions.

The invention provides an efficient robust self-adapting beam forming method of broadband. The method is applied to the field of wireless communication and comprises steps as follows: performing fast fourier transform (FFT) to received data of an array to obtain the received data on different frequency points and a covariance matrix of the received data of each frequency point; choosing a central frequency point as a reference frequency point; using a propagator thought to respectively performing matrix partitioning on the covariance matrix of each frequency point and the covariance matrix of the central frequency point so as to obtain a propagator of each frequency point and the propagator of the central frequency point; constructing a focusing transformation matrix, focusing the propagators of different frequency points onto the same reference frequency point to obtain the final propagator estimation and noise subspace; and combining with a feature space method to configure a broadband beam forming algorithm weight vector to realize robust self-adapting beam forming of the broadband. In comparison with a traditional coherent signal subspace method, the method does not need any singular value or feature value decomposition, does not need a diagonal loading technique, and can reflect a good performance with respect to an environment having low snapshots and strong desired signals. Particularly, the method has stronger robustness and reduces the complexity under a condition that the desired signal estimation has a certain error.

The invention discloses a railway bridge operation state early warning method. According to the method, data observation is performed on a bridge system to acquire observation data of the bridge system in real time; meanwhile, every time a time window with set length is slid, a stochastic subspace algorithm is adopted to calculate the observation data in the time window, and modal parameter information corresponding to the observation data in the time window is obtained; and the change rate between the modal parameter information corresponding to two adjacent time windows is calculated, and then whether the calculated change rate reaches an early warning threshold is judged. In this way, tracking and recognition of bridge modal parameters in a time domain can be realized, real-time early warning can be performed on the bridge operation state, and a guarantee is provided for railway bridge operation safety.

The invention discloses a joint estimation method of signal carrier frequency and two-dimensional DOA based on undersampling, which belongs to the field of signal processing, and aims to solve the problem that the joint estimation of the two-dimensional DOA of a broadband sparse signal and the carrier frequency under the Nyquist sampling theory has high sampling rate and information redundancy. The method comprises the following steps: using a dual L-shaped array sensor to acquire signals; obtaining an undersampled value of a three-axis array by modulating a wideband converter; using the rotation invariant subspace algorithm to perform singular value decomposition on the obtained undersampled value to obtain a two-dimensional DOA and frequency parameters of the signal; overcoming the pairing of 3D parameters; and restoring the signal. The invention is applicable to the estimation of signals.

The invention provides an unmanned aerial vehicle detection method based on a time modulated array, and belongs to the technical field of unmanned aerial vehicle detection. The unmanned aerial vehicledetection method comprises the steps of: combining received multi-channel signals into a one-channel signal after periodical modulation of each antenna unit; performing down-conversion on the combined signal to an intermediate frequency analog signal, and performing sampling through an ADC module to obtain a corresponding digital signal sequence s(n); separating and recombining the digital signals into a plurality of groups of digital signals to obtain harmonic components of each group of signals; combining each group of the harmonic components into a new array signal harmonic vector, and solving an autocorrelation matrix of the new array signal harmonic vector; obtaining a DOA (Direction of Arrival) estimation angle of the signal according to a subspace algorithm of DOA estimation; and resolving the position of a unmanned aerial vehicle according to the estimated DOA of a unmanned aerial vehicle flight control map transmission signal. According to the unmanned aerial vehicle detection method of the invention, the combined one-channel signal is split in a baseband digital signal and time-frequency conversion is performed respectively, thereby improving the direction-finding precision and eliminating false spectral peaks by improving an antenna response vector.

The invention provides a single radio frequency channel spatial spectrum estimation direction finding system and method. Aiming to solve the problem of radio direction finding in wireless communication, radar, navigation and radio spectrum management, the method comprises the steps: employing a plurality of antenna units and a single radio frequency channel for achieving the super-resolution spatial spectrum estimation direction finding; periodically modulating the radio signal received on each antenna unit; combining the signals into a radio frequency signal by using a power combiner, performing down-conversion and analog-to-digital conversion, and performing spectral analysis on the radio frequency signal in a digital domain; constructing an autocorrelation matrix of the signal by usinga fundamental component and a harmonic component generated by periodic modulation, and estimating an incoming wave direction by using a multi-signal classification or rotation invariant subspace algorithm and the like. The complexity and cost of an existing spatial spectrum estimation direction finding system can be remarkably reduced, and the method is particularly suitable for a low-cost and high-precision radio direction finding system.

The invention discloses a predictive control method based on a nuclear norm subspace method and an augmented vector method. The method comprises on the basis of MPC, using an ADMM algorithm to construct a CD player mechanical arm system model, and solving a nuclear norm optimization problem by collecting the historical input and output data of the ADMM algorithm to the system model so as to obtaina group of optimized output values; setting a measurement input value, and constructing the state space model of a system by the measurement input value and an output value by means of a subspace method algorithm technique; establishing a Hankel matrix based on the measurement input value and the output value, and constructing a discrete state space model of the system based on the augmented vector method; calculating the multi-step predicted output value of the discrete state space model by a multi-step iterative method; constructing the cost function of the system and setting the referencetrajectory of the system; calculating the optimal control amount of the system, and obtaining the output value of a controller in the system. The system model avoids a large demand for actually measured data and shortens calculation time.

The invention relates to an arrival direction estimation method, and in particular relates to an arrival direction estimation algorithm based on the rotation characteristic of a signal subspace. The invention belongs to the technical field of signal processing. The invention proposes to combine the idea of signal subspace rotation transformation with a rotation invariant subspace algorithm to acquire a new arrival direction estimation algorithm. According to the invention, the rank loss characteristic of a matrix is used to divide the signal subspace in rows, and rotation transformation is carried out to acquire a new signal subspace with lower dimension; the arrival direction angle is solved by combining the rotation invariant characteristic of the signal subspaces; compared with the prior art, the problem of poor estimation stability of an ESPRIT algorithm under low signal to noise ratio and low number of beats is solved; the method reduces the dimension of the signal subspace by using rotation transformation; and the computational complexity of the algorithm is reduced.

The invention discloses a sparse subspace clustering algorithm based on semi-supervision. The sparse subspace clustering algorithm comprises the steps that data prior information is converted into a constraint matrix suitable for a sparse subspace model in the form of point pair constraint; interference of flag-free bits is eliminated in the form of Hadamard product, the state of the coefficient represented by different constraint conditions is also considered and corresponding constraint terms are established; and a semi-supervised sparse subspace model of two hard threshold and soft threshold forms is established by using the constraint terms, and a semi-supervised framework is accordingly established on the sparse subspace clustering algorithm. The clustering accuracy of the sparse subspace algorithm can still be maintained by the algorithm without prior information. Meanwhile, the performance advantages of the sparse subspace clustering algorithm are also absorbed so that the high-dimensional clustering problem containing interference information data can be directly and effectively processed, the clustering performance is ensured to be effectively enhanced under the condition of less known prior information and thus the algorithm applicability can be increased.

In an orthogonal frequency division multiplexing (OFDM) multiple-input multiple-output (MIMO) wireless communication system, a method is provided for channel estimation using a sub-space method suitable for computer implementation. The system has both a transmitter and a receiver including a plurality of antennas. The method comprising the step of: a receiver using at least one pseudo noise (PN) to correlate desired information relating to a received symbol; transforming the correlated information into frequency domain; and performing channel estimation using a sub-space method suitable for computer implementation.

The invention discloses a far-field coherent signal direction-of-arrival estimation method based on sub-space and interpolation transformation. The method specifically includes: calculating correlation vectors on all sub-intervals; calculating optimal interpolation matrices; carrying out solving on correlation vectors of a virtual matrix by the interpolation matrices and the correlation vectors ofa real array; then constructing correlation matrices of the real array and the virtual array; then constructing a linear operator and an orthogonal projection operator; and finally, calculating polynomial roots, and obtaining a far-field coherent signal direction-of-arrival estimation by solving. The linear array sub-space is constructed, solving is carried out on the correlation vectors, and feature decomposition operation of computation complexity in a traditional sub-space algorithm is avoided on the premise of ensuring accuracy; interpolation is carried out on non-uniform arrays to form uniform arrays, then a root seeking algorithm is used to estimate a direction-of-arrival angle, and frequency spectrum searching of computation complexity is avoided; and in addition, the method generalizes an application range from uniform linear arrays to arbitrary linear arrays, and is effective even when the number of sensors is larger or a real-time estimation is needed.

The invention discloses a near-field non-circular information source parameter estimation method based on a fourth-order cumulant. The method comprises the steps of constructing a symmetric uniform linear array at a receiving end; modeling a received signal of the symmetric uniform linear array; selecting receiving data of a specific array element to calculate a fourth-order cumulant matrix; constructing a matrix containing signal non-circular information by utilizing the fourth-order cumulant matrix; calculating a signal subspace of the extended matrix; obtaining a first diagonal matrix onlycontaining information source angle information through a global least squares rotation invariant subspace algorithm (TLS-ESPRIT); obtaining a second diagonal matrix only containing information sourceangle information; obtaining a high-precision direction-of-arrival estimation result through a defuzzification algorithm; constructing an augmented received data covariance matrix; and using a one-dimensional multi-signal classification (1-D MUSIC) method to obtain the distance of a near-field information source. Under the condition that the number of array elements is limited, the number of detectable information sources is greatly increased, the information source parameter estimation precision can be improved, and the method is suitable for passive positioning.

The invention relates to a method for identifying modal parameters of an oscillation signal. The method comprises the following steps: acquiring an original signal of a system, processing the original signal through an empirical mode decomposition method, extracting an intrinsic mode component, and constructing a new signal; processing the new signal by using a random subspace identification method to obtain the frequency and damping ratio of the system; processing the new signal by adopting a Prony method to obtain the frequency, the amplitude and the phase angle of the system; and on the basis of the same frequency rule, pairing the modal parameters obtained by the random subspace identification method and the Prony method to obtain complete and accurate modal parameters. Compared with the prior art, the method has the advantages that the oscillation signal is subjected to stabilization processing by using the empirical mode decomposition method, the sensitivity of the Prony algorithm to noise is overcome, the false mode generated by the random subspace algorithm in processing nonlinear and non-stable signals is avoided, the mode identification is performed by fusing the random subspace identification method and the Prony method, and the mode parameter identification precision is improved.

The invention relates to a multi-target vital sign detection method based on a subspace method and a DBF, and the method comprises the steps: carrying out the preliminary screening and positioning of a whole indoor three-dimensional space through employing the subspace method, and extracting a candidate target signal subspace; and performing digital beam forming and weighting on the angle in the candidate target signal subspace by using a DBF method, and further judging the specific accurate position information of the target through the amplitude. According to the method, a subspace algorithm is utilized, the defects that existing digital beam forming is extremely prone to being interfered by external signals, low in signal-to-noise ratio, poor in imaging effect and dependent on a large array scale are overcome, indoor high-precision multi-target life body positioning is achieved, and further vital sign detection is carried out on the target existing position.

The invention provides a method, a device, a medium and equipment for improving a fault characteristic signal, and the method comprises the steps: building initial Hankel matrixes of different dimensions based on a one-dimensional vibration signal of a rolling bearing through employing a self-determination matrix reconstruction algorithm; obtaining a high-dimensional vibration signal in the target Hankel matrix corresponding to the optimal embedded dimension; decomposing the high-dimensional vibration signal into a stationary-state vibration signal and an initial non-stationary-state vibration signal by using a steady-state subspace algorithm; extracting a target non-stationary state signal based on the kurtosis value of the initial non-stationary state signal; determining a 1.5-dimensional third-order energy spectrum of the target non-stationary state signal; in this way, a steady-state subspace algorithm is utilized to decompose a steady part and a non-steady part of the vibration signal from the high-dimensional vibration signal, and a fault signal is extracted; extracting an optimal target non-stationary state signal based on the kurtosis value; and the 1.5-dimensional third-order energy spectrum of the target non-stationary state signal is determined, so that residual noise in the optimal fault characteristic signal can be suppressed, and the fault characteristic is improved.

The invention discloses a face recognition system based on a manifold learning subspace algorithm. The system comprises a data analysis and processing center, the data analysis and processing center comprises a data processing module, an image comparison module and a data receiving module, the data processing module comprises an irrelevant linear discriminant embedding module and an orthogonal unsupervised discriminant mapping module, and a server is arranged at the output end of the data analysis and processing center. According to the invention, the filter is arranged, and the filter is setto a Log-Gabor filter and can transfer a function as a high function on a logarithmic frequency scale. the frequency response of the original image can be reflected more truly; the filter can have nodirect component, the bandwidth is not limited, the filter with any bandwidth and zero DC component can be constructed, the natural image can be better represented, the filter works to expand the feature dimension of the video image, a data processing center can conveniently carry out dimension reduction calculation on the video image, and a more accurate result is obtained.