36 results about "Filter theory" patented technology

The invention discloses a synthetic aperture radar automatic target recognition method which belongs to the target recognition field and mainly solves the problem that the space complexity of the existing synthetic aperture radar automatic target recognition technology is higher and single classifier has low recognition rate. The method comprises the following recognition steps: preprocessing, extracting characteristics, training classifiers and identifying target, wherein the step of extracting characteristics is to extract PCA characteristics of the synthetic aperture radar image, elliptic Fourier descriptor and two-dimensional Fourier transform; the step of training classifiers is based on the extracted three characteristics to separately use K-nearest neighbor method, support vector machine and MINACE filter theory to train three classifiers; and the step of identifying target is to input the extracted synthetic aperture radar image to be identified in the trained three classifiers for classification and finally adopting the Dempster-Shafer evidence theory to fuse the recognition results of the three classifiers. The method has the advantages of high recognition rate and low space complexity and can be used in the target tracking of the military or civilian field.

The invention discloses a distributed passive radar target detection method under a direct wave-free condition, and belongs to the technical field of distributed passive radar target detection. A conventional passive radar target detection method is based on the classical matched filtering theory, and approximately optimal detection properties can be achieved on premise that a direct wave with a relatively high signal-to-noise ratio can be acquired in real time and an emitted signal can be estimated with high quality by using a direct wave signal received by a reference channel. To solve the problem of detecting a target which cannot receive the direct wave signal in practical, the invention discloses the distributed passive radar target detection method under a direct wave-free condition, and a concentrated target detector under the direct wave-free condition is established, so that target detection can be implemented when the signal-to-noise ratio of the direct wave signal is relatively low or the direct wave signal cannot be received in a multiple-input multiple-output geometric structure, and meanwhile the target matching problem which is very hard to solve among different receiving stations of a distributed passive radar is indirectly avoided.

The invention concerns a method for determining the fouling ratio of at least one filter of a ventilation and / or air handling system(s), comprising the following steps: choosing at least one filter to be analyzed B, determining the theoretical filtration capacity of the filter selected in step A by depending on the theoretical efficiency of the filter and the maximum amount of dust which can be theoretically retained by the filter, C collecting pollution data of at least one air entering the ventilation system and passing through the selected filter, D measuring the air flow rate entering the ventilation system, E calculating the amount of dust retained by the at least one filter selected, from the pollution data of the entering air of step C and from the measurement of the entering air flow rate of step D, F repeating step E for each time step t, G. determining the accumulated dust amount retained by the selected filter from the results of steps E and F. H determining the fouling ratio of the selected filter by comparing the accumulated dust amount determined at step G with the theoretical capacity of the filter determined in step B.

The invention discloses a longitudinal control method for a hypersonic flight vehicle based on online identification of an aerodynamic parameter. The method firstly employs an iteration extension Kalman filtering algorithm combining an extension Kalman filtering algorithm and an iteration filtering theory to perform the online identification on the aerodynamic parameter, and provides a more accurate model for subsequent control; and then, based on the identified aerodynamic parameter, a adaptive filtering back-step controller is designed to precisely controls a longitudinal equation of the hypersonic flight vehicle. Compared with a traditional extension Kalman filtering algorithm, the iteration extension Kalman filtering algorithm employed by the method is higher in precision, and can identify the aerodynamic parameter more accurately. A control strategy proposed by the invention can effectively overcome a phenomenon of "inconsistent parameters of heaven and earth", the accuracy of a model is improved through the online identification, and the pressure of a control system is alleviated.

The invention discloses a radiation source individual characteristic enhancement method based on a time-varying filtering theory, and relates to a radiation source individual characteristic enhancement method. The objective of the invention is to solve the problems of low individual identification accuracy of an existing radiation source and identification failure caused by change of main signal parameters. The method comprises the following steps of: 1, performing time-frequency distribution calculation on a multi-component signal, and extracting time-frequency information to obtain time-frequency information of each signal component in a main signal component; 2, recovering and separating signal components in the main signal components one by one based on a time-varying filtering algorithm of order time-varying short-time fractional Fourier transform to obtain an estimation result of the sum of the main signal components; 3, subtracting an estimation result of the sum of the main signal components from the radiation source signal to obtain a residual component of the multi-component signal; 4, performing feature extraction on the residual component, and constructing a feature vector; and 5, inputting the constructed feature vectors into a classifier, and outputting a classification and recognition result. The method is applied to the field of radiation source individual characteristic enhancement.

A joint precoding method for a multi-user MIMO relay system including a direct transmission link is disclosed. In the scene of transmitting / receiving end multi-user MIMO relay communication, with MSMSE as a design criterion, the invention provides a joint precoding method with the centralized optimization of all users of a transmitting end. According to the MMSE receiving filter theory, the optimal expression of a receiving filter matrix is directly solved. The sub-optimization problem of a sub-precoding matrix of a transmitting end user is obtained by solving a K.K.T condition, the sub-optimization problem of a relay transceiver matrix is transformed into a standard SDP problem for optimization. Finally, the sub-precoding matrix of the transmitting end user, the relay transceiver matrix and the receiving filter matrix are jointly iterated to convergence, and an optimized precoding matrix is obtained. A simulation analysis verifies that a mean square error and bit error rate performance of the proposed method are good, and the system performance is optimal especially in the case of a low signal-to-noise ratio.

The invention discloses a synthetic aperture radar automatic target recognition method which belongs to the target recognition field and mainly solves the problem that the space complexity of the existing synthetic aperture radar automatic target recognition technology is higher and single classifier has low recognition rate. The method comprises the following recognition steps: preprocessing, extracting characteristics, training classifiers and identifying target, wherein the step of extracting characteristics is to extract PCA characteristics of the synthetic aperture radar image, elliptic Fourier descriptor and two-dimensional Fourier transform; the step of training classifiers is based on the extracted three characteristics to separately use K-nearest neighbor method, support vector machine and MINACE filter theory to train three classifiers; and the step of identifying target is to input the extracted synthetic aperture radar image to be identified in the trained three classifiers for classification and finally adopting the Dempster-Shafer evidence theory to fuse the recognition results of the three classifiers. The method has the advantages of high recognition rate and low space complexity and can be used in the target tracking of the military or civilian field.

The invention discloses a distributed passive radar target detection method under a direct wave-free condition, and belongs to the technical field of distributed passive radar target detection. A conventional passive radar target detection method is based on the classical matched filtering theory, and approximately optimal detection properties can be achieved on premise that a direct wave with a relatively high signal-to-noise ratio can be acquired in real time and an emitted signal can be estimated with high quality by using a direct wave signal received by a reference channel. To solve the problem of detecting a target which cannot receive the direct wave signal in practical, the invention discloses the distributed passive radar target detection method under a direct wave-free condition, and a concentrated target detector under the direct wave-free condition is established, so that target detection can be implemented when the signal-to-noise ratio of the direct wave signal is relatively low or the direct wave signal cannot be received in a multiple-input multiple-output geometric structure, and meanwhile the target matching problem which is very hard to solve among different receiving stations of a distributed passive radar is indirectly avoided.

The invention discloses a production technology of an anti-mite pillow interior. The technology comprises the following steps of design preparation of mite-separation fabric and preparation of the anti-mite pillow interior. According to the production technology of the anti-mite pillow interior, based on a filter theory, fiber in a filter material is thinner, the specific surface area is larger, the pore diameter is smaller, and the filter precision and the filter effect are better. Therefore, the density of the fabric is increased, invasion of mites or other allergens is prevented, and then the anti-mite aim is achieved.

The invention discloses a permanent magnetic synchronous motor rotor speed estimation method. The method includes establishing a permanent magnetic synchronous motor mathematical model, and performing vector transformation to decouple; performing order reduction on a system model, and establishing a simplifying filter model estimating the rotor speed and position only; adding stator parameters into the system model, and acquiring an augmented system model containing the generator electromechanical state and rotor parameters; finally, according to the established augmented model, acquiring an estimation value of the permanent magnetic synchronous motor rotor speed and stator parameters by the H8 filter theory. By the aid of the method, the rotor speed and stator parameters can be estimated accurately, the calculation amount is reduced, and the system robustness against parameter uncertainty is improved effectively.

The invention discloses a generator dynamic state estimation method based on adaptive traceless H-infinity filtering, which is used for enhancing the robustness of a dynamic state estimator on model uncertainty and improving the dynamic state estimation precision. According to the method, based on the unscented transformation and H-infinity filtering theories, by introducing an adaptive technology, a state estimation error covariance adaptive adjustment updating strategy capable of inhibiting the influence of model uncertainty on a state estimation result is provided. According to the method,the dynamic state estimation precision of the generator can be effectively improved, and the problem that upper bound parameters are difficult to select due to model uncertainty constraint in a traditional robust state estimation method is avoided. According to the method, the actual engineering background is considered, the implementation process is clear, and high engineering application value is achieved.

The present invention provides a low-illuminance image / video enhancement method based on adaptive multi-scale filtering. The present invention is based on the atmospheric light scattering model, adaptive filtering based on regional characteristics, and adaptive brightness restoration theory, in the field of low-illuminance video and image enhancement , the adaptive multi-scale filtering theory and method proposed can improve the brightness of the processed image, and at the same time improve the contrast and sharpness of the image or video, with the classic mean filter, low computational complexity; adaptively Calculation of internal parameters has a wide range of adaptation, and can be applied to color images or grayscale images, and has good versatility.

The invention relates to a distributed H infinity filter based cooperative spectrum sensing method. The distributed H infinity filter based cooperative spectrum sensing method comprises the steps that each sensing user in a cognitive wireless network firstly collects and receives wireless signals sent by an authorized user, calculates the energy values of the received signals, then builds a state transition equation of wireless channel shadow fading and performs distributed estimation on the wireless channel shadow fading; then each sensing user performs normalization processing on the energy values of the received signals by utilizing the estimation value of the wireless channel shadow fading and then performs local spectrum sensing; each sensing user sends the local spectrum sensing result to a sensing base station; and finally, the sensing base station performs decision fusion based on the local spectrum sensing result sent by each sensing user to obtain a final cognitive wireless network global spectrum sensing result. As the distributed H infinity filter used for estimating the wireless channel shadow fading received by each sensing user is adopted to remove influences from the wireless channel shadow fading to the spectrum sensing result, the distributed H infinity filter based cooperative spectrum sensing method can detect the spectrum of the authorized user equipment in the cognitive wireless network in a better manner.

The invention discloses a horizontal array motion aperture synthesis positioning method and a device therefore. The horizontal array motion aperture synthesis positioning method comprises: decomposing the underwater acoustic data collected by the array elements of a horizontal array into a plurality of time units; evaluating the phases of each narrowband time unit laminated spatially between the array elements, to obtain a phase evaluation factor, processing time delay sampling and phase compensation on the array element time units; processing aperture synthesis for the time sequence obtainedby different array elements in one narrow band, to form a large aperture virtual horizontal array; according to zero point control full normal mode filter theory, processing zero point control mode filtration on the large aperture virtual horizontal array, to obtain target mode information; using the target mode information to obtain three-dimension positioning result of the target according to maximum likelihood method. The method and device can obtain high positioning accuracy, with simple and practical process. The invention can realize real-time application without much additional computing resource.