44 results about "Matrix function" patented technology

The present invention provides for systems and methods for dimming a LED matrix functioning as a backlight to an avionics display. A system according to an embodiment of the present invention comprises a processor for receiving inputs of ambient lighting and temperature, as well as light generated by the LED matrix. The processor provides modulated pulse wave signals (square waves) to two control circuits for controlling the LED matrix in two modes. At low dimming levels, the processor modulates the duty cycle of a first square wave for affecting the light level and maintains a minimal duty cycle of a second square wave. Once the highest light level is obtained by increasing the duty cycle of the first square wave, the processor then modulates a second square wave by increasing its duty cycle. The duty cycle of the second square wave is modified by a circuit to produce a voltage level which is provided as an input to control light level of the LED matrix. As the duty cycle of the second signal is increased, so is the voltage level provided to the LED matrix and the light generated by the LED matrix.

The present application describes a method for transmitting data through a MIMO-OFDM system which requires the computation of a function f(H(sk)), where H is the channel matrix and sk=exp(j*2*pi*k/N). The method splits function f into a first, nonpolynomial part, which is in turn a function of one or more polynomial functions in sk.

The invention provides a method for detecting fatigue driving around the clock. A camera is used for acquiring a human face image to primarily position the eyes of a person; afterwards, a human eye image which is well processed by light is obtained through light preprocessing, then the human eye image is matched with a trained human eye template, and therefore the state of opening and closing of human eyes is acquired; at last, a driver is judged whether to be fatigue or not through a PERCLOS fatigue judging criterion. A light preprocessing method with the contrast ratio balanced is provided, variation of image light and dark areas is used, a convolution operation is conducted according to a convolution operation function and a Gaussian difference function, and therefore contrast ratio balance is achieved. Furthermore, the invention provides a human eye positioning method with multi-feature fusion. The advantages of various features are used, fused features can more effectively present a detected object, detecting is more accurate, and adaptability is stronger.

A solid-state laser emitting material for use in conjunction with a light source includes a polymer matrix functioning as host materials, containing laser dye of rhodamine 590 or rhodamine 610 as gain materials and nano-submicron particles as scatters therein. The lowest lasing threshold of the laser emitting material is approximately 5 mJ/cm² for 585 nm emission and 2 mJ/cm² for 630 nm emission.

The invention discloses a method and a system for realizing a video matrix function. One decoder in a video matrix is a master decoder, while the other decoders are slave decoder, wherein the master decoder is used for maintaining decoding channel information of the master decoder and each slave decoder and each piece of encoding channel information at the front end. The method comprises the following steps of: detecting and receiving serial port data by using the master decoder; analyzing the serial port data and determining a control type; according to the serial port data, the decoding channel information of the master decoder and each slave decoder and each piece of encoding channel information at the front end, determining a control command and a control object; and transmitting the control command to the control object to realize the video matrix function corresponding to the control type. Through the method and the system, the centralized management and configuration of the decoders can be realized, so that the video matrix function can be realized by adopting a digital method and the engineering complexity can be reduced.

The invention provides an MIMO transmission power distribution optimizing method based on robustness design. Firstly, convexity-concavity of a matrix function is utilized, a power distribution matrix is optimized to be a diagonal matrix and the diagonal matrix is simplified to be a scalar function. Consequently, the scalar function is combined to perform optimization on MIMO transmission power distribution. During optimization of MIMO transmission power distribution, the optimal solution without limited conditions is found out and then adjusted in a fine mode. The algorithm convergence and low operation quantity are ensured. Compared with an ordinary optimal solution obtaining algorithm, the MIMO transmission power distribution optimizing method based on the robustness design can obtain the optimal solution with little operation cost. In addition, as a reference value of power distribution is fixed in the first step, the problem of algorithm divergence does not exist.

The invention discloses a real-time extraction and update method of background image in motion video, belonging to the technical field of digital image processing. The method utilizes the different change properties of the background image and a motion object image in the time domain to construct the stable matrix function of the image, automatically distinguishes background spots and foreground spots through the change of factors in the stable matrix, sets the highest and lowest saturation values for the stable matrix, so that the algorithm can automatically perceive the sudden change of the background in a short time, thus quickly establishing a background image model and carrying out real-time update. The invention has the beneficial effect that the method can quickly obtain the stable background image so as to quickly and accurately accomplish detection of the motion object.

The invention belongs to the electrical control technology field, relates to the robust control technology field and especially relates to a magnetic levitation bearing robust controller constructionmethod based on a multi-objective genetic algorithm. The method is characterized by summarizing the general formula of a robust controller weight matrix function and limiting the range of a weight function parameter; and giving the adjustment direction of manual weight parameter adjustment, and solving the blindness of a beginner choosing a robust controller weight function. In the invention, a multi-objective genetic algorithm is further adopted to optimize the weight function parameter and a designer can compromise between closed loop response performance and a control signal according to anactual demand so as to acquire an optimal robust H-infinite controller under two objectives. A user does not need to grasp excess professional knowledge and can easily and reasonably construct the robust H-infinite controller suitable for an axial (a single degree of freedom) magnetic levitation bearing system.

An image processing system includes a filtering arrangement configured to receive incoming pixel information and filter at least a first portion of the information to thereby pass a second portion of the information for further processing; circuitry configured to apply a first color correction function to the incoming pixel information to thereby produce a modified first portion; circuitry configured to apply a second color correction function to the second portion to thereby produce a modified second portion; and an adder configured to combine the modified first portion to the modified second portion.

The present invention relates to a dynamic keyboard matrix function switching method and system. The method comprises the following steps: acquiring a keyboard matrix type; and obtaining a corresponding keyboard scan code from a keyboard matrix type and keyboard scan code correspondence table pre-stored in embedded controller codes according to the keyboard matrix type. According to the dynamic keyboard matrix function switching method and system, by identifying the keyboard matrix type, the corresponding keyboard scan code is obtained from the pre-stored keyboard matrix type and keyboard scan code correspondence table for use, so that a set of embedded controller codes are adaptive to a plurality of types of keyboards, the number of the embedded controller codes is reduced, and maintenance difficulty of the embedded controller codes is reduced; and another set of embedded controller codes do not need to be refreshed for changing a keyboard when demands are different, so that work load of repeatedly refreshing the embedded controller codes is reduced, and the probability of errors caused by repeatedly refreshing the embedded controller codes is reduced.

The invention is applicable to the technical field of communication and provides a method and a device for acquiring a useful signal in full-duplex communication and a full-duplex communication system. The method comprises the following steps: acquiring a base vector and the number of dimensions of a null space of a self-interference channel matrix according to the self-interference channel matrix between a local transmitting antenna and a local receiving antenna in full-duplex communication; acquiring transmitting data flow transmitted in the null space of the self-interference channel matrix and the number of transmitting data flows transmitted in a row vector space of the self-interference channel matrix according to the base vector and the number of dimensions of the null space as well as channel condition and service information; and acquiring a locally-received useful signal according to the transmitting data flow transmitted in the null space of the self-interference channel matrix, a transmission pre-coding matrix function and the self-interference channel matrix when the number of the transmitting data flows transmitted in the row vector space of the self-interference channel matrix is not equal to 0, thereby achieving the purpose of acquiring a useful signal more flexibly.

The invention discloses a method to make handwriting screen of mobile phone have functions of character inputting and image inputting. In the storage of the mobile phone, font containing characters and symbols and database containing dot matrix function are set to make MPU of mobile phone access to character input mode and image input mode according to the selected instruction while user is inputting by handwriting pen. When in character inputting mode, MPU can automatically search the corresponding font or symbol according to the inner code of the inputted font or symbol and display onto handwriting screen. When in image inputting mode, MPU can call dot matrix function corresponding to the image and display. Thus, the handwriting screen can not only input normal character and symbol, but also input image.

The invention provides an ultra-short base line orientation estimation method for any formation. The method comprises the steps of initializing a delay adjustment matrix according to a received signal and a delay signal thereof, thus acquiring a delay adjustment matrix function, and initializing a distributed search network; detecting the arrival time of the signals, determining synchronous moments, sampling the synchronous moments, building a synchronous sampling moment set, thus acquiring a reference array element sampling sequence of the signal arrival moments; and computing the rotated distributed search network, computing a matched peak of each direction vector under the synchronous sampling moment set according to the reference array element sampling sequence of the signal arrival moments and the delay adjustment matrix function, and traversing orientation angles, wherein the direction of the maximum matched peak is the final estimated orientation vector. The ultra-short base line orientation estimation method for any formation provided by the invention can be expanded to various complex formations, and is stable in performance when the part of array elements is invalid, and high in estimation accuracy of the distance and orientation of the target signal source.

Apparatus and systems, as well as methods and articles, may operate to use a graphics processing unit (GPU) to perform color conversions using International Color Consortium (ICC) profiles. In some embodiments, code is generated for execution by the GPU. The conversion can be represented as a series of steps mapped to particular GPU processes such as 1D texture, 3D texture and matrix functions.

The invention belongs to the electrical control technology field, relates to the robust control technology field and especially relates to a magnetic levitation bearing robust controller constructionmethod based on a multi-objective particle swarm algorithm. The method is characterized by summarizing the general formula of a robust controller weight matrix function and limiting the range of a weight function parameter; and giving the adjustment direction of manual weight parameter adjustment, and solving the blindness of a beginner choosing a robust controller weight function. In the invention, a two-objective particle swarm algorithm is further adopted to optimize the weight function parameter and a designer can compromise between closed loop response performance and a control signal according to an actual demand so as to acquire an optimal robust H-infinite controller under two objectives. A user does not need to grasp excess professional knowledge and can easily and reasonably construct the robust H-infinite controller suitable for an axial (a single degree of freedom) or a radial (multiple degrees of freedom) magnetic levitation bearing system.

The present invention relates to a method for determining hydrocarbon production for a reservoir. The method comprises

• • determining a projector matrix based on a Jacobian matrix function of the gridded model, then splitting the Jacobian matrix into subsets of consecutive lines.
  • For each subset of consecutive lines, creating a respective square matrix based on said subset.
  • A determining eigenvectors and respective eigenvalues associated with the respective square matrix and then determining relevant eigenvectors having respective eigenvalues below a predetermined threshold.
  • The projector is determined as a concatenation of the extended eigenvectors ordered according to multiple criteria: the respective order value of the subset; and the respective eigenvalue of the relevant eigenvector.