35results about How to "Fast calculation" patented technology

The invention discloses a kinematic inverse solution method and device of a mechanical arm and the mechanical arm. The kinematic inverse solution method comprises the steps that the geometric parameters of the mechanical arm are determined, wherein the mechanical arm comprises multiple rotation joints, and the rotation joints are divided into the first type of rotation joints and the second type of rotation joints; according to the geometric parameters of the mechanical arm, the first type of joint angles corresponding to the first type of rotation joints are determined by means of a geometricalgebra method; according to the first type of joint angles, the second type of joint angles corresponding to the second type of rotation joints are determined by means of the eulerian angle rotationmatrix transform method, wherein the first type of joint angles and the second type of joint angles are used for controlling operation of the mechanical arm. The kinematic inverse solution method anddevice of the mechanical arm solves the problem that the solution process is complicated in the prior art and achieves a simple solution process easy to understand, the code quantity is simplified, the calculation rate is rapid, and the precision is high.

The present invention relates to a method for noise reduction of ECG signals based on heterogeneous computing, comprising: inputting ECG signal data of several patients; decomposing each ECG signal through CEEMDAN, and each ECG signal will be obtained after decomposition A set of modalities, different ECG signal decomposition processes are performed in parallel; judge whether the obtained modalities are pure modalities; linearly add the obtained pure modalities, and finally obtain pure signals. This method is based on heterogeneous computing. With CEEMDAN, it has a certain adaptive ability, can reduce time overhead, and has a fast calculation rate and high denoising efficiency.

A subgrade compaction construction data efficient processing system, comprising a data splitting unit, a data storage queue unit, a storage task scheduling thread pool unit, an original data storage unit, a data calculation task queue unit, and a calculation task execution unit. The construction data efficient processing system aim at achieving high-concurrency rapid storage and high-concurrency rapid calculation processing of a large amount of construction original data. By designing a reasonable data fragmentation rule and a storage task scheduling policy based on a thread pool, high-concurrency rapid storage of real-time data is implemented. By means of fragmented storage of the data, high-concurrency reading, calculation and processing of the original data are implemented.

The invention discloses a seismic data reconstruction method, comprising the following steps: acquiring seismic data to be reconstructed and its acquisition matrix; performing block processing on the seismic data and acquisition matrix; establishing each small block of seismic data and its corresponding acquisition The sparse inversion model regularized by the hyperbolic tangent function of the matrix; the optimal solution of the sparse inversion model is obtained through each small block of data, and the reconstructed seismic data is obtained; data reconstruction is performed on all block data, and the All reconstructed data are stitched into a complete reconstruction. The seismic data reconstruction method disclosed by the present invention has fast calculation speed and high precision, and is not only applicable to two-dimensional data, but also applicable to three-dimensional data and five-dimensional data; not only applicable to pre-stack seismic data, but also applicable to post-stack seismic data; not only applicable to It is applicable to land seismic exploration data and also applies to marine seismic exploration data.

The invention discloses a power flow calculation method for a three-phase imbalanced medium-and-low-voltage active distribution network with a looped network. The method comprises: for the three-phaseimbalanced medium-and-low-voltage active distribution network with the looped network, acquiring known basic network data; by a virtual current compensation method, decomposing the looped network into a decomposed looped network; obtaining the power flow post-voltage of various nodes in a power grid after power flow iteration; performing power flow iterations continuously, and calculating, aftereach power flow iteration, the maximum value of the voltage differences of the nodes in the power grid before and after the power flow iteration: and determining until the requirements are satisfied so as to complete the power flow processing. The method of the invention can process the looped imbalanced network in the three-phase imbalanced medium-and-low-voltage active distribution network, performs calculation quickly, occupies small memory, has high power flow result accuracy, can optimize the construction of the three-phase imbalanced medium-and-low-voltage active distribution network, and provide accurate data for grid dispatch.