289results about How to "Reduce error accumulation" patented technology

The invention discloses a one-dimensional rotation and two-dimensional inclination working platform, which comprises a pedestal, a bottom layer board, a middle layer board, an upper layer board, an adjusting mechanism and a driving part. The bottom layer board and the upper layer board achieve the inclination of the working platform in a mutually orthogonal direction respectively, while the middle layer board achieves the rotation of the working platform in a larger range in a plane vertical to the two orthogonal directions. The pedestal is connected with the bottom layer board through a rotating pair, a wedged slide block and a spring, a linear guide way is fixed on a bottom plate, and the wedged slide block supports the bottom layer board; the bottom layer board is connected with the middle layer board through a cambered guide way and a central rotating shaft; and the middle layer board is also connected with the upper layer board through the rotating pair, the wedged slide block and the spring, the linear guide way is fixed on the middle layer board, and the wedged slide block supports the upper layer board. The adjusting mechanism comprises a coarse adjusting structure and a fine adjusting structure. The invention can achieve the work demands of the one-dimensional rotation and the two-dimensional inclination for the working platform, and can achieve the angle positioning of a sub-angle second-time level in a larger space adjusting range.

The invention relates to a twin-translation-rotation partial decoupling parallel mechanism, relating to a three-degree-freedom parallel mechanism in the field of industrial robots. A movable platform is arranged above a fixed platform; a tail-end actuator is arranged above the movable platform; the fixed platform is connected with the movable platform through three alien branches; a first branch is provided with a rotary pair, and a movable pair of the first branch connected with the fixed platform is a rotary pair; a movable pair in the first branch connected with the movable platform is a universal joint or a ball pair; a movable pair in the second branch connected with the fixed platform is a rotary pair; a movable pair in the second branch connected with the movable platform is a universal joint or a ball pair; a third branch is formed by a branch chain with six degrees of freedom and is provided with a movable pair; a movable pair of the third branch connected with the fixed platform is a universal joint; a movable pair of the third branch connected with the movable platform is a ball pair; and the movable pair, the ball pair and the universal joint in the third branch are connected into UPS branch chain. The invention is suitable for machine tools, robots, and the like.

The invention discloses a Markov chain modeling and predicating method based on wind power variable quantity. The method comprises the steps that firstly, linear transformation is carried out on existing original power data to obtain a wind power variable quantity data sample, then, according to the variable quantity data size and a statistical result of probability distribution, state space of a Markov chain model is divided as fine as possible, after the state is determined, a transition probability matrix of the variable quantity is obtained through statistical calculation, and Markov chain model construction is completed. The Markov chain model can be used for constructing a short-period and ultra-short period wind power predication method, and the theoretical basis is laid for real-time economic dispatch comprising a wind power system, optimized decision and model prediction control based on a Markov chain.

The invention provides a high precision angle measurement system and method for a spacecraft. The high precision angle measurement system comprises an alignment measurement subsystem and a mechanical servo subsystem, wherein the mechanical servo subsystem comprises a spacecraft parking frame and the like; the alignment measurement subsystem comprises a first benchmark theodolite and the like; in a using process, a spacecraft body is statically put on the spacecraft parking frame; the first benchmark theodolite and a second benchmark theodolite are fixedly arranged on a benchmark locating support; an alignment theodolite is fixed on a vertical motion support and moves up and down along the vertical motion support; and the whole vertical motion support is mounted on an annular rotating table, the annular rotating table realizes rotation in the azimuth of 0-360 degrees, and motion of the whole mechanical servo system can form a cylindrical space at one height. The high precision angle measurement system provided by the invention adopts a CCD photoelectric auto-collimator for replacing the traditional optical theodolite telescope system, so that the accuracy and stability of measured data are improved, and the high precision angle measurement system has an automatic reading function.

The invention relates to the technical field of shipbuilding, in particular to a method for gathering ships by calculating frames, half widths and heights with three-dimensional coordinates, and aims to solve the problems of low accuracy, waste of time and force, and large error accumulation in the conventional measuring method. The method comprises the following specific steps: 1) building a coordinate system; 2) building a station for an instrument freely; 3) gathering segmentally; 4) floating and positioning; 5) measuring the principal dimension of ships. The measuring method has the advantages of high accuracy, saving of time and force, and free from error accumulation; measured data can be used for fully and visually reflecting posture of shipbuilding, so that the principal dimension and linearity of ship bodies can be measured and controlled strictly.

The invention discloses a visual course identification method based on a target point identification. The method comprises a step of loading an azimuth angle and pixel correspondence table obtained through camera calibration, selecting an initial target point in a first frame image, extracting and describing SURF characteristic points in first and second interest regions, and storing the frame image as a first reference image, a step of adjusting the position and size of an interest region in a collected next frame image, carrying out characteristic matching with the reference image, screen an ideal coplanar characteristic point, calculating a homography matrix, and identifying the current position of the target point, a step of updating the current frame image as the reference image when a reference image updating condition is satisfied, screening a stable ideal coplanar feature points set which is in accordance with affine transformation and is in a infinite plane, and correcting homography matrix and target point positions, and a step of finally identifying a target point corresponding course according to the azimuth angle and pixel correspondence table.

The invention relates to a voice synthesis method and device, a storage medium and electronic equipment, and the method comprises the steps: inputting a to-be-synthesized text into an information extraction model, and obtaining voice feature information corresponding to the to-be-synthesized text; inputting the voice feature information into a voice synthesis model to obtain acoustic features corresponding to the to-be-synthesized text, the voice synthesis model comprising a duration sub-model and an acoustics sub-model, and the duration sub-model and the acoustics sub-model being subjected tojoint training to obtain the voice synthesis model; and obtaining audio information corresponding to the to-be-synthesized text according to the acoustic features. Therefore, the acoustic features can be directly obtained through the voice synthesis model according to the voice feature information corresponding to the to-be-synthesized text without cooperation of multiple models, so that the voice synthesis efficiency can be improved, the error accumulation can be effectively reduced, and the accuracy of the voice synthesis method is improved.

The invention discloses an optical motion capture data processing method based on a module piecewise linear model. Based on overall information of optical human motion capture scattered data, a data processing algorithm based on the module piecewise linear model is provided. By using the module piecewise linear model to generalize the change characteristics of different modules, the method determines a matching priority and an intra-segment fitting function of module data, effectively performs overall layered prediction and tracking for each three-dimensional motion data module, performs de-noising processing for the noise data based on the module, and provides an interpolation fitting algorithm for the missing motion data based on segmenting Newton to perform reasonable supply. The optimized method does not need manual intervention during processing, and can meet real-time requirement.

The invention discloses a medical image pulmonary nodule detection method based on a cyclic feature pyramid, and mainly solves the problem of low detection precision caused by high extraction difficulty of pulmonary nodules and ground glass nodules in the prior art. The method comprises the following implementation steps: 1) acquiring a medical image; 2) preprocessing the medical image, and expanding a sample data set; 3) constructing a circular feature pyramid detection model by combining the feature map; 4) training the detection model by using the extended data set sample to obtain a targetdetection model; and 5) inputting the test set in the data set into the trained detection model to carry out pulmonary nodule detection. The method constructs a new feature extraction network, accelerates the network training speed, enhances the sensitivity to nodules of different sizes, improves the detection precision of pulmonary nodules of medical images, and can be used for a computer-aidedmedical diagnosis system.

The invention provides a weighted DV _ Hop node positioning method based on a minimum mean square error criterion, and the method comprises the steps: setting other anchor nodes to have the same weight under the minimum mean square error criterion, and obtaining a preliminary estimation value of the average distance of each hop of the anchor nodes; setting weights for other anchor nodes for the second time, and obtaining the corrected average hop-by-hop distance of the anchor nodes under the minimum mean square error criterion; calculating the connectivity of each adjacent anchor node and theaverage distance error of each hop; calculating the average hop distance of the unknown node and the estimated distance from the unknown node to the anchor node according to the average hop distance of the adjacent anchor nodes and the normalized weight value; and setting a linear solution of a weighted least square method as an iterative initial value, minimizing a weighted distance estimation error objective function through a nonlinear weighted iterative method, and taking an iterative result as a positioning coordinate of an unknown node. According to the method, the more accurate averagehop distance between the anchor node and the unknown node can be obtained, the calculated amount is small, and the positioning accuracy and stability are improved.

Disclosed is an intelligent wheelchair static gesture identification method. The method includes the steps of 1) collecting scene depth information through Kinect, 2) separating gestures in the scene depth information according to the depth information, 3) extracting the feature vector of the gestures by using normalization central moment, and 4) conducting gesture identification through a DAGSVM classifier according to the feature vector of the gestures extracted in the step 3). According to the intelligent wheelchair static gesture identification method, in the generating process of the DAGSVM classifier, the between-class distance and the class standard deviation of each SVM classifier are calculated, SVM classifiers are arrayed in the descending order according to the between-class distances, and the SVM classifier with the largest between-class distance is chosen to be used as a root node classifier of the DAGSVM classifier. Similarly, the SVM classifiers with the largest between-class distance at the positions of the rest nodes can be chosen. The phenomenon of error accumulation can be effectively reduced, the correctness of a gesture recognition result is improved, and the safety and the reliability of an intelligent wheelchair man-machine interaction system are furthest guaranteed.

The invention belongs to the field of intelligent navigation devices and discloses a simultaneous localization and mapping and inertial navigation combined error fusion system. The system comprises a carrier, an inertial navigation device, a simultaneous localization and mapping device and an error fusion device. The inertial navigation device comprises an accelerometer, a gyroscope and an inertial navigation driving device. The simultaneous localization and mapping device comprises laser radar and an environment map building device. The error fusion device comprises a linear discrete simultaneous localization and mapping error module, an inertial navigation error module and a weight fusion module. Therefore, high accuracy and reliability in positioning and navigation of robots can be realized.

The invention discloses a method and device for modeling a superficial layer in an earth surface based on seismic wave data. The method includes the steps that seismic wave original data are preprocessed, a first model is obtained, the travel time and the offset distance of early-reached waves are determined based on the first model and the interesting depth; observation data about the early-reached waves are extracted from the original data according to the travel time and the offset distance; calculating data about the early-reached waves are generated according to the determined travel time ts and the determined offset distance on the basis of the first model; data field fitting is performed on the observation data and the calculating data so that the gradient can be obtained, and therefore a speed updating model is further obtained; the first model and the speed updating model are overlapped to obtain a second model, and the steps are executed repeatedly until the obtained speed updating model reaches the global optimum. Only part of early-reached wave information is used, nonlinearity of a target function is reduced, small offset distance information is adopted, error accumulation is reduced, and therefore by means of the method and device for modeling the superficial layer in the earth surface based on the seismic wave data, the global optimal solution can be converged more easily.

The invention discloses an air quality prediction method based on multi-step recursive prediction, and the method comprises the following steps: obtaining and preprocessing air quality data and weather data, and constructing input data and output data, wherein the input data of the encoder comprises pollutant data and historical meteorological data, the input data of the decoder comprises the output result of the encoder, the weather forecast data and the pollutant data at the previous moment; dividing the data into training data and test data; using the training data to train the Seq2Seq model; and using the test data to test the prediction result. According to the method, the air quality is predicted by using the Seq2Seq model. An RNN of an encoder is replaced by using a full connectionlayer, and a time sequence relation of an input sequence is reserved by using position coding, so that the effect of accelerating training while maintaining the prediction precision is achieved, erroraccumulation can be effectively reduced, and the prediction precision is improved.

The invention relates to a time window online correction-based multi-AGV (Automated Guided Vehicle) path planning method. The method includes the following steps that: the operation point sites of anAGV are extracted according to workshop layout information and path information, independent road sections are obtained through division according to the operation point sites; a scheduling system updates an AGV status list and a real-time task list in real time, assigns a task in the real-time task list to an idle AGV, and informs the AGV of a start point and an end point; a feasible path list isestablished, whether a feasible path exists is judged, a task corresponding to the feasible path is added in a time window model, the AGV executes the tasks and feeds back current AGV position information in real time; and whether the current AGV position information is consistent with the position information of the AGV in the time window model is judged, and the time window model is corrected online and updated synchronously. With the time window online correction-based multi-AGV (Automated Guided Vehicle) path planning method of the invention adopted, error accumulation caused by actual errors can be decreased; and a multi-AGV path conflict problem can be solved. The time window online correction-based multi-AGV (Automated Guided Vehicle) path planning is a new dynamic path planning solution suitable for any environment and any number of AGVs, and is suitable for being popularized.

The invention discloses a multi-camera measurement method based on a refractive projection model and a beam tracking method, and used for the field of visual detection and image measurement. The process of the present invention comprises the following steps of: building a multi-camera system; global calibrating by multi-camera based on the refractive projection model and the beam tracking method;respectively capturing images from different directions by a plurality of cameras; extracting and matching features to obtain measured three-dimensional points; converting the three-dimensional pointsobtained by each subsystem into the same coordinate system to obtain global measured data; and using the obtained three-dimensional point data to calculate an indirect amount (e.g., strain, displacement). The key of the present invention is the calibration of the multi-camera system, the result will affect the subsequent three-dimensional point acquisition. According to the multi-camera measurement method based on the refractive projection model and the beam tracking method, the features in the measurement process are that the multiple cameras in the system can be calibrated at the same timeto solve the problem that a planar calibration plate limits the distribution of multiple cameras, and the measurement process is flexible.

A fine machining method of the rolling stand frame belongs to machine finishing technology field. At first determining center of turning circle position of the fine boring machine table and mounting end face multiple tooth circular dividing table thereon, the tetrahedron is mounted on the end face multiple tooth circular dividing table, and adjusting it to collimate with the photoelectricity autocollimation to determine machineable machine zero position point; the rolling stand frame is mounted on the worktable, and adjusting it to symmetry with the center of gyration; boring the frame side face of each axle hole for the rolling stand frame and the axle hole until achieving to the required precise size, when the frame rotates from one side to another side, rotating relevant angle of the worktable contrasting to graduation value of the end face multiple tooth circular dividing table, the tetrahedron is measured by the photoelectricity autocollimation, fine tuning the worktable to collimate the corresponding face of the tetrahedron with the sending beam of the photoelectricity autocollimation, at this point, the rotating angle is exact value, then processing. The invention operates conveniently and audio-visually, with reliable processing manufacturability, and reducing accumulation of error and improving process precision of the frame.

The invention belongs to the fire positioning field, particularly relates to a fire positioning method on the basis of smoke concentration, is on the basis of a wireless sensor network and carries out positioning estimation on a fire source point through utilizing smoke concentration acquired by the concentration sensor on the basis of a smoke diffusion model. The method comprises steps that 1, the smoke diffusion model under the no wind environment is established, and a relationship between gas concentration under the no wind condition and a distance is determined; 2, on the basis of weight fusion of the gas concentration and a least square method, a fire source zone is positioned; 3, on the basis of a K-means cluster method, cluster analysis on the fire source zone positioned in the step 2 is carried out, and a fire source point is determined. The positioning method is applicable to conditions that a fire is at an early stage, a roof is flat, a wind speed is relatively small, and only a single fire source point exists, a smoke flow is taken as a two-dimensional plane, through the fire positioning method, a place below the zone of a positioning result is taken as a fire generation zone. The fire positioning method has advantages of low energy consumption, timeliness, low cost, interference resistance and strong practicality.

The invention discloses an NTP-based time synchronization method and a corresponding system and belongs to the field of digital information transmission. The method comprises the following steps of dividing terminal devices into a plurality of multicast groups and implementing the corresponding system of the method, wherein one multicast group has one terminal device to synchronize the time with a server in one period, and other terminal devices take the terminal device which synchronizes the time with the server as secondary server synchronization time. The existing terminal device is taken as a subordinate server, a problem of huge performance pressure generated when the terminal devices of hundreds of thousands of orders of magnitude utilize one server is effectively solved without increasing the server cost, the probability of failed time synchronization is reduced, and meanwhile, the data traffic in a video monitoring network is reduced, the network congestion is reduced, and the error accumulation caused by multistage NTP time synchronization servers is reduced.