138 results about "Adaptive identification" patented technology

Systems and methods for analyzing documents are provided herein. A plurality of documents and user input are received via a computing device. The user input includes hard coding of a subset of the plurality of documents, based on an identified subject or category. Instructions stored in memory are executed by a processor to generate an initial control set, analyze the initial control set to determine at least one seed set parameter, automatically code a first portion of the plurality of documents based on the initial control set and the seed set parameter associated with the identified subject or category, analyze the first portion of the plurality of documents by applying an adaptive identification cycle, and retrieve a second portion of the plurality of documents based on a result of the application of the adaptive identification cycle test on the first portion of the plurality of documents.

The invention discloses an industrial robot torque compensation control method based on load adaptive identification. The industrial robot torque compensation control method comprises the steps that according to robot overall stress balance analysis, a robot dynamics model under the load working condition is established in a deduced mode through a Newton-Euler method on the basis of considering friction; the optimal excitation trajectory of robot load identification is solved; load inertial parameters are calculated; and on the basis of load inertia matching, joint driving torque consumed by the load when a robot moves under the load working condition is compensated, and further verification and optimization are conducted through experiments. According to the industrial robot torque compensation control method, the influences of the load on operation performance of the robot are considered, under the premise that no any external sensor is needed, adaptive identification of any load andtorque compensation control are completed through easy operation easy to achieve, the control precision of the robot under the load working condition is effectively improved, and very important significance for processes such as carrying and friction stir welding of the industrial robot under high-speed and large-load movement is achieved.

The invention discloses an on-line identification and control method for a parameter of an alternating current position servo system model. According to the invention, a model reference adaptive identification algorithm based on a Lyapunov stability theory is utilized; and on-line identification is carried out on a rotary inertia J and a viscous damping coefficient B of a controlled object of an alternating current servo system; after convergence is carried out on an identification parameter, an on-line design on a position controller is carried out according to J and B values and the design is automatically switched to position control. According to the invention, an efficiency on an alternating current servo system design can be substantially improved.

The invention discloses a dynamic test system of a magnetic levitation system. The system mainly comprises an excitation unit, a node selection unit, a tracking band-pass filter unit, an adaptive identification unit and a frequency property extraction unit. The excitation unit is used for generating swept-frequency excitation signals. The node selection unit is used for selecting excitation inputnodes, superposing the swept-frequency excitation signals in a control system, and simultaneously selecting two measurement signal lead-out nodes according to a measured module. The tracking band-pass filter unit is used for filtering lead-out signals. The adaptive identification unit is used for performing Robust adaptive fitting of the processed node lead-out signals by using the swept-frequency excitation signals. The frequency property extraction unit is used for obtaining a frequency property curve of the measured module according to the fitting result. The system disclosed by the invention is integrated in a magnetic levitation control system; observable points are increased; a certain module in the magnetic levitation control system can be subjected to open-loop and closed-loop dynamic test; any equipment is unnecessary to be increased; the cost is saved; and the test efficiency is increased.

The invention provides a smoke detection and fire alarm method based on trajectory identification. The method includes the following steps of 1, video information preprocessing, wherein the video information preprocessing step sequentially includes a dark channel setting stage and an image partitioning stage to increase the thin smoke and environment contrast ratio and a video time information compression stage for transforming three-dimensional images into two-dimensional projection compression images through projection transformation and building a time compression trajectory, 2, smoke detection based on trajectory identification, wherein smoke detection is conducted on the video time compression trajectory obtained in step 1, and the trajectory-identification-based smoke detection step sequentially includes a background image removing stage, a movement trajectory extracting stage and a smoke feature detection stage and 3, smoke source positioning and smoke quantitative calculation, wherein smoke source positioning is conducted on the smoke detected in step 2, and the smoke amount is calculated according to a smoke source. The smoke detection and fire alarm method has the advantages of being easy to calculate, higher in identification accuracy, high in identification speed, good in adaptability and better in identification effect.

The invention discloses a fractional order adaptive rapid terminal sliding mode control method of a micro-gyroscope. The method comprises steps of establishing a dimensionless kinetic equation mathematical model and a reference trajectory model of a micro-gyroscope system; and constructing a fractional order-based adaptive rapid terminal sliding mode controller. According to the invention, an objective of real-time tracking of the micro-gyroscope can be achieved; errors can be rapidly converged into zero in the limited time; robustness of the system is enhanced; under the condition that thereis external interference, good performance can be kept; according to a fractional order adaptive rapid terminal sliding mode surface, a fractional order adaptive rule is designed; based on the Lyapunov stability criterion, an adaptive identification method is designed; all kinds of unknown system parameters of the micro-gyroscope are estimated in an online real-time manner; compared with an integer order, adjustable items are added, so control effects and parameter estimation effects are improved; and by use of the adaptive control method, system non-determinacy and parameters of the upper bound of the external disturbance sum are estimated, and automatic tracking of manufacturing errors and external disturbance is achieved.

The invention discloses a multi-target tracking and facial feature information identification method. The method comprises the following steps: firstly, inputting a facial video and converting the facial video into a human face key frame set SF; obtaining a data set SFD through classification and data enhancement SF; extracting SFD features by using an improved miniXaction model to obtain a self-adaptive optimization facial feature recognition model FFsmodel; extracting a key frame sequence in the face data stream FEV, and adaptively aggregating face tracking features to obtain a multi-targetface key frame position set FEC; loading an FFsmodel model and inputting the FFsmodel model into FEC to generate a multi-target facial feature classification result set FECR; and finally, opening an adaptive identification interface to process a terminal request, and storing an obtained processing result set of sign-in and target feature tracking state identification in a multi-target scene in a Web server. The improved multi-target tracking and self-adaptive facial feature recognition technology is combined, the picture feature result label with the highest accuracy can be effectively obtained, and the accuracy and use value of face tracking and facial feature recognition in a multi-target scene are improved.

The invention discloses a target vehicle lane changing behavior self-adaptive identification model construction method, which comprises the following steps of 1, establishing a lane changing behaviorinfluence parameter set, and dividing samples to obtain model input; 2, based on the large-scale vehicle trajectory data set, conducting training to obtain a general model for identifying left and right target lane changing behaviors; and 3, establishing an adaptive lane changing behavior recognition model based on Bayesian inference. According to the target vehicle lane changing behavior self-adaptive identification model construction method, the lane changing behaviors of surrounding target vehicles during driving can be effectively identified through setting of the steps 1 to 3 so that information support can be provided for driving behavior decision of the vehicles, and thus driving safety can be enhanced.

The invention provides a non-contact adaptive echo cancellation device and an echo cancellation method. The device comprises an adaptive identification filter (10) and a non-contact measurement device(7), wherein an output end of the non-contact measurement device (7) is connected with an input end of the adaptive identification filter (10), and the non-contact measurement device (7) emits laseror ultrasonic wave to a speaker diaphragm (6), and measures a reflected signal of the laser or the ultrasonic wave so as to acquire and use a signal reflecting vibration information of the speaker diaphragm as a reference signal of the adaptive identification filter (10). When the echo cancellation method is used to cancel echo based on the device, adverse effects of non-linear change of transmission from a speaker driving current signal to a sound signal on the adaptive identification filter can be prevented, and the convergence performance of the filter is improved; and the echo suppressionquantity in adaptive echo cancellation is increased, so that the system performance is improved.

The invention provides a non-intrusive load adaptive identification method based on a twin network, and the method comprises the steps: taking a V-I track and active power of a load as to-be-identified load features, discriminating the similarity of the V-I track of the load through the twin network, and obtaining the load number information through the matching with a feature library, thereby achieving the load identification. Wherein the user establishes the mapping between the load number information and the actual type of the electric appliance according to the stored use time and the actual use condition of the day. Through the dynamic construction of the feature library, the method can achieve the accurate recognition of the unknown load. Finally, the validity and universality of the model are verified in a PLAID data set and a COOLL data set.

The invention discloses a method and system for identifying the moment of inertia of a permanent magnet synchronous motor. The method comprises the following steps: obtaining alpha axis current, beta axis current, an alpha axis electromotive force, a beta axis electromotive force and an electromagnetic torque of a servo control system of the permanent magnet synchronous motor; inputting the alpha axis current, the beta axis current, the alpha axis electromotive force and the beta axis electromotive force into a speed observer, and obtaining an output rotor observation speed of the speed observer; and inputting the rotor observation speed and the electromagnetic torque into a model reference adaptive identification algorithm, and outputting the moment of inertia. According to the method disclosed by the invention, the output rotor observation speed of the speed observer and the obtained electromagnetic torque are used as the input of the model reference adaptive identification algorithm to obtain the moment of inertia, the speed observer can eliminate phase hysteresis caused by obtaining the speed through position and improve the identification accuracy, and the model reference adaptive identification algorithm can guarantee good stability for the identification system and improve the user experience.

The invention discloses a lithium battery equivalent circuit model parameter online robust adaptive identification method. The method comprises the following steps of: establishing an SOC-OCV relation curve according to a battery open-circuit voltage and a state of charge; initializing system parameters, and setting initial values of theta(0) and P(0), an initial SOC state, lambda and a sampling period T; calculating Soc(k) of the k-th step by adopting an ampere-hour integral method, wherein k=1, 2,..., n; according to the collected battery current and terminal voltage observation data, calculating to obtain a matrix by combining an SOC-OCV function relationship; starting an RLsign algorithm, and calculating a1, a2, a3, a4 and a5 according to an iterative formula; further solving parameters R0, R1, R2, C1 and C2 in a battery model; and circulating the steps to carry out online identification on the model parameters so as to realize dynamic adjustment of the battery model parameters. According to the lithium battery equivalent circuit model parameter online robust adaptive identification method, the influence of non-Gaussian noise on parameter identification can be effectively reduced, and the effectiveness of parameter estimation precision is ensured; and the algorithm is stable, the calculation complexity is low, and the steady-state convergence precision is higher than that of other similar traditional algorithms under the non-Gaussian noise condition.

The invention discloses a self-adaptive identification method for particles in an asphalt mixture CT image, aiming to solve the problem of low extraction accuracy of a microstructure of a mixture caused by a black core phenomenon in the center of a test piece in an existing asphalt mixture CT image. The self-adaptive identification method comprises the following steps: 1, determining the radius and circle center position of a test piece in an asphalt mixture CT image; 2, performing gap gray level zeroing processing on the CT image; 3, performing statistical gray distribution on the CT image after the gap gray return-to-zero processing along the radial direction; 4, carrying out the homogenization of the brightness of the image after the statistical gray scale is distributed in the radial direction; 5, performing filtering and noise reduction on the image subjected to the homogenization processing; 6, performing global threshold segmentation on the image subjected to filtering and noisereduction; and 7, performing image morphological processing on the image after global threshold segmentation. The self-adaptive identification method is used for digital image processing and road engineering.

The invention discloses a plantar reflex zone adaptive identification and treatment method of an intelligent foot biological health information management and hydrotherapy health care decision-making system. The method comprises the following steps: a user puts feet on a water outlet board of a hydrotherapy instrument, and a sensing device arranged on the water outlet board collects response data of the user when the user uses the hydrotherapy instrument; the foot outline of the user is identified and drawn according to the response data, the foot outline is compared with a standard foot outline, and the actual plantar reflex zone positions of the user are calculated out according to the plantar reflex zone positions on the standard foot outline; and hydrotherapy heads corresponding to the actual plantar reflex zone positions of the user are adaptively selected to implement stimulation therapy on the plantar reflex zones of the user. Through the method, the foot outline and the plantar reflex zone positions of a user can be identified automatically, and the hydrotherapy heads in corresponding positions are automatically aligned with the reflex zone positions under control to implement stimulation therapy. The method has a wider range of application. An ideal therapeutic effect is achieved for users of different foot sizes. Family members can share one hydrotherapy instrument.

The invention discloses a high-adaptability identification counting method for multiple fields, which is suitable for counting targets and comprises the following steps of: inputting an to-be-detectedimage; analyzing a to-be-detected region of the to-be-detected image based on gray information and shape information to obtain a target area ROI; extracting array features of a ROI of the target area; judging whether the target is in regular array distribution or not according to the array characteristics, and counting the target according to a judgment result to obtain a counting value of a target. The method can automatically search a target and perform counting operation so as to meet the counting requirements of different fields, does not need repeated development and deployment, is highin adaptability and counting efficiency, can greatly shorten the development period of the counting method, and effectively improves the productivity.

The invention discloses a radar data filtering method based on a thermodynamic diagram and graph theory clustering. The method comprises the following steps: 1, obtaining and caching radar data, and mapping the radar data to Heat Map to generate a corresponding thermodynamic diagram; 2, analyzing the occurrence frequency and condition of points in each region by observing the thermodynamic diagram, setting a threshold value, and finding out and marking all points near the threshold value; 3, after a boundary threshold point set needing to be processed is obtained, the boundaries of the points are found out through Alpha Shapes, and a fixed noise point generation area is obtained; and step 4, after boundary point cluster sets are obtained through calculation by an Alpha Shapes algorithm, whether points thrown by the radar fall into the boundary point cluster sets is judged, and radar data preprocessing is completed. The method is suitable for vertex radar data processing of multiple noise points in a complex environment, has good adaptability to mobile radar data or radar data in an abrupt change environment by automatically recognizing a noise source area in real time, and is high in recognition efficiency, simple and effective.