347 results about "Dynamic time warping" patented technology

The invention discloses a parking system path planning method based on dynamic time windows, and belongs to the technical field of path planning. The method is characterized by comprising the following steps: S1, building a work environment model of AGVs in an intelligent garage in a topological method; S2, setting priority for each AGV and each car parking / picking task according to different evaluation criteria; S3, using a Dijkstra algorithm to plan a shortest possible path for an AGV accepting a task; S4, arranging feasible path time windows; S5, designing conflict resolution strategies according to different types of conflicts; and S6, planning a conflict-free optimal path for the AGV using a parking system path planning algorithm based on dynamic time windows. A time-sharing use strategy is used, and the Dijkstra algorithm and a time window method are combined effectively, so that the problem that the existing multi-AGV path planning is of poor flexibility and is prone to deadlock or collision conflict is solved effectively, and a shortest conflict-free optimal path can be planned for an AGV accepting a task. In addition, the overall operation efficiency of an intelligent three-dimensional parking system can be improved effectively, and the car parking / picking waiting time can be reduced for social members.

Systems and methods for recognizing the gestures of an entity, notably a human being, and, optionally, for controlling an electrical or electronic system or apparatus, are discussed. The system uses sensors that measure signals, preferentially representative of inertial data about the movements of said entity, and implements a process for enriching a dictionary of said gestures to be recognized and a recognition algorithm, for recognition among the classes of gestures in said dictionary. The algorithm implemented is preferably of the dynamic time warping type. The system carries out preprocessing operations, such as the elimination of signals captured during periods of inactivity of the entity, subsampling of the signals, and normalization of the measurements by reduction, and preferentially uses, to classify the gestures, specific distance calculation modes and modes for merging or voting between the various measurements by the sensors.

The invention discloses a gesture recognition method based on an acceleration sensor. The gesture recognition method based on an acceleration sensor comprises the following steps: automatically collecting gesture acceleration data, preprocessing, calculating the similarity of all gesture sample data so as to obtain a similarity matrix, extracting a gesture template, constructing a gesture dictionary by utilizing the gesture template, and carrying out sparse reconstruction and gesture classification on the gesture sample data to be recognized by adopting an MSAMP (modified sparsity algorithm adaptive matching pursuit) algorithm. According to the invention, the compressed sensing technique and a traditional DTW (dynamic time warping) algorithm are combined, and the adaptability of the gesture recognition to different gesture habits is improved, and by adopting multiple preprocessing methods, the practicability of the gesture recognition method is improved. Additionally, the invention also discloses an automatic collecting algorithm of the gesture acceleration data; the additional operation of traditional gesture collection is eliminated; the user experience is improved; according to the invention, a special sensor is not required, the gesture recognition method based on the acceleration sensor can be used for terminals carried with the acceleration sensor; the hardware adaptability is favorable, and the practicability of the recognition method is enhanced. The coordinate system is uniform, and can be adaptive to different multiple gesture habits.

The invention relates to a human body skeleton-based action recognition method. The method is characterized by comprising the following basic steps that: step 1, a continuous skeleton data frame sequence of a person who is executing target actions is obtained from a somatosensory device; step 2, main joint point data which can characterize the actions are screened out from the skeleton data; step 3, action feature values are extracted from the main joint point data and are calculated, and a feature vector sequence of the actions is constructed; step 4, the feature vectors are preprocessed; step 4, the feature vector sequence of an action sample set is saved as an action sample template library; step 6, actions are acquired in real time, the distance value of the feature vector sequence of the actions and the feature vector sequence of all action samples in the template library is calculated by using a dynamic time warping algorithm; and step 7, the actions are classified and recognized. The method of the invention has the advantages of high real-time performance, high robustness, high accuracy and simple and reliable implementation, and is suitable for a real-time action recognition system.

A method for comparing data sequences includes accepting first and second data sequences of data elements. A distance matrix is computed. The matrix includes rows and columns of matrix elements, describing distances between the data elements of the first sequence and the data elements of the second data sequence. The distance matrix is reshaped by applying successive, incremental shifts to the rows or columns so as to produce a reshaped matrix. A best-score path through the reshaped matrix is calculated using vector operations, so as to quantify a similarity between the first and second data sequences. Due to vectorization, a significant increase in computation speed is achieved in both software and hardware implementations.

A method and a system for determining a product vector for computation of a Euclidean distance for performing Dynamic Time Warping of a test signal and a template signal are provided. Low-rank factorized vectors are determined for the template signal. The low-rank factorized vectors are processed along with the test signal for determining the product vector. The product vector is thereafter usable for the determination of a Euclidean distance between the test signal and the template signal, and for performing dynamic time warping of the test signal and the template signal.

The present invention relates to an object detection and behavior recognition system using three-dimensional motion data. The system receives three-dimensional (3D) motion data of a scene from at least one sensor, such as a LIDAR sensor. An object is identified in the 3D motion data. Thereafter, an object track is extracted, the object track being indicative of object motion in the scene over time. Through Dynamic Time Warping (DTW) or other comparison techniques, the object track is compared to a database to identify the behavior of the object based on its object track.

An equipment fault diagnosis method successively comprises fault diagnosis training steps and fault diagnosis operation steps. The method focuses on a data form characteristic essence and combines a multi-dimensional segmentation fitting algorithm and an optimized dynamic time warping algorithm based on equipment fault data to carry out pattern expression of modeling and distance threshold extraction, and extract form characteristics from the found equipment abnormity data to carry out mode matching, so that an equipment data fault type can be identified and a reason diagnosis function is realized. In this way, the difficulties in current fault diagnosis technologies that similarity degrees among the fault data can hardly be described efficiently and accurately can be solved.

Factor retrieving is a major approach for pulse wave analysis. Stiffness index and cardiac output are widely used factors for cardiac risk detection. Research has been done on clinical pulse wave data which are collected by pulse oximeter. The result shows that collected factors have a positive correlation with certain cardiac risks. Some adjustments have been applied on the algorithms that increase the significance. In addition to the factor based analysis, other signal processing techniques for pulse waveforms are included such as bispectrum estimation, Wavelet transform, and weighted dynamic time warping. Bispectrum estimation and Wavelet transform have meaningful features of pulse waveforms with some special shapes. Weighted dynamic time warping compares the similarity of waveforms. It also includes medical significance into the calculation by adjusting the weight vector. This algorithm has higher accuracy when providing more samples to compare. The factor based analysis and waveform analysis compose an analytic model which can be used for risk evaluation, classification and disease detection.

The invention belongs to the field of load power use monitoring and energy management, and particularly relates to a method and system for identifying transient process of a non-intrusive load. The method comprises the following steps: by virtue of transient power waveform characteristics of multiple types, measuring similarity of a transient power waveform characteristic parameter sample sequence of an unfixed length and a template sequence by using a dynamic time warping algorithm within a time domain, on the basis, establishing a load transient process classification identification scheme based on a nearest neighbor classification strategy, and confirming power use equipment with the load transient process, thereby achieving artificial working state identification on the non-intrusive type power use equipment. The method and the system have the beneficial effects that accuracy and robustness of power load transient process identification can be improved, the cost of a monitoring system can be effectively controlled, and the practicability of the monitoring system can be effectively improved, therefore, practical popularization of an NILM (Non-Intrusive Load monitoring) technique can be greatly improved.

The invention discloses a multidimensional weighted 3D recognition method for dynamic gestures. At the training stage, firstly, standard gestures are segmented to obtain a feature vector of the standard gestures; secondly, coordinate system transformation, normalization processing, smoothing processing, downsampling and differential processing are performed to obtain a feature vector set of the standard gestures, weight values of all joint points and weight values of all dimensions of elements in the feature vector set, and in this way, a standard gesture sample library is constructed. At the recognition stage, by the adoption of a multidimensional weighted dynamic time warping algorithm, the dynamic warping distances between the feature vector set Ftest of the gestures to be recognized and feature vector sets Fc =1,2,...,C of all standard gestures in the standard gesture sample library are calculated; when the (m, n)th element S(m, n) of a cost matrix C is calculated, consideration is given to the weight values of all the joint points and the weight values of all the dimensions of the elements, the joint points and coordinate dimensions making no contribution to gesture recognition are removed, in this way, the interference on the gesture recognition by joint jittering and false operation of the human body is effectively removed, the anti-interference capacity of the algorithm is enhanced, and finally the accuracy and real-time performance of the gesture recognition are improved.

The invention relates to a method for extracting and identifying characteristics of an electro-ocular signal, which is characterized by comprising three stages, namely pretreatment of the electro-ocular signal, extraction of characteristic parameters of the electro-ocular signal and mode identification of the electro-ocular signal, wherein in the pretreatment stage, endpoint detection and bandpass filtering on the electro-ocular signal are carried out; in the characteristic parameter extraction stage, frame separation and windowing on the electro-ocular signal are carried out, and an electro-ocular characteristic parameter sequence changed along with time after converting the continuous electro-ocular signal into a plurality of sections of short-time electro-ocular signals is extracted; and in the mode identification stage, similarity comparison between the input electro-ocular characteristic parameter sequence and each template of a template base sequentially through a dynamic time normalization method is carried out so as to judge the corresponding eye movement of the operator. The method has the characteristics of high identification accuracy of the electro-ocular signal, certain antijamming capacity, strong application value and the like.

The invention discloses a 4D track generating method. The method includes the following steps of receiving aircraft meteorological data sent by an aircraft to download AMDAR data, conducting testing and quality control on the AMDAR data, generating a flight height profile and a flight speed profile according to the track historical information in the AMDAR data and the large elliptic distance method, generating a nominal flight profile as an initial 4D track generating profile according to the flight height profile and the flight speed profile and through a dynamic time warping (DTW) algorithm and an aircraft basic performance database (BADA) model, modifying weather forecast GRIB data according to actually-measured meteorological data in the AMDAR data, and calculating aircraft airway point-to-point time and combining the aircraft airway point-to-point time with the nominal flight profile to generate a 4D track.

The invention discloses an identification method for a horn of a special vehicle based on dynamic time warping (DTW) and hidden markov model (HMM) evidence integration. The identification method for the horn of the special vehicle based on dynamic time warping (DTW) and hidden markov model (HMM) evidence integration comprise a first step of building a vehicle-horn sample library. A second step of preprocessing step. A third step of extracting and dimensionality reduction disposing parameters of vehicle-horn characteristics. A fourth step of identifying the horn of the special vehicle based on the evidence integration, and gaining the DTW identification result and the HMM identification result by respectively adopting DTW algorithm and HMM algorithm. If the DTW identification result and the HMM identification result are consistent, the final identification result is kept consistent with the DTW identification result or the HMM identification result. If the DTW identification result and the HMM identification result are different, the final identification result should be output by an identification decision reasoning with a data set (DS) evidence theory. The identification method for the horn of the special vehicle based on dynamic time warping (DTW) and hidden markov model (HMM) evidence integration adopts integration identification technology and identification rate is high.

The invention discloses an indoor fusion positioning method based on multi-features. The whole process of the method is divided into three parts such as Wi-Fi fingerprint probability calculation, inertial navigation calculation and particle filtering fusion, wherein the Wi-Fi fingerprint probability calculation uses a cosine similarity matching algorithm, the inertial navigation calculation is divided into three parts such as step number calculation, step size calculation and walking direction estimation, the step number calculation uses methods such as walking detection, center moving equating, window wave peak detection, dynamic time regularization, derivative analysis filtering and wave peak resurrection, the step size calculation uses a step size and step frequency relationship model, the walking direction estimation uses the time and frequency zone analysis of a horizontal direction acceleration and strategy adjusting based on a gravity direction, and finally in the particle filtering fusion part, the hidden markov model constructed based on the inertial navigation and the Wi-Fi fingerprint probability uses the particle filtering method and combines with the map information to calculate the position probability so as to finally obtain the positioning result.

The invention discloses a quick human movement identification method oriented to human-computer interaction. The method comprises the following steps that: (1) collecting bone point coordinate information; (2) selecting a key point; (3) extracting movement characteristics; (4) carrying out movement identification; and (5) using a robot. The invention provides the quick human movement identification method oriented to the human-computer interaction. The whole system consists of a terminal computer, Kinect human movement input equipment, a Bluetooth communication module and a robot. Firstly, theKinect is used for capturing a human body, effective nodes which can represent whole-body movement are extracted from 20 joint nodes and are calculated to obtain movement characteristics, and a movement template is formed and stored into a TXT text. In an identification stage, a movement sequence to be tested and a standard template are quickly matched through an F-DTW (Fast Dynamic Time Warping)algorithm, and an identification result is given. According to a movement identification result, the robot makes different responses. By use of a quick algorithm, movement identification speed is greatly improved, and quickly control for the robot is optimized.

The invention relates to a voice keyword retrieval method based on an audio template. The method comprises the following steps: to begin with, converting a voice sample template and voice to be retrieved into a probability distribution sequence; then, carrying out matching on the voice sample template and the voice to be retrieved through dynamic time warping to obtain keyword starting and ending time point in the voice to be retrieved and acoustic confidence measure scores of each appearance position; and finally, carrying out warping on the scores obtained by different voice sample templates, and carrying out ranking to obtain a retrieval result. In the retrieval process, information of a specific language is not required at all, thereby maximizing universality and transportability; and meanwhile, operation amount in the retrieval process is reduced, and keyword retrieval speed is accelerated.

Input handwritten characters are classified as print or cursive based upon numerical feature values calculated from the shape of an input character. The feature values are applied to inputs of an artificial neural network which outputs a probability of the input character being print or cursive. If a character is classified as print, it is analyzed by a print character recognizer. If a character is classified as cursive, it is analyzed using a cursive character recognizer. The cursive character recognizer compares the input character to multiple prototype characters using a Dynamic Time Warping (DTW) algorithm.

The invention relates to a performance evaluation method for a rolling bearing under multiple operating conditions based on geometric measurement. The performance evaluation method comprises the steps of firstly converting rolling bearing vibration signals with different fault degrees under different operating conditions into a recurrence plot, then stable extracting geometric features of the converted recurrence plot by using scale invariant feature transform method, carrying out singular value decomposition on the geometric features extracted from the bearing signals to acquire feature curves of the rolling bearing with different fault degrees under the different operating conditions, and then calculating the similarity of different feature curves by using dynamic time warping so as to work out a health degree of the rolling bearing. Case verification is carried out by using bearing data of the Case Western Reserve University in the United States, and an experimental result shows that the performance evaluation method can evaluate the health degree of the rolling bearing under multiple operating conditions effectively.

The invention discloses a detection and matching mechanism for recognition of handwritten letters using WiFi signals. The invention mainly makes use of the characteristics of wireless signal multipath propagation, and acquires CSI (Channel State Information) data that carry information reflecting external environmental characteristics, including hand movements when letters are written. The CSI data after de-noising (via Butterworth filtering and principal component analysis filtering) show that in the stage of writing letters, the CSI data fluctuate significantly, and the CSI data are relatively stable in the stage without handwriting, and on the basis of the characteristics, the invention provides a waveform feature extraction algorithm to detect and extract waveforms only including information on 26 handwritten letters. Finally, the invention provides a solution combining feature matching and context error correction, takes the waveform as a matching feature, carries out machine learning classification for the extracted 26 feature waveforms, matches the waveforms by using a DTW (dynamic time warping), and improves the recognition accuracy of the system through the context error correction algorithm.

The invention discloses an isolated word speech recognition method based on an HRSF (Half Raised Sine Function) and an improved DTW (Dynamic Time Warping) algorithm. The isolated word speech recognition method comprises the following steps that (1), a received analog voice signal is preprocessed; preprocessing comprises pre-filtering, sampling, quantification, pre-emphasis, windowing, short-time energy analysis, short-time average zero crossing rate analysis and end-point detection; (2), a power spectrum X(n) of a frame signal is obtained by FFT (Fast Fourier Transform) and is converted into a power spectrum under a Mel frequency; an MFCC (Mel Frequency Cepstrum Coefficient) parameter is calculated; the calculated MFCC parameter is subjected to HRSF cepstrum raising after a first order difference and a second order difference are calculated; and (3), the improved DTW algorithm is adopted to match test templates with reference templates; and the reference template with the maximum matching score serves as an identification result. According to the isolated word speech recognition method, the identification of a single Chinese character is achieved through the improved DTW algorithm, and the identification rate and the identification speed of the single Chinese character are increased.

The present invention discloses an embedded index based hydrological time series similarity searching method. The method is carried out by the following steps: calculating corresponding embedded index vectors of each position in an original time series in an off-line preparation stage, wherein in the off-line preparation stage, the hydrological time series flood peak segmentation, serial clustering, initial reference sequence set generation, reference set training and time series embedded index calculation are realized; and calculating the index vectors by using a query sequence and a reference set sequence in an on-line searching stage, performing searching in an embedded index Euclidean vector space of the original series to find a relatively similar point as a candidate point set, refining candidate points, and then performing original DTW (dynamic time warping) measurement to find a final similar sequence. According to the embedded index based hydrological time series similarity searching method, the similarity searching is mapped to the Euclidean vector space to perform searching, thereby greatly improving the search efficiency.

Provided is an encoding device (1) including: a pitch contour analysis unit (101) which detects information, a dynamic time-warping unit (102) which generates, based on the information, pitch change ratios (Tw_ratio in FIG. 18) within a range (86) including a range (86a) of the pitch change ratios corresponding to absolute pitch differences of 42 cents or larger; a first lossless coding unit (103) which codes the generated pitch parameters (102x); a time-warping unit (104) which shifts a pitch of a signal according to the information; and a second encoding unit which codes a signal (104x) obtained by the shifting.

A method of generating synthetic data from time series data, such as from handwritten characters, words, sentences, mathematics, and sketches that are drawn with a stylus on an interactive display or with a finger on a touch device. This computationally efficient method is able to generate realistic variations of a given sample. In a handwriting or sketch recognition context, synthetic data is generated from real data in order to train recognizers and thus improve recognition accuracy when only a limited number of samples are available. Similarly, synthetic data can also be used to test and validate such recognizers. Also discussed is a dynamic time warping based approach for both segmented and continuous data that is designed to be a robust, go-to method for gesture recognition across a variety of modalities using only limited training samples.

The invention provides a continuous character gesture track recognizing method based on a structural dynamic time domain wrapping algorithm (STDW) for solving the problems in the prior art. Through the automatic cutting problem of continuous gesture tracks, extraction of angular point new characteristics, the recognition of the STDW and the processing between character gesture tracks, natural, real-time and accurate recognition on continuous character gesture tracks is achieved. Compared with the prior art, the interaction mode is more natural, the constraint of users is smaller, and the requirement of man-machine interaction is met. Compared with the traditional situation that multiple gesture actions are more diversified in type, under the condition that accuracy and real-time performance are considered at the same time, the recognition effect of an original algorithm is improved by means of the STDW formed by combining a character gesture track structure and a dynamic time wrapping algorithm, the recognition rate is increased, and the method is high in accuracy and good in man-machine interaction experience and can meet the requirement for real-time performance.

The invention discloses a failure diagnosis and health evaluation method based on wavelet power, manifold dimension reduction and dynamic time warping, and aims to improve the feature separability of bearing failure, impeller failure and the mixed failures of a centrifugal pump and realize diagnosis and health evaluation of various states. The method comprises the following steps: firstly, decomposing collected vibration signals of the centrifugal pump into 8 wavelet components by applying wavelet packet conversion; extracting wavelet energy of each component to be taken as a failure feature to obtain an eight-dimensional failure feature vector; then conducting dimension reduction on the eight-dimensional feature by applying a manifold learning method to obtain a three-dimensional feature vector with better separability, simplicity and stability; finally, based on the feature vector, measuring the distance of test data and training data by applying a dynamic time normalization method so as to determine the current failure state and realize failure diagnosis of a bearing. The distance value can also reflect the health degree of the current state, can realize evaluation of the health state of the centrifugal pump, and has the excellent practical engineering application value.

The embodiment of the invention discloses a method and device for determining similar audios. The method includes the steps of determining an appointed audio characteristic value sequence of target audios; according to the dynamic time warping, respectively calculating DTW distances between the appointed audio characteristic value sequence of the target audios and an appointed audio characteristic value sequence of pre-determined N basic audios; determining the obtained N DTW distances to be audio fingerprints of the target audios; according to a preset formula, calculating the similarity between the audio fingerprints of the target audios and audio fingerprints of standard audios; if the similarity between the audio fingerprints of the target audios and the audio fingerprints of the standard audios is larger than a preset pre-threshold value, determining that the target audios are similar to the standard audios. Compared with the prior art, generation of a large number of feature vectors is not needed, so that in the overall audio fingerprint matching process, storage and searching of a large number of features are not needed, and the robot resource cost is small. In addition, the problem that the local robustness in the prior art is not high enough can be solved, and the overall robustness can be improved.

The invention belongs to the field of data mining and discloses an intra-domain traffic flow pattern discovery method based on an improved similarity search technology in order to perform data mining on the intra-domain specific traffic flow pattern in a better way. The method is mainly based on the improved DTW (dynamic time warping) similarity search technology, and the calculation overhead of the DTW technology is reduced. The core of the improved DTW similarity search technology includes the combination of lower bound function technology and role exchange concept, also increases the calculation early termination threshold Epsilon on this basis and reduced the calculation of bound exceeding. According to the intra-domain traffic flow pattern discovery method based on improved similarity search technology, the calculation overhead of the DTW technology, the speed and accuracy rate of intra-domain specific traffic flow pattern are accordingly improved. The study and judgment of the current specific pattern state are finished through the information migration, so that the recurrence of significant threatening events in urban road safety is prevented.