Showing papers on "Dynamic time warping published in 2022"

Unsupervised hierarchical methodology of maritime traffic pattern extraction for knowledge discovery

Abstract: Owing to the space–air–ground integrated networks (SAGIN), seaborne shipping has attracted increasing interest in the research on the motion behavior knowledge extraction and navigation pattern mining problems in the era of maritime big data for improving maritime traffic safety management. This study aims to develop a novel unsupervised methodology for feature extraction and knowledge discovery based on automatic identification system (AIS) data, allowing for seamless knowledge transfer to support trajectory data mining. The unsupervised hierarchical methodology is constructed from three parts: trajectory compression, trajectory similarity measure, and trajectory clustering. In the first part, an adaptive Douglas–Peucker with speed (ADPS) algorithm is created to preserve critical features, obtain useful information, and simplify trajectory information. Then, dynamic time warping (DTW) is utilized to measure the similarity between trajectories as the critical indicator in trajectory clustering. Finally, the improved spectral clustering with mapping (ISCM) is presented to extract vessel traffic behavior characteristics and mine movement patterns for enhancing marine safety and situational awareness. Comprehensive experiments are conducted and implemented in the Chengshan Jiao Promontory in China to verify the feasibility and effectiveness of the novel methodology. Experimental results show that the proposed methodology can effectively compress the trajectories, determine the number of clusters in advance, guarantee the clustering accuracy, and extract useful navigation knowledge while significantly reducing the computational cost. The clustering results are further explored and follow the Gaussian mixture distribution, which can help provide new discriminant criteria for trajectory clustering.

A Fused Method of Machine Learning and Dynamic Time Warping for Road Anomalies Detection

Abstract: To discover the condition of roads, a large number of detection algorithms have been proposed, most of which apply machine learning methods by time and frequency processing in acceleration and velocity data. However, few of them pay attention to the similarity of the data itself when the vehicle passes over the road anomalies. In this article, we propose a method to detect road anomalies by comparing the data windows with various length using Dynamic Time Warping(DTW) method. We propose a model to prove that the maximum acceleration of a vehicle passing through a road anomaly is linear with the height of the road barrier, and it’s verified by an experiment. This finding suggests that it is reasonable to divide the window by threshold detection. We also apply a brief random forest filter to roughly distinguish normal windows from anomaly windows using the aforementioned theory, in order to reduce the time consumption. From our study, a system is proposed that utilizes a series of acceleration data to discover where might be anomalies on the road, named as Quick Filter Based Dynamic Time Warping (QFB-DTW). We show that our method performs clearly beyond some existing methods. To support this conclusion, experiments are conducted based on three data sets and the results are statistically analyzed. We expect to lay the first step to some new thoughts to the field of road anomalies detection in subsequent work.

DeFall: Environment-Independent Passive Fall Detection Using WiFi

Abstract: Fall is recognized as one of the most frequent accidents among elderly people. Many solutions, either wearable or noncontact, have been proposed for fall detection (FD) recently. Among them, WiFi-based noncontact approaches are gaining popularity due to the ubiquity and noninvasiveness. The existing works, however, usually rely on labor-intensive and time-consuming training before it can achieve a reasonable performance. In addition, the trained models often contain environment-specific information and, thus, cannot be generalized well for new environments. In this article, we propose DeFall, a WiFi-based passive FD system that is independent of the environment and free of prior training in new environments. Unlike previous works, our key insight is to probe the physiological features inherently associated with human falls, i.e., the distinctive patterns of speed and acceleration during a fall. DeFall consists of an offline template-generating stage and an online decision-making stage, both taking the speed estimates as input. In the offline stage, augmented dynamic time-warping (DTW) algorithms are performed to generate a representative template of the speed and acceleration patterns for a typical human fall. In the online phase, we compare the patterns of the real-time speed/acceleration estimates against the template to detect falls. To evaluate the performance of DeFall, we built a prototype using commercial WiFi devices and conducted experiments under different settings. The results demonstrate that DeFall achieves a detection rate above 95% with a false alarm rate lower than 1.50% under both line-of-sight (LOS) and non-LOS (NLOS) scenarios with one single pair of transceivers. Extensive comparison study verifies that DeFall can be generalized well to new environments without any new training.

Towards Similarity-Aware Time-Series Classification

Abstract: We study time-series classification (TSC), a fundamental task of time-series data mining. Prior work has approached TSC from two major directions: (1) similarity-based methods that classify time-series based on the nearest neighbors, and (2) deep learning models that directly learn the representations for classification in a data-driven manner. Motivated by the different working mechanisms within these two research lines, we aim to connect them in such a way as to jointly model time-series similarities and learn the representations. This is a challenging task because it is unclear how we should efficiently leverage similarity information. To tackle the challenge, we propose Similarity-Aware Time-Series Classification (SimTSC), a conceptually simple and general framework that models similarity information with graph neural networks (GNNs). Specifically, we formulate TSC as a node classification problem in graphs, where the nodes correspond to time-series, and the links correspond to pair-wise similarities. We further design a graph construction strategy and a batch training algorithm with negative sampling to improve training efficiency. We instantiate SimTSC with ResNet as the backbone and Dynamic Time Warping (DTW) as the similarity measure. Extensive experiments on the full UCR datasets and several multivariate datasets demonstrate the effectiveness of incorporating similarity information into deep learning models in both supervised and semi-supervised settings. Our code is available at https://github.com/daochenzha/SimTSC

Motion Similarity Evaluation between Human and a Tri-Co Robot during Real-Time Imitation with a Trajectory Dynamic Time Warping Model

Abstract: Precisely imitating human motions in real-time poses a challenge for the robots due to difference in their physical structures. This paper proposes a human–computer interaction method for remotely manipulating life-size humanoid robots with a new metrics for evaluating motion similarity. First, we establish a motion capture system to acquire the operator’s motion data and retarget it to the standard bone model. Secondly, we develop a fast mapping algorithm, by mapping the BVH (BioVision Hierarchy) data collected by the motion capture system to each joint motion angle of the robot to realize the imitated motion control of the humanoid robot. Thirdly, a DTW (Dynamic Time Warping)-based trajectory evaluation method is proposed to quantitatively evaluate the difference between robot trajectory and human motion, and meanwhile, visualization terminals render it more convenient to make comparisons between two different but simultaneous motion systems. We design a complex gesture simulation experiment to verify the feasibility and real-time performance of the control method. The proposed human-in-the-loop imitation control method addresses a prominent non-isostructural retargeting problem between human and robot, enhances robot interaction capability in a more natural way, and improves robot adaptability to uncertain and dynamic environments.

D-LSTM: Short-Term Road Traffic Speed Prediction Model Based on GPS Positioning Data

Abstract: Short-term road traffic speed prediction is a long-standing topic in the area of Intelligent Transportation System. Apparently, effective prediction of the traffic speed on the road can not only provide timely details for the navigation system concerned and help the drivers to make better path selection, but also greatly improve the road supervision efficiency of the traffic department. At present, some researches on speed prediction based on GPS data, by adding weather and other auxiliary information, using graph convolutional neural network to capture the temporal and spatial characteristics, have achieved excellent results. In this paper, the problem of short-term traffic speed prediction based on GPS positioning data is further studied. For the processing of time series, we innovatively introduce Dynamic Time Warping algorithm into the problem and propose a Long Short-Term Memory with Dynamic Time Warping (D-LSTM) model. D-LSTM model, which integrates Dynamic Time Warping algorithm, can fine-tune the time feature, thus adjusting the current data distribution to be close to the historical data. More importantly, the fine-tuned data can still get a distinct improvement without special treatment of holidays. Meanwhile, considering that the data under different feature distributions have different effects on the prediction results, attention mechanism is also introduced in the model. Our experiments show that our proposed model D-LSTM performs better than other basic models in many kinds of traffic speed prediction problems with different time intervals, and especially significant in the traffic speed prediction on weekends.

Recognition of Upper Limb Action Intention Based on IMU

Abstract: Using motion information of the upper limb to control the prosthetic hand has become a hotspot of current research. The operation of the prosthetic hand must also be coordinated with the user’s intention. Therefore, identifying action intention of the upper limb based on motion information of the upper limb is key to controlling the prosthetic hand. Since a wearable inertial sensor bears the advantages of small size, low cost, and little external environment interference, we employ an inertial sensor to collect angle and angular velocity data during movement of the upper limb. Aiming at the action classification for putting on socks, putting on shoes and tying shoelaces, this paper proposes a recognition model based on the Dynamic Time Warping (DTW) algorithm of the motion unit. Based on whether the upper limb is moving, the complete motion data are divided into several motion units. Considering the delay associated with controlling the prosthetic hand, this paper only performs feature extraction on the first motion unit and the second motion unit, and recognizes action on different classifiers. The experimental results reveal that the DTW algorithm based on motion unit bears a higher recognition rate and lower running time. The recognition rate reaches as high as 99.46%, and the average running time measures 8.027 ms. In order to enable the prosthetic hand to understand the grasping intention of the upper limb, this paper proposes a Generalized Regression Neural Network (GRNN) model based on 10-fold cross-validation. The motion state of the upper limb is subdivided, and the static state is used as the sign of controlling the prosthetic hand. This paper applies a 10-fold cross-validation method to train the neural network model to find the optimal smoothing parameter. In addition, the recognition performance of different neural networks is compared. The experimental results show that the GRNN model based on 10-fold cross-validation exhibits a high accuracy rate, capable of reaching 98.28%. Finally, the two algorithms proposed in this paper are implemented in an experiment of using the prosthetic hand to reproduce an action, and the feasibility and practicability of the algorithm are verified by experiment.

Wetland Vegetation Classification through Multi-Dimensional Feature Time Series Remote Sensing Images Using Mahalanobis Distance-Based Dynamic Time Warping

Abstract: Efficient methodologies for vegetation-type mapping are significant for wetland’s management practices and monitoring. Nowadays, dynamic time warping (DTW) based on remote sensing time series has been successfully applied to vegetation classification. However, most of the previous related studies only focused on Normalized Difference Vegetation Index (NDVI) time series while ignoring multiple features in each period image. In order to further improve the accuracy of wetland vegetation classification, Mahalanobis Distance-based Dynamic Time Warping (MDDTW) using multi-dimensional feature time series was employed in this research. This method extends the traditional DTW algorithm based on single-dimensional features to multi-dimensional features and solves the problem of calculating similarity distance between multi-dimensional feature time series. Vegetation classification experiments were carried out in the Yellow River Delta (YRD). Compared with different classification methods, the results show that the K-Nearest Neighbors (KNN) algorithm based on MDDTW (KNN-MDDTW) has achieved better classification accuracy; the overall accuracy is more than 90%, and kappa is more than 0.9.

Uncertainty-DTW for Time Series and Sequences

Abstract: Dynamic Time Warping (DTW) is used for matching pairs of sequences and celebrated in applications such as forecasting the evolution of time series, clustering time series or even matching sequence pairs in few-shot action recognition. The transportation plan of DTW contains a set of paths; each path matches frames between two sequences under a varying degree of time warping, to account for varying temporal intra-class dynamics of actions. However, as DTW is the smallest distance among all paths, it may be affected by the feature uncertainty which varies across time steps/frames. Thus, in this paper, we propose to model the so-called aleatoric uncertainty of a differentiable (soft) version of DTW. To this end, we model the heteroscedastic aleatoric uncertainty of each path by the product of likelihoods from Normal distributions, each capturing variance of pair of frames. (The path distance is the sum of base distances between features of pairs of frames of the path.) The Maximum Likelihood Estimation (MLE) applied to a path yields two terms: (i) a sum of Euclidean distances weighted by the variance inverse, and (ii) a sum of log-variance regularization terms. Thus, our uncertainty-DTW is the smallest weighted path distance among all paths, and the regularization term (penalty for the high uncertainty) is the aggregate of log-variances along the path. The distance and the regularization term can be used in various objectives. We showcase forecasting the evolution of time series, estimating the Fréchet mean of time series, and supervised/unsupervised few-shot action recognition of the articulated human 3D body joints.

A Multistep Paroxysmal Atrial Fibrillation Scanning Strategy in Long-Term ECGs

Abstract: Atrial fibrillation (AF) is a progressive disease often initially manifested by intermittent episodes spontaneously terminating and is an insidious disease. The previous work trained the support vector machine (SVM) classifier on multiple RR interval features. The trained AF detector was tested on the fourth China Physiological Signal Challenge (CPSC 2021) database, achieving 97.59% and 89.83% for sensitivity and specificity on dataset 1, respectively. The test results were 96.46% and 78.64% on dataset 2, respectively. The results show that the AF detector based on rhythm is insufficient for the recognition of non-AF. Therefore, this work developed AF scanning algorithm integrating rhythm and P-wave information for long-term ECGs. The proposed algorithm is divided into three steps. First, utilize $a$ priori knowledge to locate suspected AF, and then, employ a trained rhythm-based AF detector to detect AF. Finally, adopt the dynamic time warping (DTW) and an autoencoding (AE) network to quantize the P-wave information to identify the non-AF signal from AF. The results on dataset 1 were 97.44% and 98.50% and on dataset 2 were 96.13% and 87.42%. This work scanned 11 patients with 24-h paroxysmal AF (PAF). The best result is that the detection accuracy is 99.33%, and the false detection is 0.01%, and the worst outcome is that the detection accuracy is 88.75%, and the false detection is 14.06%. The results proved that the proposed method could provide reliable scanning for PAF events.

The effects of Landsat image acquisition date on winter wheat classification in the North China Plain

Abstract: Accurate winter wheat distribution provides critical information for yield estimation and agricultural resources monitoring, which is associated with food security and agricultural ecosystem sustainability. The increasingly high spatiotemporal resolutions of globally available satellite images (e.g., Landsat) create new possibilities for generating accurate datasets. Although winter wheat classification using Landsat time-series data has been increasingly acknowledged in the literature, the contributing effects of a single acquisition date on classification remain poorly explored. Accordingly, this study aimed to evaluate the effects of a single acquisition date (nine acquisition dates were obtained throughout the growing season) on winter wheat classification using multi-temporal Landsat 8 imagery. We adopted Dynamic time warping (DTW) to discriminate winter wheat and determined the most relevant acquisition dates for classification accuracy with different numbers of images. The most significant advantage of DTW from other similarity measurement methods is that DTW can identify the optimal alignment of two time-series datasets with an unequal number of images. Then, we calculated the importance of each acquisition date using the random forest algorithm and quantified their contributions to the classification using variance analysis. The conditional combination of our best accuracy was 84.51% for overall accuracy (OA) and 84.62% for F-score when all nine images were used for the classification. While the lowest accuracy was 46.2% for OA and 51.3% for F-score through the classification of two image combinations from the wintering and heading images. It was shown that the initial and late growth images were the most discriminating dates for time series classification of winter wheat in ∼ 91% of the total combinations explored (i.e., 456 combinations). The main effects of the late growth images and their interactions significantly affected model accuracy, while the interactions between the initial growth images also played an important role, particularly for greater numbers of acquisition dates. The relative contributions explained by the main effects of late growth images gradually decreased with the increasing image acquisition dates, while those explained by the internal interactions of both initial and late growth images significantly increased. The wide distribution of winter wheat in many parts of the world and its strong consequences for global food security suggest that better monitoring of wheat is necessary, and the results indicated that appropriate selection of image dates was significant for accomplishing this task.

Designs of human–robot interaction using depth sensor-based hand gesture communication for smart material-handling robot operations

Abstract: With rapid developments in biometric recognition, a great deal of attention is being paid to robots which interact smartly with humans and communicate certain types of biometrical information. Such human–machine interaction (HMI), also well-known as human–robot interaction (HRI), will, in the future, prove an important development when it comes to automotive manufacturing applications. Currently, hand gesture recognition-based HRI designs are being practically used in various areas of automotive manufacturing, assembly lines, supply chains, and collaborative inspection. However, very few studies are focused on material-handling robot interactions combined with hand gesture communication of the operator. The current work develops a depth sensor-based dynamic hand gesture recognition scheme for continuous-time operations with material-handling robots. The proposed approach properly employs the Kinect depth sensor to extract features of Hu moment invariants from depth data, through which feature-based template match hand gesture recognition is developed. In order to construct continuous-time robot operations using dynamic hand gestures with concatenations of a series of hand gesture actions, the wake-up reminder scheme using fingertip detection calculations is established to accurately denote the starting, ending, and switching timestamps of a series of gesture actions. To be able to perform typical template match on continuous-time dynamic hand gesture recognition with the ability of real-time recognition, representative frame estimates using centroid, middle, and middle-region voting approaches are also presented and combined with template match computations. Experimental results show that, in certain continuous-time periods, the proposed complete hand gesture recognition framework can provide a smooth operation for the material-handling robot when compared with robots controlled using only extractions of full frames; presented representative frames estimated by middle-region voting will maintain fast computations and still reach the competitive recognition accuracy of 90.8%. The method proposed in this study can facilitate the smart assembly line and human–robot collaborations in automotive manufacturing.

A modified Kohonen map algorithm for clustering time series data

Abstract: Time Series clustering is a domain with several applications spanning various fields. The concept of vector quantization, popularly used in signal processing to approximate a large number of signals, can be used to cluster signals and thereby time series data. Though a popular clustering algorithm such as K-Means is capable of performing vector quantization, the averaging technique to compute centroids in the algorithm is not well suited to handle time series data. The ability of Self Organizing Map algorithm, has, therefore, been explored in this work to perform clustering of time series data by adopting several modifications in the original steps of the algorithm. By initializing the prototype vectors using a farthest neighbors’ approach instead of random initialization and using the dynamic time warping distance measure to calculate similarity between signals, a novel procedure has been proposed to apply the Self Organizing Map algorithm to cluster time series data. The proposed algorithm is first tested on 119 data sets and its performance is compared to that of Agglomerative Clustering and k medoids clustering using 3 validation measures. Next, their scalability is compared by looking at their time of computation on the data sets. Performance of the proposed algorithm in terms of the fluctuations involved due to initialization and the parameters of the algorithm are studied next using 3 more validation measures. The results showcase that the modified Self Organizing Map is not only a better algorithm than Agglomerative Clustering in terms of clustering performance, but also more scalable in terms of taking less time to compute clusters as it performs them in lesser time that k medoids while having similar cluster quality.

Dynamic Time Warping Under Product Quantization, With Applications to Time-Series Data Similarity Search

Abstract: The similarity search on sensor data generated by a myriad of sensing devices is a frequently encountered problem in the era of the Internet of Things (IoT). This sensor data generally appear in the form of time series, a temporally ordered sequence of real numbers obtained regularly in time. It has been widely accepted that the dynamic time warping (DTW) currently is the most prevalent similarity measure in the time-series mining community, mainly due to its flexibility and broad applicability. However, calculating DTW between two time series has quadratic time complexity, leading to unsatisfactory efficiency when performing the similarity search over the large time-series data set. The main contribution of this article is to propose a method called product quantization (PQ)-based DTW (PQDTW) for fast time-series approximate similarity search under DTW. The PQ, a well-known approximate nearest neighbor search approach, is used in PQDTW. Nevertheless, the conventional PQ is developed with the Euclidean distance and is not designed for DTW. To solve this problem, the DTW barycenter averaging (DBA) technique is utilized to adapt the PQ for DTW before using it. We employ PQDTW along with the filter-and-refine framework to efficiently and accurately perform the time-series similarity search. Our method can reasonably reduce many DTW computations in the filtering phase; thus, the query process is accelerated. We compare PQDTW with related popular algorithms using public time-series data sets. Experimental results verify that the proposal achieves the best tradeoff between query efficiency and retrieval accuracy compared to the competitors.

Training Robust Deep Models for Time-Series Domain: Novel Algorithms and Theoretical Analysis

Abstract: Despite the success of deep neural networks (DNNs) for real-world applications over time-series data such as mobile health, little is known about how to train robust DNNs for time-series domain due to its unique characteristics compared to images and text data. In this paper, we fill this gap by proposing a novel algorithmic framework referred as RObust Training for Time-Series (RO-TS) to create robust deep models for time-series classification tasks. Specifically, we formulate a min-max optimization problem over the model parameters by explicitly reasoning about the robustness criteria in terms of additive perturbations to time-series inputs measured by the global alignment kernel (GAK) based distance. We also show the generality and advantages of our formulation using the summation structure over time-series alignments by relating both GAK and dynamic time warping (DTW). This problem is an instance of a family of compositional min-max optimization problems, which are challenging and open with unclear theoretical guarantee. We propose a principled stochastic compositional alternating gradient descent ascent (SCAGDA) algorithm for this family of optimization problems. Unlike traditional methods for time-series that require approximate computation of distance measures, SCAGDA approximates the GAK based distance on-the-fly using a moving average approach. We theoretically analyze the convergence rate of SCAGDA and provide strong theoretical support for the estimation of GAK based distance. Our experiments on real-world benchmarks demonstrate that RO-TS creates more robust deep models when compared to adversarial training using prior methods that rely on data augmentation or new definitions of loss functions. We also demonstrate the importance of GAK for time-series data over the Euclidean distance.

Clustering sequential navigation patterns in multiple-source reading tasks with dynamic time warping method

Abstract: Background Data-driven investigations of how students transit pages in digital reading tasks and how much time they spend on each transition allow mapping sequences of navigation behaviours into students' navigation reading strategies. Objectives The purpose of this study is threefold: (1) to identify students' navigation patterns in multiple-source reading tasks using a sequence clustering approach; (2) to examine how students' navigation patterns are associated with their reading performance and socio-demographic characteristics; (3) to showcase how the navigation sequences could be clustered on the similarity measure by dynamic time warping (DTW) methods. Methods This study draws on process data from a sample of 16,957 students from 69 countries participating in the PISA 2018 study to identify how students navigate through a multiple-source reading item. Students' navigation sequences were characterized by two indicators: the page sequence that tracks the page transition path and the time sequence that records the time duration on each visited page. K-medoid partitioning clustering analyses were conducted on pairwise distance similarity measures computed by the DTW method. Results and conclusions Students' navigation patterns were found moderately associated with their reading proficiency levels. Students who visited all the pages and spent more time reading without rush transitions obtained the highest reading scores. Girls were more likely to achieve higher scores than boys when longer navigation sequences were used with shorter reading time on transited pages. Students who navigated only limited pages and spent shorter reading time were averagely at the lowest rank of socio-economic status. Implications This study provides evidence for the exploration of students' navigation patterns and the examination of associations between navigation patterns and reading scores with the use of process data.