Article
Engineering, Civil
Yulang Huang, Dianhai Wang, Wang Xu, Zhengyi Cai, Fengjie Fu
Summary: This paper introduces an innovative map matching method for mobile phone signaling data (MSD), leveraging an incremental Hidden Markov Model (HMM) algorithm. The proposed method addresses several challenges associated with MSD, and achieves better accuracy in extracting people's travel trajectories and providing optimal paths on digital maps, compared to existing approaches.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Computer Science, Information Systems
Junyi Cheng, Xianfeng Zhang, Peng Luo, Jie Huang, Jianfeng Huang
Summary: In this paper, an unsupervised method called UPAPP is proposed for the semantic place annotation of individual trajectories using spatiotemporal information. The method decomposes the spatiotemporal probability into spatial probability, duration probability, and visiting time probability, and integrates spatial information from two geospatial data sources. The method achieves accurate place annotation without any annotated data, as demonstrated by experimental results.
INFORMATION SCIENCES
(2022)
Article
Computer Science, Information Systems
Ge Cui, Wentao Bian, Xin Wang
Summary: Map matching is a crucial preprocessing step for many GPS trajectory-based applications. The conventional map matching methods based on hidden Markov model can suffer from decreased effectiveness and efficiency in dense road networks. This study proposes a segment-based hidden Markov model method to improve the performance of map matching by segmenting GPS trajectories and searching for candidate road segment sequences.
Article
Computer Science, Information Systems
Ruijie Tian, Jiajun Li, Weishi Zhang, Fei Wang
Summary: In this paper, we propose DFST, an efficient framework for semantic trajectory similarity join in distributed systems, which achieves a 13.6% improvement of join performance compared to existing methods. DFST utilizes ITS index and summary index to prune dissimilar trajectory pairs and supports most existing similarity functions to quantify spatial similarity. Experimental results on real world datasets demonstrate the effectiveness of DFST.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Automation & Control Systems
Renhai Chen, Shimin Yuan, Chenlin Ma, Huihui Zhao, Zhiyong Feng
Summary: This article discusses the advantages and disadvantages of GPS and cellular-based positioning. It highlights the challenges of cellular-based positioning and proposes a novel algorithm called THMM to improve its accuracy. The algorithm is optimized based on the characteristics of cellular-based data and the experimental results demonstrate its effectiveness.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2022)
Article
Multidisciplinary Sciences
Ha Yoon Song, Jae Ho Lee
Summary: With the advancement of geopositioning systems and mobile devices, much research is being conducted on geopositioning data. Map matching, a core preprocessing technique for trajectory data, is gaining attention, particularly the use of Hidden Markov Model (HMM) for map matching. However, the HMM model simplifies the dependency of time series data excessively, leading to incorrect matching results. In this research, a new algorithm called trendHMM map matching is proposed, which improves upon the assumptions of HMM by considering a wider range of dependencies and incorporating neighboring data.
Article
Computer Science, Artificial Intelligence
Alaa Zreik, Zoubida Kedad
Summary: This paper introduces a method to predict the physical state of documents based on their conservation-restoration histories. The paper presents a representation of the conservation-restoration trajectory, a matching process, and an analysis process, and proposes a prediction model based on the analysis process.
DATA & KNOWLEDGE ENGINEERING
(2022)
Article
Physics, Multidisciplinary
Jing Zhao, Yi Zhang, Shiliang Sun, Haiwei Dai
Summary: The Hidden Markov model is crucial for trajectory recognition, and the sampled BP-HMM model, while effective, is inconvenient for classification and slow to converge. To improve trajectory recognition performance, a novel variational BP-HMM model has been proposed, which can share information among different classes.
Article
Computer Science, Information Systems
Xintai He, Qing Li, Runze Wang, Kun Chen
Summary: This study proposes a spatio-temporal feature trajectory clustering algorithm based on deep learning, which improves the clustering performance by combining image matching technology with trajectory temporal features.
Article
Computer Science, Artificial Intelligence
Mohsen Zand, Ali Etemad, Michael Greenspan
Summary: Conditional Normalizing Flows (CNFs) are flexible generative models capable of representing complicated distributions with high dimensionality and large interdimensional correlations, making them appealing for structured output learning. MotionFlow, a new normalizing flow approach, combines deterministic and stochastic representations with CNFs to create a probabilistic neural generative approach that can model the variability seen in high-dimensional structured spatio-temporal data.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Automation & Control Systems
Konstantinos E. Nikolakakis, Dionysios S. Kalogerias, Anand D. Sarwate
Summary: This paper provides high-probability sample complexity guarantees for exact structure recovery and accurate predictive learning in tree-shaped graphical models. By quantifying the impact of noise in the hidden model on tasks like structure recovery and marginal distribution estimation, the study presents upper and lower bounds on sample complexity, generalizing prior work and recovering noiseless cases.
JOURNAL OF MACHINE LEARNING RESEARCH
(2021)
Article
Computer Science, Artificial Intelligence
Arpan Man Sainju, Wenchong He, Zhe Jiang
Summary: This paper introduces a novel spatial structured model that captures the dependency structure between samples based on their locations in space. The model outperforms existing methods in classification performance and can be used as a post-processor to refine the results of other models.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2022)
Article
Green & Sustainable Science & Technology
Zhengang Xiong, Bin Li, Dongmei Liu
Summary: This study proposes a new map matching method that combines the widely used Hidden Markov Model with decision makers' path choice preferences, aiming to improve matching accuracy, especially for higher frequency locating trajectory. The algorithm is tested in Beijing, China, showing that it can enhance matching accuracy while considering travelers' route choice preferences.
Article
Computer Science, Information Systems
Xiaoxin Du, Hui Zhu, Yandong Zheng, Rongxing Lu, Fengwei Wang, Hui Li
Summary: With the widespread use of IoT, LBS providers have collected large volumes of individuals' trajectories, which can be valuable for certain applications. However, directly publishing these trajectories may violate privacy and lead to data loss. To address this issue, this article proposes a semantic-preserving scheme for synthesizing and publishing trajectories under differential privacy.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Information Systems
Thouraya Sakouhi, Jalel Akaichi
Summary: This study explores enriching individuals' mobility data with contextual information from geographic data and social media, presenting a novel approach that integrates these three data sources for trajectory semantic annotation. Experimental results show that the proposed approach improves the precision of annotation words while maintaining similar recall rates, enhancing the quality of trajectory semantics by combining both data sources.
PERVASIVE AND MOBILE COMPUTING
(2021)