Approximate Similarity Measurements on Multi-Attributes Trajectories Data

Abstract

With the development of global positioning technology, sensor networks, and smart mobile terminal, a large number of trajectory data are accumulated. Trajectory data contains a wealth of information, including spatiality, time series, and other external descriptive attributes (i.e., features, travelling mode, and so on). Trajectory analysis and mining show the great value. The research of trajectory similarity measurement is the basis of trajectory data management and mining, which plays an important role in trajectory computing. Most trajectory similarity work only focuses on the spatial-temporal features. The addition of multi-attributes to the trajectories changes the trajectory similarity. However, there are few researches focusing on multi-attributes trajectory similarity. In this paper, we propose two novel trajectory similarity measurements, i.e. maximum-minimum trajectory distance and sum of minimum trajectory distance and analyze the correlation among the spatial-temporal similarity and textual similarity. Finally, the measurement validity is verified and visualized through clustering, by both a simulation dataset and a real dataset. INDEX TERMS Trajectory similarity measurement, clustering, big trajectory data, trajectory computing.

Detecting Travel Modes Using Rule-Based Classification System and Gaussian Process Classifier

“…. Thereafter, we take trials from p(f * |X, x * , f), softmax them according to equation (4), and average them to derive the travel mode for each test segment [46]. The number of trials is determined for each test segment by taking trials repeatedly until the observed probability fluctuation of travel modes has ''stabilized''.…”

“…In the past few decades, dedicated Global Positioning System (GPS) devices and smartphones have been increasingly applied to gather location-based data in GPS-based travel surveys, which are widely considered promising alternatives to conventional travel surveys. This type of travel survey method is advantageous because of its minimal burden on respondents [1], high reliability of GPS data, and valuable information obtained from such data [2]- [4]. Nevertheless, some key travel characteristics, including trip rates, trip purposes, and travel modes, may not be directly derived from GPS data.…”

Outlier Detection of Crowdsourcing Trajectory Data Based on Spatial and Temporal Characterization

“…As an emerging type of spatio-temporal big data based on positioning technology and navigation devices, vehicle-based crowdsourcing data has become a valuable trajectory data resource. The analysis and mining of spatio-temporal trajectory data is fundamental content in the field of urban management and human activity, which consists of trajectory clustering [1], trajectory correlation analysis [2], trajectory prediction [3], target motion pattern recognition [4], and outlier detection [5], etc. However, crowdsourcing trajectory data is collected by non-professionals and multiple measurement terminals, resulting in some errors in the data collection.…”