My Background

I am a graduate student in Transportation Engineering. I know basic Statistics but have never done any time-series/ spatio-temporal data analysis.

Data Format

I conducted an experiment on a driving simulator which produced vehicle trajectory data on a test road network. These data (for each driver) contain time frames, vehicle position (x, y, z Cartesian coordinates), distance to lead vehicle, etc. I use R language for analysis.


I want to find similarities in the vehicle trajectories. Let's say, I have 10 trajectories, each representing the journey completed by an individual driver. I want to see how many trajectories are "similar" in terms of time and space. The road network consists of various ramps, so I expect that driver behavior is likely to be different close to ramps than on the highway. Additionally, speeds/accelerations of different drivers might be (dis)similar at different locations.


I have extensively searched similarity/clustering/segmentation algorithms for time-series and spatio-temporal data. In time-series examples, I find that mostly seasonality and decomposition is discussed. And the data contain date/time stamp. However, my data only contain the time frames.
In spatio-temporal data examples, I find latitude, longitude variables. In my case, I only know about proximity to ramps, Cartesian coordinates, distance to lead vehicle, etc. Therefore, my questions are: Can I treat my data as spatio-temporal data? What properties of time-series are applicable to trajectory data sets? What similarity measures can be used on these multi-variate data? Please help with links/examples/implementations in R. Thanks in advance.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.