More about HKUST
Compression Methods for Spatio-temporal Data Of Moving Objects
PhD Qualifying Examination
Title: "Compression Methods for Spatio-temporal Data Of Moving Objects"
by
Mr. Yudian JI
Abstract:
Advanced location acquisition technologies have led to a wide range of
location based applications and services supported by location acquisition
systems like GPS and sensor networks. The popularization of location based
applications, in turn, has caused the explosion of spatio-temporal data of
moving objects, consequently making the spatio-temporal data compression
techniques desired. The spatio-temporal data are data with location
information and time series viewed from different solutions about data
compression. Spatio-temporal data of moving objects are basically stored as
strings, which means traditional string compression techniques can be applied
for data compression. Traditional string compression methods are good in some
cases because they are simple and fast. However, these methods ignored the
spatio-temporal characteristics of data, which may lead to a better
performance of compression. More importantly, the data lose all the utility
after compression, thus being very inconvenient for those data that need to
be queried often. The limitations here make traditional compression
techniques based on information theory not sufficient for the cases where
optimized compression ratio is important or frequent queries are needed.
Taking the spatio-temporal characteristics into consideration, there are also
two ways of looking at the problem. The data could be seen as trajectories
with geometry instincts, or prefix tree-like different ways of different
methodologies. For example, the line simplification techniques mainly deal
with the curvilinear features of the data, while map-matched compression
techniques focus on graph theory based problems like shortest path. Different
aspects will have their pros and cons respectively. In this survey, we bring
a deep insight into the field of spatio-temporal data compression by giving a
classification of current spatio-temporal data compression techniques and
reviewing the typical methods of each kind. We study their basic ideas,
methodologies and evaluations. Comments on pros and cons will be given
respectively.
Date: Friday, 12 June 2015
Time: 2:30pm - 4:30pm
Venue: Room 3584
Lifts 27/28
Committee Members: Prof. Lionel Ni (Supervisor)
Dr. Qiong Luo (Chairperson)
Dr. Lin Gu
Dr. Ke Yi
**** ALL are Welcome ****