Techniques for Information Extraction From Compressed GPS Traces : Final Report
-
2015-12-31
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final report, 1/1/2014 - 12/31/2015
-
Corporate Publisher:
-
Abstract:Developing techniques for extracting information requires a good understanding of methods used to compress the traces. Many techniques for compressing trace data consisting of position (i.e., latitude/longitude) and time values have been developed. Since current vehicles are equipped with many on-board instruments, traces generated by such vehicles contain many attributes in addition to position and time. The problem of compressing such multi--atribute traces is currently being studied by a number of researchers. We consider the Multiple Attribute Trajectory Compression program with defined error bounds on attributes. Our focus is on solving this problem using attribute partition methods. Such methods partition attributes into groups with the aim of reducing the total storage cost after compression methods are applied to each group individually. We present a comprehensive overview of various trajectory compression algorithms, concentrating on the most recent works since SQUISH-E [20]. New accuracy metrics for measuring the difference between a trajectory and its compressed representation are also explained. Lastly, we present some preliminary experimental results on two real trajectory data sets with multiple attributes using two known compression algorithms, namely PROXIMUS and CompreX.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
File Type:
Supporting Files
-
No Additional Files
More +