Find the latest in DOE-sponsored Scientific and Technical Software

SOFTWARE ABSTRACT

Request Information Back

PACKAGE ID	002703IBMPC00 TRACC
KWIC Title	TRACC: Algorithm for Predicting and Tracking Barges on Inland Waterways
AUTHORS	Duminda I.B. Randeniya
LIMITATION/AUDIENCE CODE	COPY/LIM
COMPLETION DATE	04/23/2010
PUBLICATION DATE	04/23/2010
DESCRIPTION	Algorithm developed in this work is used to predict the location and estimate the traveling speed of a barge moving in inland waterway network. Measurements obtained from GPS or other systems are corrupted with measurement noise and reported at large, irregular time intervals. Thus, creating uncertainty about the current location of the barge and minimizing the effectiveness of emergency response activities in case of an accident or act of terrorism. Developing a prediction algorithm become a non-trivial problem due to estimation of speed becomes challenging, attributed to the complex interactions between multiple systems associated in the process. This software, uses systems approach in modeling the motion dynamics of the barge and estimates the location and speed of the barge at next, user defined, time interval. In this work, first, to estimate the speed a non-linear, stochastic modeling technique was developed that take local variations and interactions existing in the system. Output speed is then used as an observation in a statistically optimal filtering technique, Kalman filter, formulated in state-space to minimize numerous errors observed in the system. The combined system synergistically fuses the local information available with measurements obtained to predict the location and speed of traveling of the barge accurately.
PACKAGE CONTENTS	OAK RIDGE NATIONAL LABORATORY WILL BE HANDLING ALL SOFTWARE DISTRIBUTION.
SOURCE CODE INCLUDED?	Y
MEDIA QUANTITY	1
COMPUTER	IBM PC
OPERATING SYSTEMS	Windows XP, Vista, 7
PROGRAMMING LANGUAGES	Matlab 7.8.0.347 (R2009a)
SOFTWARE LIMITATIONS	The software is sensitive to the measurement time interval; the rate of convergence depends on the time interval. Also, prediction model is sensitive to the outliers. Although, there is a function to eliminate the outliers from corrupting the predictions, there is no proper technique to handle outliers. At this stage, the software can only be used on post-processing, however, to make the software capable of handling real-time data is not difficult and can be performed easily.
RELATED SOFTWARE	User Guide
ABSTRACT STATUS	Released As-Is 11/29/2011
SPONSOR	USDOE
RESEARCH ORG	Oak Ridge National Laboratory
PACKAGE TYPE	AS-IS

Request Information Back