Using Self Organizing Networks for Moving Object Trajectory Prediction V ijay S
Rajpurohit* and Manohara Pai M. M.** Abstract: Trajectory prediction for moving objects in a Robotic navigational environment is a challenging problem where, a Robot either follows a given trajectory or keeps track of a moving object. This paper proposes a method for predicting the trajectory and the final destination of a moving object in a Fuzzy navigational environment. Trajectory prediction is done using Fuzzy based Self Organizing Map Networks. The proposed technique learns the typical motions of moving objects and predicts the trajectory of a moving object by observing its partial motion. The environmental uncertainties are effectively addressed by the Fuzzy navigational system. Self Organizing Map efficiently clusters the learned motion trajectories and classifies the observed partial motion trajectory to the nearest cluster class in a short duration. Results are tested for Real Life data sets. Performance of the proposed predictor is better in response time and comparable in relative error to the existing algorithms. Keywords: Moving Object Trajectory Prediction, Fuzzy Navigational Environment, Fuzzy Self Organizing Map, Trajectory clustering, Partial Trajectory.
(
BibTex @ARTICLE{P1120906633, AUTHOR = {V ijay S Rajpurohit and Manohara Pai M. M.}, TITLE = {Using Self Organizing Networks for Moving Object Trajectory Prediction},
JOURNAL = {ICGST International Journal on Artificial Intelligence and Machine Learning,
AIML},
YEAR = {2009}, VOLUME = {9}, ISSUE ={I}, PAGES = {27--34} } ( |
|||
|
P1120906633
