PhD Qualifying Examination


A Survey of Graphical Models for Tracking and Plan Recognition in
Pervasive Computing

Mr. Junfeng Pan


Abstract:

With the recent advance in pervasive computing technology and mobile
devices, location-sensing and context-aware systems have attracted
intense attention. Location-based systems have a variety of
applications from object tracking to people monitoring. This survey
focuses on two popular topics in Pervasive Computing Environment:
tracking users' locations as well as recognizing their activities.
Yet these two tasks are correlated in nature. A user's trajectory
may contain rich information which reflects his activities. However,
his activity routines could act as a feedback to refine his
locations. This survey mainly reviews machine learning approaches,
especially graphical models such as kalman filter, particle filter,
hidden markov model, dynamic bayesian network, markov random field
and dynamic markov network. These models well capture the sequential
and structural characteristics of tracking and plan recognition
problems.


Date:     		Monday, 9 January 2006

Time:                   3:00p.m.-5:00p.m.

Venue:                  Room 4480
			lifts 25-26

Committee Members:      Dr. Qiang Yang (Supervisor)
			Prof. Lionel Ni (Chairperson)
			Dr. James Kwok
			Dr. Yunhao Liu


**** ALL are Welcome ****