PhD Studentship: Particle Methods for Estimation and Control
Funded by EPSRC and British Aerospace Systems
Stipend £14600 pa plus fees
More information on: http://www.dcs.lancs.ac.uk/site/1/5/Vacancies
There is a substantial interest in developing autonomous systems, as part of
remote surveillance of the environment. Such systems need to possess
computational intelligence, be able to sense the environment, collect
information and perform path planning. Applications are invited for a fully
funded PhD position on algorithmic and theoretical aspects of target
tracking systems. This PhD project will focus on the development of novel
Monte Carlo methods (particle filtering) for estimation and control
purposes. Between the problems that will be investigated are motion
estimation of a single and multiple objects from a large amount of sensor
data. The proposed research will investigate novel particle filtering
methods to solve non-linear, non-Gaussian, estimation/control problems where
current techniques are inadequate. Immediate applications relate to path
planning control of Unmanned Aerial vehicles (UAVs), but more generally the
work is relevant to control of vehicles and robots in uncertain situations.
UAVs are remotely controlled or self-piloted aircraft that carry different
sensors and are equipped with communication facilities. The work will go
beyond the usual realm of Particle Filtering techniques. There are three
strands to the work: estimation in closed loop environments; the use of
particles to solve optimal stochastic control problems related with path
planning and the combination of particle filters with derivative-free methods.
The studentship will be supervised by Dr Mila Mihaylova (Lancaster
University) and it is anticipated that the student will be based at
Lancaster University, offering a highly collegiate and stimulating
environment for research career development as well as undertaking a
programme of original research.
|