Electrical and Electronic Engineering - Theses
Permanent URI for this collection
Search Results
1 - 2 of 2
-
ItemTarget tracking using sequential Monte Carlo methods( 2004)The objective of target tracking is to estimate the current state of one or multiple targets using a series of sensor measurements, usually received at discrete instants of time. In many cases, the target dynamics and measurement relation are modelled as a hidden Markov model. The Bayesian recursion can then be used to estimate the target state sequentially in time. An example of an algorithm employing the Bayesian recursion is the celebrated Kalman filter. However, the Kalman filter assumes that the target dynamics and measurement relation are linear Gaussian, and does not work well in non-linear, non-Gaussian and multitarget cases. Sequential Monte Carlo (SMC) methods, or particle filtering techniques are simulation based methods that can be used for estimation in non-linear and non-Gaussian environments. In this approach, the distributions of interest are approximated using a large number of random samples generated via a. sequence of sequential importance sampling (SIS) and resampling steps. It has gained popularity over recent years due to increase in computational power. This thesis presents a review of estimation theory, some commonly used approaches for target tracking and also an introduction to SMC methods. A novel single target tracking algorithm based on SMC methods is proposed. In this target tracking scenario, the target dynamics are linear Gaussian. However, the measurements are highly non-linear unthresholded pixels. The proposed algorithm uses Rao-Blackwellisation, the optimal importance function and the Kalman recursion to arrive at estimates of the target. We show in this thesis that this algorithm performs better than the traditional bootstrap filter, especially at low SNRs. A multitarget tracking algorithm is also reviewed. This algorithm is a departure from other multitarget tracking algorithms because the multitarget tracking system is modelled as a jump Markov system. This algorithm estimates the states of the multiple targets based on the auxiliary particle filter. Simulation results are included in this thesis to illustrate the outcomes of these two algorithms.
-
ItemBearings only tracking( 1994)This thesis addresses the problem of tracking a single target when only bearing information is available from the sensor. The aim is to determine target information, such as heading, speed and position. A number of existing techniques are described. A new HMM based algorithm is proposed as a solution to the bearings only tracking problem. Track initiation, maintenance and termination are automatically provided in the HMM framework. The key issue of the new algorithm is the representation of a continuous state process, such as target position and velocity, using a discrete time finite state process. Rules and assumptions for constructing a finite state HMM are presented. Bearing measurement imperfections are modeled by additive white noise with state dependent standard deviation reflecting the target-observer geometry. Observer maneuvering strategies, that assist in the quality of track, are not addressed. A large number of simulations have been performed and the results show the excellent tracking capability of the new algorithm.