Please use this identifier to cite or link to this item: http://hdl.handle.net/2381/10048
Title: Algorithms for Motion Planning and Target Capturing
Authors: Kothari, Mangal
Supervisors: Postlethwaite, Ian
Award date: 1-Oct-2011
Presented at: University of Leicester
Abstract: This thesis addresses the development and implementation of algorithms for unmanned air vehicles (UAVs). With advances in technology, it is relatively easy to manufacture and operate UAVs that are particularly useful for dull, dirty and dangerous operations. The success of such autonomous missions depends heavily on the planning algorithms used. A key consideration in the thesis is the path planning problem for single and multiple UAV systems in obstacle rich environments in the presence of uncertainty. Recently, rapidly-exploring random trees (RRTs) have been applied to find feasible trajectories quickly in complex motion planning problems. We use RRTs to construct trees of kinematically feasible trajectories made of waypaths, and feasibility is evaluated by checking for collisions with the predicted trajectories. When there are uncertainties acting on the system, we can identify probabilistic feasible paths by growing trees of state distributions and ensuring that the probability of constraint violation is below a pre-defined value. In addition to this, a guidance law is designed combining a pursuit law with a line-of-sight law, to track the path generated by the path planner with minimum deviation. In the penultimate chapter, an application is presented where a multi-UAV system captures a more capable target by forming a target centred formation around it. The approach combines a consensus algorithm with a controller to develop a robust distributed control law for formation control. Under certain conditions, it is shown that a set of UAVs can form a target centred formation even when target information is not known. The effectiveness of this algorithm is demonstrated using numerical results. In the appendix, a decentralised scheme is described for target tracking using consensus theory in conjunction with a data fusion algorithm which guarantees perfect fault detection and isolation. This scheme was developed by the Leicester team. As part of this thesis the theoretical algorithm was tested experimentally on a set of real robots.
Links: http://hdl.handle.net/2381/10048
Type: Thesis
Level: Doctoral
Qualification: PhD
Rights: Copyright © the author, 2011.
Appears in Collections:Theses, Dept. of Engineering
Leicester Theses

Files in This Item:
File Description SizeFormat 
main.pdf3.46 MBAdobe PDFView/Open


Items in LRA are protected by copyright, with all rights reserved, unless otherwise indicated.