| || || Drone aircraft -- Control systems|
| || || A swarm model for planar formations of multiple autonomous unmanned aerial vehicles (UAVS) |
Author: Sharan, Ashna
Institution: University of the South Pacific.
Subject: Drone aircraft -- Control systems, Drone aircraft -- Automatic control, Drone aircraft -- Mathematical models, Vehicles, remotely piloted
Call No.: Pac UG 1242 .D7 S53 2013
Copyright:40-60% of this thesis may be copied without the authors written permission
Abstract: This thesis addresses the common ndpath problem for a multi-agent robotic sys- tem, the solution of which has been inspired from observations of multi-agent dynamical systems in nature including certain species of insects, birds and sh. For this reason the introductory chapter is devoted to understanding relevant con- cepts of swarming in the context of biological systems in nature which then serve as the building blocks of the emerging eld of swarm robotics. The primary ob- jective is to control multiple autonomous quadrotor Unmanned Aerial Vehicles (UAVs) simultaneously. In order to meet this objective, a general swarm model for a system of simple point-like rigid bodies is designed. The general swarm model is then adapted so that it is applicable to a system of quadrotor UAVs. The Direct Method of Lyapunov is used in designing control algorithms to ensure the stabil- ity of the entire system. Lyapunov functions are constructed and new velocity controllers are derived as feedback controllers via consideration of their gradients. Two cases are considered, in the rst case the conguration space is free of static obstacles and in the second it is cluttered with static obstacles. Once the velocity controllers have been dened, the swarm model is simulated for verication of its functionality. Basic patterns of formation which are similar to emergent behaviors distinctive of swarming dynamical systems in nature are demonstrated. Some ad- vanced maneuvering such as split and rejoin and tunnelling are also observed in the stationary obstacle collision avoidance scenario. Dierent emergent pat- terns are obtained with variation in the control parameters. Time evolution of the translational and angular velocities derived are also evaluated to understand the emergent patterns exhibited. v