Browsing by Author "Gillis, Colin"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
Item Dynamic model development and simulation of an autonomous active AUV docking device using a mechanically actuated mechanism to recover AUVs to a submerged slowly moving submarine in waves(University of New Brunswick, 2014) Gillis, Colin; Carretero, Juan; Dubay, Rickey; Jeans, Tiger; Watt, GeorgeAutonomous Underwater Vehicles (AUVs) are presenting an ever expanding range of applications that enhance human capabilities and mitigate human risk. Development of a successful subsurface autonomous launch and recovery system would expand the functional use of AUVs in many fields. Defence Research and Development Canada (DRDC) is leading a collaborative project with the University of New Brunswick (UNB) to develop such a system that would recover AUVs to a slowly moving submerged submarine. This thesis provides an overview of the design, dynamic modelling, and preliminary control in simulation of an electro-mechanically actuated AUV dock concept, which operates without using hydrodynamic fluid power to provide motive force. The device is partially faired and has a R⊥R⊥P serial manipulator architecture. In short, the device is referred to as the mechanically actuated RRP serial manipulator. High speed actuation of the device is required to compensate for relative trajectory errors between the submarine and AUV during significant sea states in littoral waters. Hydrodynamic forces present in water cannot be ignored and will be modelled using the Morison Equation. Unimodal Linear Wave Theory is used to simulate AUV kinematics, establishing end effector design trajectories, as well as providing wave kinematics for hydrodynamic modelling. Alterations to the recursive Newton-Euler derivation of manipulator dynamics are explained, and results of simulations are presented. Model Predictive Control (MPC) of the mechanically actuated RRP serial manipulator is simulated using a Dynamic Matrix Control (DMC) architecture. The dynamic models are verified analytically and provide accurate evaluation without lose of generality. Dynamic modelling shows the actuator loads for the proposed device are significant. Drag is the largest contributer and indicates the device must be streamlined. The link diameter used for simulation is overly conservative and should be optimized to reduce its size, this will decrease the required actuator loads. The control simulation shows the DMC controller is a robust design for tracking, however it needs to be combined in a cascading architecture to control both position and velocity for precise control. Overall, the mechanically actuated RRP serial manipulator is a viable design but requires further modelling and development. The device becomes more promising as it is streamlined and reduced in overall length.