Optimal Control Model for an Autonomous Underwater Vehicle

Abstract: The main goal with this thesis was to develop an optimal control model for anautonomous underwater vehicle (AUV) and evaluate whether reference trackingusing Model Predictive Control (MPC) based on a linear dynamics modeldescribing all six degrees of freedom (DOF) is a suitable method for waypointnavigation. MPC is an advanced receding horizon optimal control method capableof including constraints in the optimization. The drawback with this methodis that it is computationally heavy and since the dynamic AUV behavior is bothcomplex and nonlinear there are two open research questions: Does MPC improvereference tracking and is it computationally feasible?Optimal control requires accurate dynamics models of the targeted system tofunction and ensure robustness. At the start of this thesis the targeted AUV wasan unmodelled vehicle and therefore modeling, model analysis and linearizationtechniques constitutes a big part of this thesis and creating a nonlinear dynamicsmodel was a big task in this project.To ensure that reference tracking using optimal control strategies was feasible,state-feedback solutions together with standard techniques for stability analysis,controllability and observability were investigated. A linear quadratic regulator(LQR) was designed using a linear time-invariant (LTI) dynamics modelaugmented with integral action and error dynamics to create a performancereference for the MPC implementation. The final step during this thesis wasto implement MPC reference tracking using integral action on both states andinputs and the results from this controller was then compared to the resultsfrom the LQR controller in terms of performance.A model generator capable of creating 6 DOF nonlinear dynamics models withvarying complexity has been designed. This generator was used to derive twodi↵erent linear models describing the dynamic AUV behavior. These linearmodels were analyzed and the necessary model parameters were estimated usingsimulation and then compared to physical data and live testing. Resultsfrom the LQR controller using an augmented error dynamics model is promisingin terms of reference tracking. Due to a time varying reference and anunder-actuated vehicle the resulting performance of the MPC controller doesnot match the results from the LQR controller. Overall, model-based optimalcontrol show potential as a method for dynamic control and waypoint navigationfor AUV applications, but further work with parameter estimation, modelvalidation and control objective definition is required to ensure feasibility in aphysical implementation.