Enhancing Reduced Order Model Predictive Control for Autonomous Underwater Vehicle

Performing control actions to complete the mission depending on the feedback measurements is a crucial task in case of Autonomous Underwater Vehicle due to dependency on navigation system. For such task a Reduced Order Model Predictive Control (ROMPC) has been implemented using highly nonlinear model of AUV to control motion in thee dimensional space under the assumption that the feedback measurements at every iteration are clearly available to solve the quadratic problem. But in real-time scenario, navigation system collects measurements from the sensors installed on the hardware part of the AUV which may fail due to vulnerability of sensors to onboard equipment noise or poor signal during diving operation resulting in failure of ROMPC furthermore mission, hence proper state estimator or observer is required for real time operation to support navigation system. This work proposes a solution, based on the optimal estimation property of Extended Kalman Filter in the presence of process and measurement noise or missing measurements to estimate position and orientation of AUV for successful and enhanced feedback control application.