Adaptive Filter Applications to Autonomous Underwater Vehicle

A powerful adaptive filter algorithm for estimation of forces acting on an Autonomous Underwater Vehicle (AUV), and the corresponding adaptive control law which determines the desired heading commands are considered in this paper. Due to the complicated nature of AUV’s task and incorporation of the ocean current, uncertain parameters arise in thc equations of motion that constitute the AUV’s dynamic model. Based on the Lainiotis Partitioning Theory, a highly adaptive scheme is used to estimate the vclocity components, for controlling the vehicle. Extensive simulations demonstrate the robustness and the effcctiveness of the new scheme in a variety of different environments. I . I N T R O D U C T I O N Autonomous Underwater Vehicles (AUVs) are an important tool for various undersea tasks such as construction [ 11, repairing of underwater equipment [2], waste management, pollution monitoring, undersea research and data collection [3]. Due to the complicated nature of the tasks and the potential for assuming various payloads, there is a need for tight control and maneuverability of the AUV under large variations in the operating conditions. A robust control algorithm that can enhance the efficiency of the vehicle is therefore highly desirable [2],[3],[4]. However, due to the ocean current and the nature of the vehicle most of the available methods have limited applicability [2],[4]. The dynamics of the autonomous underwater vehicle are nonlinear in nature due to rigid body coupling and the hydrodynamic forccs of thc vehicle. Moreover most autonomous vehicles perform maneuvers in different directions using thrusters, which makes the uncertainties even more severe. Recently, several control suategies for the AUVs have been discussed. Most of these methods require linearized models. The main problem in these methods is that the robustness of the control scheme with respect to parameter uncertainties cannot be guaranteed, since the linearized set of equations have to be corrected almost continuously, to account for frequent changcs in the operating point of the vehicle. Hence an adaptive contol scheme is required that will account for these hydrodynamic uncertainties and variable loading conditions. The proposed strategy is called the Partitioned Adaptive Control. Based on the general theory of Partitioning algorithms, the new method uses the Adaptive Lainiotis Filter for estimation and a quadratic performance criterion, to calculate the required control input [71,[91,[101. Specifically, the paper is organized as follows: Section I1 describes the model representing the AUV motion. In Section 111 the Adaptive Lainiotis approach to estimation and control is discussed. The simulation results and conclusions are presented in Sections IV and V, respectively.