Wave driven robot multibody yawing filter under uncertain model perturbation

Abstract

Abstract Wave driven robot (WDR) is a new type of ocean vehicle with special multibody structure. Aiming at the WDR multibody heading filtering problem under model perturbation, the WDR multibody yawing filter (WYF) with dynamic adaptive model was proposed in this paper. The WYF algorithm included adaptive model construction loop and filter calculation loop. The tracking differentiator (TD) with adaptive tracking speed was designed to fit the generalized rudder angles of the floating body (Float) and the submerged body (Glider). Based on the Taylor series expansion, the dynamic adaptive yawing response model of WDR with delay compensation was proposed. The filter calculation loop was developed within the extended Kalman filter (EKF) framework, which could estimate the multibody heading information of WDR and modify the dynamic adaptive yawing response model at the same time. The WYF algorithm did not rely on WDR dynamic mechanism model, but only extracted system information from the output data of yawing system. Simulation experiments and sea trials were carried out and the results showed that the WYF could keep fast tracking speed and smooth the noisy data effectively under both precise model and perturbed model.