Advanced Stewart Flight Simulator Motion Cueing Algorithm Design with Parallel Architectures and Adaptive Switching Mechanism

With the excellent mechanism characteristics of high stiffness and high maneuverability, the 6-Degree-of-Freedom (DoF) Stewart platform has been widely applied to build flight simulator platforms for replicating the motion sensation during pilot training. The fidelity of dynamic simulation depends on the quality of the Motion Cueing Algorithm (MCA). However, conventional MCAs neglect the simulator operating boundary constraints and demonstrate limitations in handling multi-constraint problems. In this paper, a MCA method for Switchable Nonlinear Model Predictive Control (S-NMPC)-based on parallel architecture and Adaptive Switching Mechanism(ASM) is proposed to guarantee solution availability and reduce control errors due to the differences in the terminal state constraints (TSCs) of NMPCs. Within the operating range of the simulator, accurate tracking can be achieved by an NMPC-based MCA, whose TSC has an Equational Constraint (EC); Beyond the operating range of the simulator, optimal approximate tracking can be achieved by a NMPC-based MCA, whose TSC does not possess EC. Moreover, due to the use of the Extended Kalman Filter (EKF) in the ASM, the optimality and smoothness of the system when switching between multiple modes of operation is ensured. Additionally, a method for assessing the fidelity of kinetic simulation based on Multiple Index Fusion Weighting (MIFW) was established by combining the characteristics of objective and subjective assessment. A horizontal stall experiment meeting the MIFW evaluation criteria suggests that the proposed S-NMPC-based MCA was 11.63% and 37.57% higher than the NMPC-based MCA and Classical Washout Filter (CWF)-based MCA, respectively.