An optimal washout filter for motion platform using neural network and fuzzy logic. (February 2022)

• Record Type: Journal Article
• Authors: Qazani, Mohammad Reza Chalak; Asadi, Houshyar; Mohamed, Shady; Lim, Chee Peng; Nahavandi, Saeid
• Is Part Of: Engineering applications of artificial intelligence. Volume 108(2022)
• Publication Date: 2022-02
• Subjects: Motion cueing algorithm -- Optimal washout filter -- Time-varying weights -- Fuzzy logic controller -- Riccati equation -- Neural network; Engineering -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Ingénierie -- Informatique -- Périodiques; Intelligence artificielle -- Périodiques; Systèmes experts (Informatique) -- Périodiques; Artificial intelligence; Engineering -- Data processing; Expert systems (Computer science); Periodicals; 620.00285
• DOI: 10.1016/j.engappai.2021.104564
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: To experience the motion sensation of a real vehicle through a motion simulator, a motion cueing algorithm (MCA) is required to transform the vehicle motions to the driving motion platform (DMP) while respecting the physical limitations of DMP. In this aspect, the optimal washout filter (WF) extracts the optimal motion signals including linear accelerations and angular velocities for the DMP with consideration of the human vestibular model and DMP motion states using the linear quadratic regulator (LQR) technique. The LQR technique is employed to obtain the optimal and pre-defined higher order transfer functions by solving the Riccati equation. However, the Riccati equation is solved using fixed weights, leading to an inconvenient usage of the DMP workspace. In this research, a new optimal WF model is designed and developed using a neural network (NN) and a fuzzy logic controller (FLC). The NN is introduced to solve the Riccati equation online while the FLC model is designed to extract the weighting matrices of the LQR technique. The proposed technique considers the physical DMP limitations online and reproduces accurate motion signals with a high degree of fidelity. The results demonstrate the efficiency of the developed optimal WF model as compared with those of existing optimal WF models. Highlights: A novel optimal washout filter is proposed using soft computing techniques. A feedforward NN is trained to solve the Riccati equation online. An FLC is designed to extract the appropriate weights of the LQR technique. The RMSE of the sensed specific force reduces 14.44% along the x-axis. The proposed method increases the motion fidelity via efficient usage of workspace.