Robust Control Design for Autonomous Vehicles Using Neural Network-Based Model-Matching Approach

Abstract

In this paper, a novel neural network-based robust control method is presented for a vehicle-oriented problem, in which the main goal is to ensure stable motion of the vehicle under critical circumstances. The proposed method can be divided into two main steps. In the first step, the model matching algorithm is proposed, which can adjust the nonlinear dynamics of the controlled system to a nominal, linear model. The aim of model matching is to eliminate the effects of the nonlinearities and uncertainties of the system to increase the performances of the closed-loop system. The model matching process results in an additional control input, which is computed by a neural network during the operation of the control system. Furthermore, in the second step, a robust H∞ is designed, which has double purposes: to handle the fitting error of the neural network and ensure the accurate tracking of the reference signal. The operation and efficiency of the proposed control algorithm are investigated through a complex test scenario, which is performed in the high-fidelity vehicle dynamics simulation software, CarMaker.

Tuning Path Tracking Controllers for Autonomous Cars Using Reinforcement Learning

“…Literature on the subject of autonomous driving control is populated with Neural Network (NN) based controllers in the role of supervisors of the tracking function (see for instance [27] on NNs to match the model and subsequent robust control design). For the proposed architecture, the argument on dependability developed in the previous section shows that the overall system has a stability property (which amounts to safe driving).…”

Prediction for Future Yaw Rate Values of Vehicles Using Long Short-Term Memory Network

“…In addition, the roll rate has an important role not just in the automotive industry but in the shipping industry [ 9 ]. Artificial neural networks are widely used in the yaw rate control of the vehicle, but only as a control algorithm and not as an estimation method for the yaw rate [ 10 , 11 ]. The number of articles on yaw rate prediction is limited.…”

“…For measuring the performance of the LSTM network, the following loss measures were calculated: mean squared error (MSE), root mean squared error (RMSE), mean absolute error (MAE) and R 2 score. The calculation of the error measures is given by Equations ( 7)- (10).…”

Analysis and Synthesis of Coordinated Control Systems for Automated Road Vehicles

“…Using data within the vehicle control design process on the level of autonomous vehicle modeling is another approach. In [3], a neural-network-based model-matching process was developed, the aim of which is to handle the nonlinear characteristics of vehicle dynamics. Thus, a novel robust control algorithm is proposed, which exploits the advantages of both the classical control and machine-learning-based methods.…”