Mobility Prediction of Off-Road Ground Vehicles Using a Dynamic Ensemble of NARX Models

Abstract

Mobility prediction of off-road Autonomous Ground Vehicles (AGV) in uncertain environments is essential for their model-based mission planning especially in the early design stage. While surrogate-modeling methods have been developed to overcome the computational challenge in simulation-based mobility prediction, it is very challenging for a single surrogate model to accurately capture the complicated vehicle dynamics. With a focus on vertical acceleration of an AGV under off-road conditions, this paper proposes a surrogate modeling approach for AGV mobility prediction using a dynamic ensemble of Nonlinear Autoregressive Network with Exogenous inputs (NARX) models over time. Synthetic vehicle mobility data of an AGV are first collected using a limited number of high-fidelity simulations. The data are then partitioned into different segments using a variational Gaussian mixture model in order to represent different vehicle dynamic behaviors. Based on the partitioned data, multiple surrogate models are constructed under the NARX framework with different numbers of lags. The NARX models are then assembled together dynamically over time to predict the mobility of the AGV under new conditions. A case study demonstrates the advantages of the proposed method over the classical NARX models for AGV mobility prediction.

Explicit model predictive control of magnetorheological suspension for all-terrain vehicles with road preview

“…Then, the MRD was tested via the MTS platform (figure 2 Mechanical test results of the MRD show that the characteristics of the MRF lead to a significant hysteresis nonlinearity in the developed damper, which can lead to significant errors if approximated by a conventional linear model. The dynamic NARX neural network integrates the delayed network outputs into the network inputs, incorporates the historical system state information into the recursive iterations, and continuously enhances the network's resistance to sample disturbances such as noise [30]. As a result, the NARX neural network can effectively characterize complex nonlinear timevarying systems and exhibit robustness and high accuracy in prediction [31].…”