Deep Truck : A deep neural network model for longitudinal dynamics of heavy duty trucks

Abstract

This article demonstrates the use of deep neural networks (NN) and deep reinforcement learning (deep-RL) for modeling and control of longitudinal heavy duty truck dynamics. Instead of explicit use of analytical model derived information or parameters about the truck, the deep NN model is fitted to data using a brief set of historical data collected from an arbitrary driving cycle. The deep model is used in this article to design a cruise controller for the truck using model-free deep-RL. The deep model and the control loop performances are demonstrated both using state-of-the-art commercial simulation software, and using a real-physical truck. Model and control performances are compared to classical physics-based modeling and control design approaches. The deep NN model is shown to capture latent nonlinear state dynamics and the deep-RL cruise controller is shown to achieve comparable results to a carefully designed and calibrated controller.