Information Theoretic Model Predictive Control: Theory and Applications to Autonomous Driving

TL;DR

This paper introduces an information theoretic approach to stochastic optimal control, leading to a sampling-based model predictive control algorithm applied to aggressive autonomous driving, demonstrating its effectiveness over existing methods.

Contribution

The paper develops a novel information theoretic model predictive control framework and applies it to autonomous driving, showing improved performance over traditional methods.

Findings

IT-MPC outperforms cross-entropy based MPC in aggressive driving scenarios

The approach provides a general sampling-based optimization scheme

Demonstrates effectiveness in real-world autonomous vehicle control

Abstract

We present an information theoretic approach to stochastic optimal control problems that can be used to derive general sampling based optimization schemes. This new mathematical method is used to develop a sampling based model predictive control algorithm. We apply this information theoretic model predictive control (IT-MPC) scheme to the task of aggressive autonomous driving around a dirt test track, and compare its performance to a model predictive control version of the cross-entropy method.