PRIDE: a hierarchical, integrated prediction framework for autonomous on-road driving

Abstract

In this paper, we present PRIDE (prediction in dynamic environments), a hierarchical multi-resolutional framework for moving object prediction that incorporates multiple prediction algorithms into a single, unifying framework. PRIDE is based upon the 4D/RCS (real-time control system) architecture and provides information to planners at the level of granularity that is appropriate for their planning horizon. The primary contribution of this paper is the approach that is used to provide a more synergistic framework, in which the results from one prediction algorithm are used to strengthen/weaken the results of another prediction algorithm. In particular, we describe how the results of an estimation-theoretic short-term (ST) prediction algorithm are used to validate the results of a situation-based long-term (LT) prediction algorithm. We show the results of the long-term prediction both before and after the results of the short-term predictions are integrated, and demonstrate how the latter provide better predictions of vehicle future positions using the AutoSim simulation package