Multi-Predictor Fusion: Combining Learning-based and Rule-based Trajectory Predictors

TL;DR

This paper introduces multi-predictor fusion (MPF), a method that combines learning-based and rule-based trajectory predictors to improve autonomous vehicle planning in complex traffic scenarios, demonstrating superior and consistent performance.

Contribution

The paper proposes MPF, a novel algorithm that probabilistically fuses learning- and rule-based predictors, enhancing trajectory prediction accuracy for autonomous vehicles.

Findings

MPF outperforms standalone predictors on multiple metrics.

MPF provides more consistent trajectory predictions.

Fusion improves safety and efficiency in AV planning.

Abstract

Trajectory prediction modules are key enablers for safe and efficient planning of autonomous vehicles (AVs), particularly in highly interactive traffic scenarios. Recently, learning-based trajectory predictors have experienced considerable success in providing state-of-the-art performance due to their ability to learn multimodal behaviors of other agents from data. In this paper, we present an algorithm called multi-predictor fusion (MPF) that augments the performance of learning-based predictors by imbuing them with motion planners that are tasked with satisfying logic-based rules. MPF probabilistically combines learning- and rule-based predictors by mixing trajectories from both standalone predictors in accordance with a belief distribution that reflects the online performance of each predictor. In our results, we show that MPF outperforms the two standalone predictors on various metrics and delivers the most consistent performance.