Motion Planning under Uncertainty: Integrating Learning-Based Multi-Modal Predictors into Branch Model Predictive Control

TL;DR

This paper introduces a real-time motion planning framework for autonomous vehicles that integrates learning-based multi-modal predictions with branch model predictive control, improving safety and comfort in complex traffic scenarios.

Contribution

It presents a novel BMPC framework that efficiently incorporates multi-modal predictions and adaptive decision postponing for better handling of uncertainty.

Findings

Enhanced safety and comfort in traffic scenarios

Real-time capability with minimal prediction sets

Improved decision robustness through postponing strategies

Abstract

In complex traffic environments, autonomous vehicles face multi-modal uncertainty about other agents' future behavior. To address this, recent advancements in learningbased motion predictors output multi-modal predictions. We present our novel framework that leverages Branch Model Predictive Control(BMPC) to account for these predictions. The framework includes an online scenario-selection process guided by topology and collision risk criteria. This efficiently selects a minimal set of predictions, rendering the BMPC realtime capable. Additionally, we introduce an adaptive decision postponing strategy that delays the planner's commitment to a single scenario until the uncertainty is resolved. Our comprehensive evaluations in traffic intersection and random highway merging scenarios demonstrate enhanced comfort and safety through our method.