Delayed-Decision Motion Planning in the Presence of Multiple Predictions

2025-02-28

AI Summary

Autonomous driving systems face safety challenges in motion planning due to the unknown intentions and highly stochastic behaviors of traffic participants. To address this, this paper proposes a receding-horizon motion planning framework grounded in multi-hypothesis probabilistic future prediction. Its core innovation lies in the first integration of the maximum entropy principle with delayed decision-making, enabling a safety-oriented, flexible behavior planning model; it further formalizes the decision optimization problem under multi-hypothesis predictions. Additionally, we devise an MPC-inspired quadratic programming reformulation that admits real-time solution. Extensive evaluations on both simulation benchmarks and a real-world mobile robot platform demonstrate that the proposed method significantly enhances planning safety—particularly in complex interactive scenarios—while maintaining computational efficiency suitable for real-time deployment.

Technology Category

• Multiagent Systems: Multiagent Planning
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Replanning and Plan Repair

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Agentic search

Abstract

Reliable automated driving technology is challenged by various sources of uncertainties, in particular, behavioral uncertainties of traffic agents. It is common for traffic agents to have intentions that are unknown to others, leaving an automated driving car to reason over multiple possible behaviors. This paper formalizes a behavior planning scheme in the presence of multiple possible futures with corresponding probabilities. We present a maximum entropy formulation and show how, under certain assumptions, this allows delayed decision-making to improve safety. The general formulation is then turned into a model predictive control formulation, which is solved as a quadratic program or a set of quadratic programs. We discuss implementation details for improving computation and verify operation in simulation and on a mobile robot.

Problem

• Addresses behavioral uncertainties in automated driving
• Formalizes planning with multiple possible future behaviors
• Enhances safety via delayed decision-making and predictive control

Innovation

• Delayed decision-making for safety
• Maximum entropy behavior planning
• Model predictive control formulation