Reasoning Knowledge-Gap in Drone Planning via LLM-based Active Elicitation

TL;DR

This paper presents a novel LLM-based active elicitation framework for UAV planning that reduces human intervention by explicitly reasoning about knowledge gaps, improving success rates in complex search-and-rescue tasks.

Contribution

It introduces MINT, a neuro-symbolic reasoning mechanism that formulates minimal binary queries to efficiently resolve environmental ambiguities in UAV planning.

Findings

Significant improvement in success rates for search-and-rescue tasks.

Reduced human interaction compared to baseline methods.

Validated in both simulation and real-world UAV deployments.

Abstract

Human-AI joint planning in Unmanned Aerial Vehicles (UAVs) typically relies on control handover when facing environmental uncertainties, which is often inefficient and cognitively demanding for non-expert operators. To address this, we propose a novel framework that shifts the collaboration paradigm from control takeover to active information elicitation. We introduce the Minimal Information Neuro-Symbolic Tree (MINT), a reasoning mechanism that explicitly structures knowledge gaps regarding obstacles and goals into a queryable format. By leveraging large language models, our system formulates optimal binary queries to resolve specific ambiguities with minimal human interaction. We demonstrate the efficacy of this approach through a comprehensive workflow integrating a vision-language model for perception, voice interfaces, and a low-level UAV control module in both high-fidelity NVIDIA Isaac simulations and real-world deployments. Experimental results show that our method achieves a significant improvement in the success rate for complex search-and-rescue tasks while significantly reducing the frequency of human interaction compared to exhaustive querying baselines.