Search papers, labs, and topics across Lattice.
This paper introduces a framework for human-AI collaboration in UAV planning that uses active information elicitation to address environmental uncertainties. The core of the framework is the Minimal Information Neuro-Symbolic Tree (MINT), which structures knowledge gaps as queryable formats. By leveraging LLMs to formulate optimal binary queries, the system minimizes human interaction while resolving ambiguities, leading to improved success rates in search-and-rescue tasks compared to exhaustive querying.
LLMs can orchestrate human input to UAVs, dramatically improving mission success rates while minimizing human interaction.
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.
