Hallucination-Aware Generative Pretrained Transformer for Cooperative Aerial Mobility Control

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ahn, Hyojun, Oh, Seungcheol, Kim, Gyu Seon, Jung, Soyi, Park, Soohyun, Kim, Joongheon
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916691210928128
author Ahn, Hyojun
Oh, Seungcheol
Kim, Gyu Seon
Jung, Soyi
Park, Soohyun
Kim, Joongheon
author_facet Ahn, Hyojun
Oh, Seungcheol
Kim, Gyu Seon
Jung, Soyi
Park, Soohyun
Kim, Joongheon
contents This paper proposes SafeGPT, a two-tiered framework that integrates generative pretrained transformers (GPTs) with reinforcement learning (RL) for efficient and reliable unmanned aerial vehicle (UAV) last-mile deliveries. In the proposed design, a Global GPT module assigns high-level tasks such as sector allocation, while an On-Device GPT manages real-time local route planning. An RL-based safety filter monitors each GPT decision and overrides unsafe actions that could lead to battery depletion or duplicate visits, effectively mitigating hallucinations. Furthermore, a dual replay buffer mechanism helps both the GPT modules and the RL agent refine their strategies over time. Simulation results demonstrate that SafeGPT achieves higher delivery success rates compared to a GPT-only baseline, while substantially reducing battery consumption and travel distance. These findings validate the efficacy of combining GPT-based semantic reasoning with formal safety guarantees, contributing a viable solution for robust and energy-efficient UAV logistics.
format Preprint
id arxiv_https___arxiv_org_abs_2504_10831
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Hallucination-Aware Generative Pretrained Transformer for Cooperative Aerial Mobility Control
Ahn, Hyojun
Oh, Seungcheol
Kim, Gyu Seon
Jung, Soyi
Park, Soohyun
Kim, Joongheon
Artificial Intelligence
Robotics
68T05
This paper proposes SafeGPT, a two-tiered framework that integrates generative pretrained transformers (GPTs) with reinforcement learning (RL) for efficient and reliable unmanned aerial vehicle (UAV) last-mile deliveries. In the proposed design, a Global GPT module assigns high-level tasks such as sector allocation, while an On-Device GPT manages real-time local route planning. An RL-based safety filter monitors each GPT decision and overrides unsafe actions that could lead to battery depletion or duplicate visits, effectively mitigating hallucinations. Furthermore, a dual replay buffer mechanism helps both the GPT modules and the RL agent refine their strategies over time. Simulation results demonstrate that SafeGPT achieves higher delivery success rates compared to a GPT-only baseline, while substantially reducing battery consumption and travel distance. These findings validate the efficacy of combining GPT-based semantic reasoning with formal safety guarantees, contributing a viable solution for robust and energy-efficient UAV logistics.
title Hallucination-Aware Generative Pretrained Transformer for Cooperative Aerial Mobility Control
topic Artificial Intelligence
Robotics
68T05
url https://arxiv.org/abs/2504.10831