Scholarlink

<p>Advanced air mobility (AAM) holds promise for alleviating urban traffic congestion. However, for winged electric vertical takeoff and landing (eVTOL) vehicles, the transition between vertical and cruise flight remains a critical challenge. This maneuver requires a precise change in tilt angle that is difficult for a pilot to perform manually, necessitating a robust autonomous control. This study addresses this gap by implementing an AI autonomous tilting solution for eVTOLs. Reinforcement learning (RL) framework is used, leveraging the Soft Actor-Critic (SAC) algorithm to manage the continuous control problem under various operational constraints. The simulation results demonstrate that the eVTOL can undergo stable tilting within specified limits, validating the SAC algorithm as a viable solution for this task. This research represents a foundational step toward developing the autonomous control systems essential for the safety and operational viability of eVTOLs. Further research is needed to apply this approach to full-scale eVTOLs in more challenging environments.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Implementation of autonomous tilting for AAM eVTOL using a soft actor-critic reinforcement learning approach

  • 期刊论文

  • 发表日期：

    2026年2月9日

查看全文
  • Youngjae Lee,
  • Cheol-kyun Choi,
  • Jungseok Hyun,
  • Hyeon Jun Lee,
  • Yeunun Choo,
  • Donghyun Yoon,
  • L. V. Thang Nguyen,
  • Jae-Woo Lee

摘要

Advanced air mobility (AAM) holds promise for alleviating urban traffic congestion. However, for winged electric vertical takeoff and landing (eVTOL) vehicles, the transition between vertical and cruise flight remains a critical challenge. This maneuver requires a precise change in tilt angle that is difficult for a pilot to perform manually, necessitating a robust autonomous control. This study addresses this gap by implementing an AI autonomous tilting solution for eVTOLs. Reinforcement learning (RL) framework is used, leveraging the Soft Actor-Critic (SAC) algorithm to manage the continuous control problem under various operational constraints. The simulation results demonstrate that the eVTOL can undergo stable tilting within specified limits, validating the SAC algorithm as a viable solution for this task. This research represents a foundational step toward developing the autonomous control systems essential for the safety and operational viability of eVTOLs. Further research is needed to apply this approach to full-scale eVTOLs in more challenging environments.