Scholarlink

<p>In recent years, quadrotor aircraft have been extensively studied due to their simple control algorithms and broad range of applications. However, the mobility limitations resulting from their underactuated nature cannot be ignored. To address this, this paper develops a quadrotor with bidirectional thrust to enhance maneuverability. First, the dynamic quadrotor model is established, and the expanded torque range and necessary stability enabled by bidirectional thrust are analyzed. Subsequently, innovative task scenarios, including slope landing, inverted hovering and ceiling adhesion are explored. The task trajectories are formulated as optimal control problems using the pseudospectral method, and the planning results are learned through Dynamic Movement Primitives, facilitating rapid generalization from different initial conditions to simulate diverse tasks. Fractional-order PID controllers and a control allocation method optimized for bidirectional thrust are implemented to ensure that the control performance aligns with planning expectations. The simulation results demonstrate the versatility of the bidirectional thrust quadrotor in performing tasks within novel scenarios. The learning and generalization algorithms highlight the adaptability of such quadrotors to environmental conditions based on planning, while the controller ensures stable and precise control.</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

The Motion Planning, Learning and Control of a Bidirectional Thrust Quadrotor with Special Tasks

  • 期刊论文

  • 发表日期：

    2026年2月13日

查看全文
  • Lihao Xu,
  • Zhiduan Cai,
  • Yuling Wang,
  • Ruiqi Cai,
  • Yiling Liu

摘要

In recent years, quadrotor aircraft have been extensively studied due to their simple control algorithms and broad range of applications. However, the mobility limitations resulting from their underactuated nature cannot be ignored. To address this, this paper develops a quadrotor with bidirectional thrust to enhance maneuverability. First, the dynamic quadrotor model is established, and the expanded torque range and necessary stability enabled by bidirectional thrust are analyzed. Subsequently, innovative task scenarios, including slope landing, inverted hovering and ceiling adhesion are explored. The task trajectories are formulated as optimal control problems using the pseudospectral method, and the planning results are learned through Dynamic Movement Primitives, facilitating rapid generalization from different initial conditions to simulate diverse tasks. Fractional-order PID controllers and a control allocation method optimized for bidirectional thrust are implemented to ensure that the control performance aligns with planning expectations. The simulation results demonstrate the versatility of the bidirectional thrust quadrotor in performing tasks within novel scenarios. The learning and generalization algorithms highlight the adaptability of such quadrotors to environmental conditions based on planning, while the controller ensures stable and precise control.