用LSTM增强的TD3和行为克隆用于无人机轨迹跟踪控制.
Yuanhang Qi1, Jintao Hu2, Fujie Wang2
1School of Computer Science, University of Electronic Science and Technology of China, Zhongshan Institute, Zhongshan 528402, China.
Biomimetics (Basel, Switzerland)
|September 26, 2025
概括
一个新的生物灵感深度强化学习 (DRL) 算法增强了无人机 (UAV) 轨迹跟踪. 在复杂的环境中,TD3-LSTM-BC框架提高了控制精度和稳定性.
科学领域:
- 机器人技术 机器人技术 机器人技术
- 人工智能的人工智能
- 控制系统 控制系统
背景情况:
- 由于非线性动态和外部干扰,无人机在动态环境中难以准确地跟踪轨迹.
- 现有的控制方法往往需要准确的系统模型,这些模型对于复杂的无人机操作来说很难获得.
研究的目的:
- 开发一种新的生物灵感深度强化学习 (DRL) 算法,用于高精度的无人机轨迹跟踪.
- 加强无人机控制政策的学习,而不依赖于准确的系统动态模型.
主要方法:
- 在双延迟深确定性政策梯度 (TD3) 算法 (TD3-LSTM-BC) 中整合行为克隆 (BC) 和长短期记忆 (LSTM) 网络.
- 利用LSTM进行时间模式识别,以提高适应轨道变化和应对干扰的弹性.
- 雇佣BC进行DLR政策的预培训,并通过专家演示来加快融合.
主要成果:
- 拟议的TD3-LSTM-BC方法在模拟实验中表现出卓越的稳定性和跟踪精度.
- LSTM模块增强了政策网络捕获时间状态模式的能力,优于无内存网络.
- 行为克隆有效地初始化了政策,减少了探索时间并提高了学习效率.
结论:
- TD3-LSTM-BC框架为在具有挑战性的环境中使用自主无人机提供了强大而准确的控制解决方案.
- 生物启发的深度强化学习,结合模仿和自适应优化,显示了无人机控制的巨大潜力.
- 该方法成功地模仿了自然的学习过程,平衡了先天的指导与经验性适应,以提高绩效.
相关概念视频
01:24Absolute Motion Analysis- General Plane Motion
538
Visualize a drone, with its propellers spinning rapidly, hovering mid-air. The fascinating movements and operations of this drone can be comprehended by applying the principle of general plane motion.
As the drone's propellers rotate, an upward force is generated that counteracts the force of gravity, enabling the drone to lift off from the ground. This initial movement of the drone is along a straight path, representing a form of translational motion. In this phase, every point on the...
As the drone's propellers rotate, an upward force is generated that counteracts the force of gravity, enabling the drone to lift off from the ground. This initial movement of the drone is along a straight path, representing a form of translational motion. In this phase, every point on the...
538
01:26Orthogonal Trajectories
8
Orthogonal trajectories describe the geometric relationship between two families of curves that intersect each other at right angles. One illustrative case involves a family of parabolas that open sideways along the x-axis. These curves share a common shape but differ by a scaling parameter, resulting in a set of curves that all pass through the origin and widen at different rates.Determining Orthogonal TrajectoriesTo identify the orthogonal trajectories for these parabolas, the first step...
8