06/10/2025, 05:27 AM UTC
AI助力自主无人机保持航向AI Helps Autonomous Drones Stay On Path
➀ 麻省理工学院研究人员开发了基于神经网络的无人机控制系统，利用元学习实时适应风力变化；
➁ 系统仅需15分钟飞行数据即可自主学习，无需预置风场模型，并动态选择优化策略；
➂ 测试中轨迹误差减少50%，未来将扩展应对多重干扰，支持山火监测和复杂载重场景应用。
➀ MIT researchers developed an AI-powered drone controller using neural networks and meta-learning for real-time adaptation to wind changes;
➁ The system learns from 15 minutes of flight data without requiring pre-defined wind models, selecting optimization strategies dynamically;
➂ In tests, it reduced tracking errors by 50% and shows potential for wildfire monitoring and payload-carrying applications.
---
本文由大语言模型（LLM）生成，旨在为读者提供半导体新闻内容的知识扩展（Beta）。
