智能体
hackernoon.com ·
@vishwa vimukthi gammuduwaththage
用于自主导航的 AI 无人机 Agent
文章介绍自主无人机导航的技术栈：感知（相机、IMU、GPS/RTK、SLAM、VIO）、机载 AI 计算、控制（PX4/ArduPilot 经 MAVLink）与通信四层。AI 层涵盖用于感知的 CNN、RNN、Transformer 与用于导航策略的强化学习，路径规划使用 A*、RRT、PRM。2025 年硬件以 NVIDIA Jetson Orin/Xavier 级算力、全局快门/HDR/红外相机和 MEMS 或光纤 IMU 为主。文章指出 VLN 与 VLA 模型（如 AutoFly、GRaD-Nav++）正从实验室走向机载，但在高频反应式控制上仍有困难。
AI推荐理由 系统梳理自主无人机导航的四层架构与 AI 技术栈，提及 VLN、VLA 模型（AutoFly、GRaD-Nav++）及 Jetson Orin/Xavier 等硬件，对具身智能与边缘算力有参考价值；但为综述性内容，未给出具体实验数据或来源引用。
66
正文
How Companion.energy Reduced Query Latency 25x & Compressed Terabytes to Gigabytes with Tiger Cloud
See how Companion.energy migrated from Azure PostgreSQL to Tiger Cloud, achieving 25x faster queries, 98% storage savings, and faster real-time analytics.
by Tiger Data (creators of TimescaleDB)
Sep 23, 2026 · 8 min read
#COMPANION.ENERGY-TIGER-CLOUD
The Best ETL Tools for Real-Time Data Integration in 2026
Real-time doesn't mean the same thing to every ETL tool. Compare Estuary, Confluent, Fivetran, and more to find the right fit for your data.
by Pipelineguy
Sep 23, 2026 · 16 min read
#ETL
Make a dedicated Wake-on-LAN server with Tailscale
Learn how to combine Tailscale with Wake-on-LAN to wake sleeping PCs and servers remotely using a Raspberry Pi, UpSnap, or Docker.
by ameliepangolin
Sep 23, 2026 · 6 min read
#TAILSCALE-WAKE-ON-LAN