Exploring deep space and deep ocean with AI guidance
summary
Describes concrete systems and results: uses Laser Induced Breakdown Spectroscopy (LIBS) and Raman for in‑situ composition, develops on‑board ML encoders to compress and classify data, and crowdsources labels via Zuniverse to train terrain classifiers. Terrain relative navigation raises landing success from 85% to 99 percent, Perseverance’s enhanced auto‑nav runs ~six times faster reaching ~120 meters per hour, Ingenuity achieves powered flight and 625‑meter sorties, and an INVADER payload targets 1,500 m deep hydrothermal vents for real‑time spectroscopy.
AI-written summary · about
chapters (17)
- 0:00EXPLORING DEEP SPACE AND DEEP OCEAN WITH AI GUIDANCE
- 0:22OUTLINE
- 1:50IMPOSSIBLE SENSING TECHNOLOGY
- 2:25SPECTROSCOPY
- 3:43MOTIVATION-EXPLORATION
- 4:40MOTIVATION-AIADWANCEMENT
- 6:58MARS EXPLORATION ROVER
- 9:17MER ENTRY DESCENT & LANDING
- 13:56CURIOSITY
- 18:23PERSEVERANCE
- 24:20RUTURE ROVERAI
- 28:10SPACE CHALLENGES
- 29:52RUTURE SPACE MISSIONS
- 31:13OCEAN CONCEPT TESTING
- 32:12OCEAN EXPLORATION-VIPER
- 34:20CONCLUSIONS
- 35:07IMPOSSIBLE SENSING TEAM
related talks
27mAI-ML and space exploration at NASA
24mAI in maritime - Autonomous ships
49mMonitoring global events with AI-driven satellite swarms
28mTed Willke, Sr. Principal Engineer, Intel
30mAI On-board of Satellites, Towards Autonomous Scientific Instruments
31mFrom Radar Echoes To Real-Time Recognition: Deep Learning on Another Planet -An Open Source Workflow
01h 00m(Deep) Learning to Fly
21mML Infrastructure for Autonomous Vehicles