AstroMind + MarswalkAI — Demo Video | IBM watsonx × IBM Bob

Advance Space Exploration with AI · NASA Space Apps Challenge 2024

🚀

AstroMind + MarswalkAI

AI-Powered Space Mission Suite · Deep Space Planner + Mars Surface Explorer
Built with IBM Bob · Powered by IBM watsonx Granite

IBM watsonx Granite IBM Bob NASA Mars Data MOLA · CRISM · HiRISE Orbital Mechanics AI EVA Planning

🛸 AstroMind Demo

🔴 MarswalkAI Demo

▶ Press Play to watch both · ← → to navigate · Click buttons above to jump

The Problem

Deep space mission planning
is broken for most teams

NASA and ESA use billion-dollar tools and thousands of specialists to plan missions. University research teams, emerging agencies, and private innovators are completely locked out of this process — creating a massive capability gap.

10,000+

Specialist hours per mission plan

$10M+

Typical planning cost

~50%

Mars landing success rate historically

The Solution

AstroMind makes expert
mission planning instant

🚀 Mission Profile Generator

Δv budgets, launch windows, transit times, life support margins — generated in seconds by IBM Granite

⚠ AI Risk Matrix

Systematically surfaces critical, high and medium failure modes with actionable mitigation strategies

💬 Natural Language Q&A

Ask "What's the optimal 2027 Mars launch window?" and get an expert-level quantified answer instantly

📄 Mission Brief Generator

Auto-drafts formal mission documents ready for submission, review boards, or presentations

🛸

ASTROMIND

AI Deep Space Mission Planner

IBM WATSONX Powered by IBM Bob

Dashboard

Trajectory

Risk Matrix

AI Chat

—

Δv Budget (km/s)

—

Transit (days)

—

Peak Dist (AU)

—

Comm Delay (min)

🛸 ASTROMIND WATSONX AI

AI Architecture

Domain-grounded watsonx pipeline

┌──────────────────────────────────────┐ │ AstroMind Frontend │ │ HTML5 / CSS3 / Vanilla JS │ └───────────────┬──────────────────────┘ │ REST API + IBM IAM Auth ▼ ┌──────────────────────────────────────┐ │ IBM watsonx.ai │ │ ibm/granite-13b-instruct-v2 │ │ • Greedy decoding (deterministic) │ │ • Domain constants injected │ │ • Chain-of-thought for STEM │ └───────────────┬──────────────────────┘ │ ▼ ┌──────────────────────────────────────┐ │ Mission Context Engine │ │ • Orbital mechanics constants │ │ • Structured prompt templates │ │ • Response → mission card render │ └──────────────────────────────────────┘

IBM Bob

Primary development tool — start to finish

1

Architecture Design

Bob evaluated watsonx Granite vs other models for STEM reasoning and designed the domain-grounding strategy

2

Prompt Engineering

Crafted the orbital mechanics constants injection layer that prevents hallucination of physically impossible values

3

Code Generation

All 650+ lines of prototype code — watsonx API, IAM auth, SVG orbit viz, risk cards, typing animation

4

Documentation

README, architecture notes, demo scripts and this submission page all co-authored with Bob

5

Risk Framework

Generated the failure-mode taxonomy covering launch, transit, EDL, surface ops and crew health

Real-World Impact

From weeks to minutes

60–80%

Reduction in preliminary mission assessment time

6

Destinations supported: Mars · Europa · Titan · Asteroid · Lunar Gateway · Venus

5

AI-powered modules: Profile · Risk · Chat · Timeline · Brief

0

Server required — runs in any browser, API key optional (Demo Mode)

AstroMind — because the future
of space exploration
belongs to everyone

github.com/walsal201/ai-astromind

IBM watsonx IBM Bob Advance Space Exploration 2024

NASA Space Apps Challenge 2024 · New Feature

🔴

MarswalkAI

Layered NASA Mars Surface Explorer
Integrated multi-mission data for human EVA planning

MOLA Topography CRISM Mineralogy HiRISE 25cm/px THEMIS Thermal MCS Weather Perseverance Data

The Mars Challenge

The first astronauts on Mars need the best map possible

NASA has explored Mars robotically for decades — collecting topography, mineralogy, atmospheric, and terrain data. But none of that data has ever been unified into a single tool designed for a human explorer on the surface, in a suit, making real-time decisions about where to walk, what to sample, and when to turn back.

9

NASA missions · data sources integrated

10+

EVA planning modules in one tool

0

Prior tools unifying all layers for human EVA

The Solution

MarswalkAI — one layered view, every data source

🗺 Layered Surface Map

MOLA topography + CRISM minerals + THEMIS thermal + HiRISE terrain — all toggled on one interactive canvas

🚶 AI Route Planner

IBM watsonx generates phase-by-phase EVA plans with waypoints, hazard avoidance, and science stops timed to the minute

⚠ Hazard Intelligence

Boulder fields from HiRISE, soft regolith from THEMIS, radiation from NOAA — flagged automatically on the route

📋 Mission Brief

Formal NASA-style EVA Operations Order auto-generated — ready for a mission planning review board submission

🔴

MARSWALKAI

NASA Mars Surface Explorer · IBM watsonx

NASA DATA IBM WATSONX

🗺 Surface Map

📡 NASA Data

🚶 EVA Plan

⚠ Hazards

18.4°N

Latitude

−2,900m

MOLA Elev

−63°C

Surface Temp

Tau 0.5

Dust Opacity

🟫 MOLA Topo🟢 Minerals (CRISM)🔵 Route🔴 Hazards🟡 Science Targets🟣 HiRISE Coverage

🔴 MARSWALKAI WATSONX AI

NASA Multi-Mission Data

9 missions, one integrated view

MOLA · MGS

1/128° elevation maps. Jezero floor at −2,900m — ancient lake basin confirmed

CRISM · MRO

Mg-carbonates + olivine detected at delta scarp — prime biosignature preservation

THEMIS · Odyssey

Thermal inertia 420 TIU — compact regolith. Low-TI zones = soft dust → AVOID

HiRISE · MRO

25 cm/px imagery — boulders, craters, dust devil tracks resolved at human scale

MCS · MRO

Atmospheric profiles — dust τ=0.5 clear, wind 12 m/s, 700 Pa. EVA: GO

Perseverance · M2020

1,200+ sol traverse — RIMFAX subsurface, MEDA climate, 24 sample tubes cached

Curiosity MSL · Heritage REMS data InSight · HP³ Thermal Phoenix MECA · Perchlorate chemistry

EVA Operations

AI-generated EVA Timeline + Mission Brief

LMST EVA SCHEDULE — JEZERO CRATER · SOL +5 · 09:00 START

Resource Budget

O₂: 0.84 kg used · 0.42 kg reserve ✅
Battery: 86% consumed → NOMINAL ✅
Water: 1.5L consumed · 0.5L reserve ✅
Samples: ~800g (6 cores + 3 rover tubes)

GO / NO-GO Criteria

✅ Dust τ = 0.5 < 1.5 (CLEAR)
✅ Wind 12 m/s < 25 m/s (SAFE)
✅ No SPE alert (space weather QUIET)
✅ EVA window: 09:00–15:00 LMST

MarswalkAI + AstroMind
—
From orbit to bootprint,
AI guides the way

walsal201.github.io/ai-astromind

MarswalkAI AstroMind IBM watsonx IBM Bob NASA Space Apps 2024

Launch MarswalkAI → · Back to Home