On-board vs Ground Spectral Unmixing Simulator

Monte Carlo simulation comparing on-board spectral processing vs ground-based analysis. Analyze trade-offs between bandwidth, latency, and survey efficiency for asteroid prospecting.

Simulation Parameters

10 55 100
1 25 50
3 yr 9 yr 15 yr
5 25 50

Key Trade-offs

  • + On-board: Fast decisions, low bandwidth
  • - On-board: Limited processing power
  • + Ground: Powerful analysis, detailed results
  • - Ground: High latency, bandwidth intensive

Metrics Comparison

Run simulation to see comparison chart

Comparison Results

Configure the parameters and run the simulation to compare on-board vs ground processing.

Simulation Methodology

This simulation models NEA encounter opportunities and analyzes how processing location affects survey success rates. Each encounter has a limited observation window during which data must be captured, processed, and acted upon.

  • On-board processing: Downlinks only processed results (~10 MB), enabling fast autonomous decisions
  • Ground processing: Downlinks raw hyperspectral data (~100s MB), requires round-trip communication
  • Encounters are missed if bandwidth or latency exceeds observation window
  • Latency includes transmission time, light delay, and processing time

Results help determine whether on-board edge computing is worth the development cost for prospecting missions.

This simulator was built to investigate research question RQ-0-2: On-board vs ground spectral unmixing

View Prospecting Satellites BOM Item
D
Project Dyson

A non-profit organization dedicated to realizing a Dyson swarm through detailed planning, research aggregation, and multi-LLM collaboration.

Resources

Community

© 2026 Project Dyson. Open source under MIT license.

Built with Svelte, powered by AI collaboration