Just-in-Time Collision Avoidance

Dust-based remediation for orbital debris conjunction events

Summary

  • Dust-based remediation architecture for debris conjunctions.
  • Forward-only propagation with Lambert targeting and event catalogues.
  • Uncertainty-aware optimization and sigma-point dispersion modeling.

Overview

Orbital debris poses an escalating threat to operational spacecraft. With over 34,000 tracked objects and millions of smaller fragments in orbit, collision risk is increasing exponentially. Traditional collision avoidance relies on maneuvering the threatened asset—but many debris objects are uncontrollable, and frequent maneuvers consume precious fuel that shortens mission lifetime.

My dissertation explores an alternative approach: dust-based just-in-time collision avoidance. The concept deploys a constellation of small "remediation satellites" that monitor predicted conjunctions. When a collision is imminent, a deployer intercepts the debris before the conjunction and releases a targeted dust cloud. The particles impart a gentle momentum change to the debris, nudging it onto a slightly different trajectory—enough to increase miss distance and reduce collision probability without creating additional trackable debris.

Research Questions

  • Constellation Design: How should deployer satellites be distributed to maximize coverage of the debris population while minimizing constellation size and delta-v budgets?
  • Dust Dispersion: How do deployed particles disperse under differential drag, solar radiation pressure, and gravitational perturbations? What deployment strategies maximize momentum transfer efficiency?
  • Uncertainty Propagation: How does uncertainty in atmospheric density, debris state, and deployment conditions affect conjunction predictions and remediation success rates?

Methodology

The evaluation pipeline uses a forward-only propagation architecture optimized for constellation design studies. Conjunction target states are pre-computed into interpolated event catalogues, eliminating expensive repeated propagation during optimization. The constellation is propagated once per event to an intercept epoch (typically 2 hours before conjunction), where a Lambert solver finds feasible transfers and computes required dust mass.

Transfer Solver

Multi-revolution Lambert targeting with configurable policy constraints (max delta-v, minimum lead time, perigee guards)

Dust Physics

Sigma-point propagation with Gaussian mixture splitting for accurate uncertainty quantification

Event Catalogue

1,000-event LEO conjunction bank generated via PCA-KDE debris surrogate sampling

Optimization

NSGA-III multi-objective optimization with adaptive evaluation policies for computational efficiency

Current Evaluation Focus

01

Prototype work

Dispersion-model comparisons in ongoing dissertation research

02

Ongoing study

Dust-dispersion evaluation under richer uncertainty propagation

03

1,000 events

LEO conjunction events in reproducible test bank

04

Rust-backed

Batch evaluation workflow for optimization studies