Venue 6 of 6 · Aerospace autonomy · Archival journal
AIAA Journal of Aerospace Information Systems
AIAA's journal for aerospace computing, information, networks and autonomy. It has a particular interest in autonomous systems and mission assurance. This is the capstone: the complete system, all four hypotheses, simulation plus hardware, in one archival paper.
JAIS's scope names autonomous systems, uncertainty quantification and machine learning for aerospace, and accepts both full papers and shorter Technical Notes. The paper page already lists it as a target.
Journal reviewers expect completeness. All of H1–H4, the ablations, the real-data check and the testbed. Submitting before that wastes a review cycle. A Technical Note on the Phase 2 single-planner result is a reasonable earlier step if you want a journal line sooner.
Venue facts
| Theme | Original archival research on aerospace computing, information, networks and communication: new theory, novel applications, case studies. Particular interest in autonomous systems, systems engineering, and safety and mission assurance. |
|---|---|
| Article types | Full-length papers and Technical Notes. |
| Format | AIAA style: English (American spelling), single-column, double-spaced 10-pt. Single-paragraph summary abstract of 100–200 words. Page and word limits are in AIAA's “Journal Page Limits and Word Count Guidelines,” which couldn't be retrieved here. Check before writing. |
| Sources | AIAA journal scopes · AIAA journal author resources |
Tailored abstract
Open-World, Deadline-Aware Hypothesis Discrimination for Distributed Space Observatories
Autonomous constellations following up fading astronomical transients must decide, without ground intervention, which spacecraft takes which measurement while the discriminating signal decays and the true explanation may be unmodeled. We formulate follow-up as open-world sequential model discrimination with time-decaying measurement value, executed by spacecraft with intermittently synchronised beliefs. The planner combines expected discrimination, expected evidence of model inadequacy, and opportunity decay. It switches to evidence preservation when inadequacy becomes probable and allocates measurements through a stale-belief-tolerant marginal-gain auction. We evaluate it on a simulated constellation propagated from real orbital elements, against reactive, uncertainty-sampling and closed-world baselines under identical budgets, and on an ESP32 hardware-in-the-loop testbed. Result sentences pending: H1–H4 outcomes from preregistered test seeds.
Paper plan (full-length)
- Introduction§1The manuscript's §1 and research question, with an aerospace framing: autonomy for time-critical science.
- Related work§2Manuscript Table 1, every entry verified and read in full.
- Formulation and planner§3–4Eqs. 1–4, the mode switch, and the auction with its exact-merge proof.
- Experimental design§5The preregistration verbatim, with seeds, scenarios and statistics.
- Results: centralised planner§6H1–H3 and the ablations.
- Results: decentralisation§7H4: communication sweep and node loss.
- Real-data validation§8Model and noise validation against ZTF, plus the counterfactual replay.
- Hardware-in-the-loop§9Testbed results and sim-vs-hardware agreement.
- Limitations, threats to validity, conclusion§10Manuscript §9.2 and Table 5, updated with what actually happened, including any refuted hypothesis.
Before you submit
- Phases 2, 3 and 4 run on preregistered test seeds, all tables generated from
research/results/. - ESP32 testbed built and measured.
- Retrieve AIAA's page-limit guideline and fit the paper to it.
- DOI-archived code release, as the manuscript's §10 promises.