{
  "abstract": "Terrain-relative navigation (TRN) lets a lander correct its position estimate against an onboard map, and a classical guidance\n\nbaseline with a TRN filter already lands precisely in nearly all default cases, leaving little room for learning. We study where\n\na learned policy could add anything, using an open 3-DoF lunar landing benchmark built from Lunar Orbiter Laser Altimeter\n\n(LOLA) polar digital elevation models (Site 04, Site 23, Haworth) with a geographic held-out split, a full navigation filter in the\n\nloop, and a realism layer. The remaining headroom appears only at large initial position error, where failures come from missing\n\nterrain fixes rather than from flying, and classical changes to the filter and sensor recover most of the loss: a terrain-match search\n\nwindow that covers the initial uncertainty, and a camera field of view of at least 40◦ for the remainder. We then compare two uses\n\nof reinforcement learning against these fixes. Residual proximal policy optimisation (PPO), which corrects the thrust command of\n\nthe baseline, did not improve it. A decision-level policy, which chooses whether to continue, hold, or detour, has real headroom:\n\na one-line detour rule raises precise landings by 11 to 12 points on a synthetic blackout benchmark, and a learned offline chooser\n\nmatches the rule but does not exceed it, while sampled PPO policies fall below it. We release the benchmark, code, and negative\n\nresults. Submitted to Acta Astronautica, October 2026. Code and benchmark: https://github.com/Weirdnemo/lunar-trnrl",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": null,
  "author_total": 1,
  "authors": [
    {
      "name": "Nimesh Chauhan",
      "orcid": "0009-0007-3934-7561"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/f14nh8pp8nrpk.bib",
  "cited_by": 0,
  "doi": "10.5281/zenodo.23642853",
  "fulltext_url": null,
  "id": "f14nh8pp8nrpk",
  "issns": [],
  "landing_url": "https://zenodo.org/records/23642853",
  "language": "eng",
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [],
  "openalex": null,
  "pdf_url": "https://zenodo.org/api/records/23642853/files/manuscript.pdf/content",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-25",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/f14nh8pp8nrpk.ris",
  "sources": [
    {
      "first_seen": "2026-11-08T12:21:11Z",
      "native_id": "23642853",
      "source": "zenodo",
      "url": "https://zenodo.org/records/23642853"
    }
  ],
  "subjects": [],
  "title": "Where Does Reinforcement Learning Help Lunar Terrain-Relative Landing? Residual Control vs. Decision-Level Policies",
  "title_text": "Where Does Reinforcement Learning Help Lunar Terrain-Relative Landing? Residual Control vs. Decision-Level Policies",
  "type": "preprint",
  "updated_at": "2026-11-08T12:21:11Z",
  "url": "https://research.draconitic.com/paper/f14nh8pp8nrpk",
  "venue": "Zenodo",
  "versions": []
}