Answers are grounded in this paper and cite their sources.
Rovers approaching lunar permanently shadowed regions must plan over uncertain elevation maps under thermal limits that apply to the whole traverse rather than to single steps. We study a grid planner, ARTP-Lunar, that converts elevation uncertainty into a closed-form per-edge probability of exceeding a slope limit and enforces a path-level cold-exposure budget through label-setting A* over (cell, cold-bin) states, retaining an admissible Euclidean heuristic. Optional dust and water-ice terms are assumed illumination-based proxies. Parameters were frozen on separate validation maps. On 40 previously unseen synthetic maps, with five comparisons fixed in advance, safe-success rose from 0.72 (fixed k-sigma) and 0.70 (terrain-aware) to 0.90 (Holm-adjusted p < 0.001 for all five), and cold-budget violations among successful routes fell from 0.32 to 0.02 relative to the same planner without the budget. Removing the uncertainty handling or the budget erased the gain, but replacing the closed-form probability with a k-sigma indicator changed safety by no measurable amount (+0.03, 95% CI [-0.03, 0.08]) and lowered route energy by about 325 J. In an exploratory sweep the advantage narrowed at 0.2 m elevation error and vanished at 0.4 m. The dust and ice terms had no practical benefit. All terrain is synthetic and all dust, thermal, and ice models are unvalidated proxies, so the results characterise planner behaviour rather than lunar performance.
No takes yet. Share an insight, caveat, or question.
Fazil et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: