Rover-Mounted vs Astronaut-Carried EchoQuilt Configurations
The Platform Question a Cave Concept Has to Answer
A JPL concept team planning a lunar lava tube mission at NIAC Phase II typically hits a fork in the decision tree: commit to a rover-carried sensor suite (distributed, autonomous, persistent) or a crew-carried one (high dexterity, short duration, curated placements). The answer has historically been "rover" for cave missions because human presence inside a subsurface void is expensive and rare, and the science return per crew-hour has been hard to justify against the alternative of multiple sols of robotic operation. That logic is changing as Artemis-era surface architecture treats EVAs as a sustained capability rather than an Apollo-style burst, and concept teams now have to answer the platform question with a more nuanced framework than the historical default. JPL's CADRE mission overview describes carry-on-bag-sized rovers mapping lunar subsurface in 3D, and CMU's Robotics Institute cave payload work keeps advancing rover autonomy for pit and lava tube entries.
Meanwhile, JPL's Axel rovers address extreme-terrain descent with meters of tether, and the LunarLeaper mission concept proposes a legged hopper under 15 kg as an alternative. On the crew side, the NASA xEVA ConOps documents the baseline for astronaut-carried operations, and the Artemis program framework lays out the period during which astronaut-carried payloads will coexist with rover platforms at the same lunar destinations.
The decision is rarely binary at concept maturity. NIAC and MatISSE proposals that survive Phase II review usually carry a primary platform plus a clear story for how the secondary platform can be added without re-baselining the science requirements. Concept teams that have already worked through PANGAEA-X exercises and crewed EVA traverses tend to converge on mixed configurations more quickly than teams that have only run rover-only or EVA-only field campaigns, because the integration friction at field campaigns surfaces early when both platforms are present, and the lessons translate directly into flight design choices. EchoQuilt was built with this trajectory in mind, and its sensor head is identical across platforms so that early concept choices do not foreclose later options.
How the Quilt Changes With Platform
EchoQuilt produces a quilt under both configurations, but the patch rate, the patch size, and the uncertainty profile all shift with platform. A CADRE-class rover carries a distributed array (typically 3-4 microphone nodes plus an onboard IMU) and produces continuous stitching during traverse. Patches are small and numerous, the quilt densifies over hours, and the traverse trajectory is chosen partly to maximize information gain. Power draw stays under 2 W continuous because the stitching runs on a dedicated DSP rather than the rover's main compute, which lets the rover's mobility planner stay within its own thermal and power envelope.
The CADRE platform's swarm dynamics open a possibility that single-platform deployments cannot exploit. When multiple rovers operate inside the same tube, EchoQuilt can fuse their independent quilts into a single shared quilt at a frequency higher than the relay window allows, because inter-rover BLE or UWB links carry several megabits per second over short ranges. This means a four-rover CADRE deployment effectively quadruples the coverage rate per sol, and the per-rover energy budget is freed up to spend on mobility rather than redundant self-localization across the rover formation.
An astronaut-carried EchoQuilt trades continuous stitching for targeted placements. The crew drops a small number of passive nodes at geometric transitions (skylight, first bend, chamber opening) and a single wearable inertial stream captures the traverse motion. Patches are fewer but placed at the most informative locations, and post-EVA reconstruction produces a quilt with larger individual patches at lower uncertainty. The two platforms are complementary: a CADRE-class rover fills the interior quilt densely, while astronaut placements anchor the low-uncertainty nodes that keep the quilt geometrically correct over multi-sortie campaigns.
A companion post on rover replanning inside new voids uses the quilt's uncertainty map to guide trajectory updates mid-traverse without waiting for ground intervention, which is the operational payoff of choosing a platform that supports continuous patch generation.
A subtler benefit of the platform-coupling design is that EchoQuilt does not lock a concept into a single platform decision irrevocably. A team that starts with a CADRE-class rover concept can later add astronaut placements without re-engineering the patch inference model, and a team that starts with an astronaut-only deployment can later add rover-based stitching without changing the quilt format. This optionality matters because Artemis-era surface architecture is itself uncertain, and a payload that imposes platform commitments early forecloses option value the broader program needs to keep open. This optionality matters at NIAC Phase II reviews, where the decision tree is rarely settled and reviewers favor concepts that retain flexibility through the next concept maturity gate.
The shared quilt format also lets two independent concept teams (one rover-led, one EVA-led) compare results on the same analog site without converting between bespoke map representations, which has shortened post-campaign analysis cycles in our experience.
Advanced Tactics for Mixed-Platform Analog Campaigns
Three practices improve the quilt when a campaign runs both platforms in one traverse. First, start the EVA placements at the boundaries between rover-reachable and rover-unreachable regions. A crew member places nodes where the rover cannot go (tight constrictions, vertical drops), which keeps the quilt connected across geometry the rover alone would have left as gaps. PANGAEA's Corona lava tube exercises have already validated this split in practice, and the quilt-coverage gains are substantial.
Second, synchronize the time base across platforms at a single known location. A fiducial ping at a mutually reachable point lets the pipeline align the rover-originated patches with the EVA-placed patches to sub-millisecond precision. Without this synchronization, the two quilt layers drift relative to each other, and the combined quilt loses the advantage of both platforms. EchoQuilt provides a built-in sync procedure that takes less than five minutes of crew time.
Third, plan the rover trajectory using the EVA-placed low-uncertainty nodes as prior anchors. A rover that knows a handful of high-confidence waypoints can navigate the interior more efficiently and spend more of its energy budget on coverage than on self-localization. For concepts that target TRL 6 via an analog campaign, this mixed-platform approach has produced higher per-sol coverage than either platform alone in every field campaign we have tracked.
A fourth tactic that has paid off across PANGAEA-X and BASALT follow-on campaigns is to treat the EVA-placed nodes as long-duration mission anchors that outlive any single rover sortie, drawing on the same logic cave diving teams use when weighing backmount configuration tradeoffs versus sidemount for persistent sensor placement. A node placed during week one of an analog campaign continues to capture passive acoustic data through every subsequent rover traverse, which means the rover does not have to re-survey already-anchored geometry on each visit. The cumulative quilt grows monotonically across rover sorties, and the late-campaign quilt is dense enough to support detailed traverse planning without consuming additional rover time. Reviewers preparing a multi-mission lunar concept respond well to seeing this asymmetry exploited explicitly, because it converts each EVA from a one-shot deployment into a persistent infrastructure investment that pays dividends across the rest of the mission lifecycle.
CTA
EchoQuilt is configuring reference platforms for JPL CADRE teams, CMU cave autonomy groups, NASA SubT alumni, and Artemis EVA planners who need a cartography payload that runs under both rover and astronaut configurations without forcing an early platform commitment. We ship the rover-mounted sensor head, the EVA node kit, the synchronization procedure, and the swarm fusion module for multi-rover CADRE deployments as one package. Join the Waitlist for Planetary Analog Researchers to book a reference configuration before your next analog campaign so the platform decision can be made with measured data rather than vendor projections. Teams with confirmed mixed-platform traverses in 2026, or those preparing PANGAEA-X follow-on field exercises, get priority seats and direct integration support from our analog field engineers.