An intelligent satellite must do nothing well
An intelligent satellite must do nothing well
This research paper challenges the conventional view that algorithm selection is the primary factor in autonomous space object inspection. The authors argue that orbital dynamics allow an inspector spacecraft to observe targets at no fuel cost if it utilizes natural trajectories, and that fuel efficiency is critically dependent on three design choices: allowing the agent to drift, recording the drifted state, and maintaining a long enough planning horizon. The study finds that removing any of these factors significantly increases fuel consumption, while model-based agents that ignore fuel constraints spend several times more fuel than necessary. Additionally, the paper emphasizes that no learned policy is inherently safe and demonstrates the use of a control barrier function filter to eliminate keep-out-zone violations during evaluation.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- Control
- Reported autonomousOpen source S2
Reported numbers
evaluation episodes
2,000
Reported trials
View original evidence
ough what it can represent: a model-based agent whose learned world model ignores fuel spends several times more of it. Finally, no learned policy is safe by construction, and a control barrier function filter beneath the decision layer removed every keep-out-zone violation in two thousand evaluation episodes.
Open source S4fuel efficiency factors
3
View original evidence
observe the target at no fuel cost. On a controlled ladder in which one design axis is varied at a time, fuel efficiency turns out to be set by three things: whether the agent is allowed to drift, whether its state records where drifting has taken it, and whether its planning horizon is long enough for the benefit to appear. Removing any one of them doubles the fuel bill or more. The algorithm still matters, but thr
Open source S3
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
arXiv:2611.02922v1 Announce Type: new Abstract: Autonomous inspection of a resident space object is usually posed as a contest between classical next-best-view planning and reinforcement learning. We argue that the choice of algorithm is the least informative of the design axes. An inspector spacecraft has no rest state: when it stops thrusting, orbital dynamics carry it along a natural trajectory that can be used to
Open source S2
ough what it can represent: a model-based agent whose learned world model ignores fuel spends several times more of it. Finally, no learned policy is safe by construction, and a control barrier function filter beneath the decision layer removed every keep-out-zone violation in two thousand evaluation episodes.
Open source S4
observe the target at no fuel cost. On a controlled ladder in which one design axis is varied at a time, fuel efficiency turns out to be set by three things: whether the agent is allowed to drift, whether its state records where drifting has taken it, and whether its planning horizon is long enough for the benefit to appear. Removing any one of them doubles the fuel bill or more. The algorithm still matters, but thr
Open source S3
Source：arXiv Robotics — research abstracts · arxiv.org