Project
AeroLoop
A simulation-only drone laboratory for diagnosing flight and landing behavior. Compare recorded experiments, inspect controller trade-offs and follow every failed acceptance gate.
Engineering reading paths
Autonomous systems under test
Follow simulated flight experiments from a failed landing to the next bounded comparison.
Start here
Key idea
Keep mission completion, paired comparisons and sustained recovery separate. These articles connect a recorded failure with the experiment and inspection tools used to understand it. Results remain simulation-only and preserve their original gates.
Curated connections
Systems
Curated connections
Engineering notes
Engineering update
AeroLoop compares landing response without adding flights
The repaired approach study records 12/18 fixed-gain and 15/18 scheduled-gain mission passes. A four-flight inspection lab keeps failed acceptance gates visible.
Engineering update
AeroLoop finds landing error before first contact
Two no-outage cases exceed the final-support position bound at contact. Retained analysis separates approach, contact and disarm while an opt-in decay candidate remains experimental.
Engineering update
AeroLoop separates flight diagnosis from a new flight result
A ten-tool workspace inspects 24 retained flights without changing their gates. Horizontal feedback helps mission completion, while paired tracking and recovery still expose failures.
Engineering update
AeroLoop finds the edge of its sensor-outage envelope
500 ms profiles recover in the retained baseline study. One-second cases are mixed and all two-second missions fail. Later prediction and landing-guard studies still report accepted: false.
Engineering update
AeroLoop adds rotor control, wind trials and contact landing
Explore measured rotor envelopes, five matched wind comparisons and a calm mission's liftoff, touchdown and motor-disarm events.
Engineering update
AeroLoop connects recorded control to replay
Compare three CPU response plots, scrub their recorded motion and inspect the separate saved-policy reproduction evidence.
Curated connections
Long-form writing
Article
Why 15 of 18 Drone Missions Passed but the Controller Still Failed Acceptance
AeroLoop's scheduled gains improve mission completion, while separate outage and recovery gates still fail. Retained flight data makes that distinction inspectable.
Article
The Drone Missed the Landing Before It Touched the Ground
AeroLoop's retained contact analysis finds two failed approaches already outside the final-support position bound at first contact. Small post-disarm changes point the next experiment upstream.
Article
How Long Can a Drone Controller Fly Blind? Measuring Recovery From Sensor Outages
AeroLoop's retained simulation study recovers from 500 ms missing captures under the tested conditions. One- and two-second gaps expose a failure boundary that prediction and a landing guard do not close.
Curated connections
Evidence
System map and source
AeroLoop: inspect the evidence
Explore the documented boundaries and pinned sources in this case study. Evidence reviewed 8 October 2026.