OUTLAST

Background

Ingenuity’s operational procedure included a ground assessment cycle after each flight to manually validate the system’s health and flight-readiness, and globally localize the helicopter on Mars. Although effective at ensuring Ingenuity’s health, the necessity of performing a ground-in-the-loop cycle post-flight necessarily precluded the possibility of multi-flight operation, i.e. the autonomous execution of multiple flights by the system without ground involvement. Autonomous multi-flight operation would enable greater science productivity in future missions not only by increasing the operational efficiency of the system, but also by enabling more intricate maneuvers and flight plans, such as multiple hops to reach challenging but scientifically interesting terrains and locations, as well as spontaneous decision-making in the presence of unexpected but scientifically interesting phenomena. To address this, we developed the software OUTLAST (Onboard aUtomaTed heaLth ASsessmenT).

OUTLAST is software capable of being run onboard that is able to perform system health assessment and diagnostics in a fully automated capacity. OUTLAST does this by compiling flight data after each flight (including telemetry and event logs) into a dataset on which a health classification model is run. The result of the classification model is a Go/NoGo decision and a compiled diagnostics file to be downlinked to ground. OUTLAST currently supports decision tree classification, which has several benefits including low complexity, decision-interpretability, and the ability to encode both manually-specified decision variables in addition to those learned from data.

Problem

The basic problem is to automatedly diagnose the system’s health and flight-readiness after a flight is executed in order to determine if the system is ready to safely perform a subsequent flight, without invoking a full ground cycle. We approach this as a classification problem on the system’s flight-readiness based on the entirety of the data collected during the prior flight.

Impact

The potential for the aerial mobility of rotorcraft on Mars demonstrated by Ingenuity fundamentally alters and expands the space of science missions that can be investigated in the future. Aerial platforms can cover multiple dimensions of exploration (as opposed to rovers which are constrained to flat, even terrain), expanding the range potential of exploration systems, enabling access to scientifically interesting but hazardous or otherwise unreachable terrains, and providing a unique opportunity for in-situ atmospheric science at different elevations. These capabilities are opportunities to address high-priority science questions identified in the Mars science community, demonstrating the vast potential of these systems in the future. We expect that future missions, like Mars Science Helicopter [1], will capitalize on this potential. However, the restriction to single-flight operations will severely limit the scientific productivity of such future missions. OUTLAST will help to remove this restriction, helping to maximize the efficiency and scientific productivity of future missions.

Status

OUTLAST was flown onboard Ingenuity on January 5, 2024 just prior to mission end.

References

[1] Tzanetos T, Bapst J, Kubiak G, Tosi LP, Sirlin S, Brockers R, Delaune J, Grip HF, Matthies L, Balaram J, Withrow-Maser S. Future of Mars rotorcraft-Mars science helicopter. In 2022 IEEE Aerospace Conference (AERO) 2022 Mar 5 (pp. 1-16). IEEE.

Publications

Team

Gerik Kubiak, Jet Propulsion Laboratory, California Institute of Technology (Project Lead)
Dr. Connor Basich, Jet Propulsion Laboratory, California Institute of Technology (Lead Developer)
Dr. Juan Delfa, Jet Propulsion Laboratory, California Institute of Technology (Advisor)
Dr. Alberto Candela-Garza, Jet Propulsion Laboratory, California Institute of Technology (Advisor)
Dr. Steve Chien, Jet Propulsion Laboratory, California Institute of Technology (Advisor)

Sponsors

Office of Technology Infusion and Strategy (OTIS)