The Sardonyx mission will demonstrate a set of technologies for agile, close-proximity (0-10 meters) operations: sensing, 6DOF movement, mapping, navigation algorithms, attachment, and detachment.
This will be accomplished by releasing an inflatable target object from the host spacecraft, then acquiring, approaching to within 10 meters, closing and finally making adhesive contact using the electric Thin Attachment Pad (eTAP).
The spacecraft will then detach and the experiment will be repeated for different closing dynamics and then with additional deployed inflatable objects with different surface materials. Long-term effects of the environment on eTAP will also be measured.
Novel sensing technologies to be tested include ranging, capacitance and touch. The propulsion and navigation systems will emphasize the ability to alter position in any direction with minimal change to attitude relative to the target.
Mission Goals
Primary Goals:
eTAP: Quantitatively and qualitatively evaluate the performance in space environment. Essential parameters include closing dynamics (angles/speeds), post-contact dynamics, charging/plasma field, vacuum (including vacuum welding), measured adhesion force, power requirements. Long-term variability of all those parameters
Education: Hands-on, systems-engineering focused undergraduate education. Students in the critical path. Emphasis for NS-12: the hands-on part (prototyping early and often).
Secondary Goals - These goals are accomplished in the process of satisfying the primary goals
Technologies for agile, close-proximity operations (0-10 meters): sensing, 6DOF propulsion, mapping, navigation algorithms
Learned autonomy: evaluate the performance of a learned 6DOF control system vs. more traditional methods
Calibrated target(s): inflatable targets of different (but known) materials will provide opportunities for other agencies to observe/track
Mission Management
Program Manager: Brody Schram
Chief Engineer: Sophia Weaver
Experiment Overview - Tag
One Experiment, Two Controllers
Deterministic Controller: Bandit
Reinforcement-Learned Controller: Alice
Test the reliability of eTAP in space using each controller
Part 1:
Release Objective
Use Deterministic Controller navigation
Downlink experiment data to train Alice
Part 2:
Release Objective
Use Alice navigation
Downlink Alice and experiment data to train new Alice model