During rover operations, when a fault occurs, it may require changes to low-level control, system usage, or the overall mission plan. In this paper, we described an integrated architecture for autonomous decision making, using diagnostic and prog- nostic information. The architecture is modular and allows different technologies for the LLC, DX, PX and DM modules. We demonstrated the approach on several scenarios in simulation.
n future work, we plan to further develop each module of the integrated architecture, and evaluate which algorithms perform best in each module of the architecture, according to a set of criteria which will include computation time and robustness to input noise, among others. PX and DM algorithms will be extended to better accommodate uncertainty in system modeling and prognostic predictions, allow handling of multiple degrading components, and incorporate system parameter/constraint optimization. The hardware testbed will also be further developed in order to demonstrate and validate our architecture in complex real-world scenarios.
Nothing to see here at the moment. Check back later.