NASA Logo, National Aeronautics and Space Administration



The goal of an ACM system is to increase uptime in the event of a potentially critical failure by carrying out system reconfigurations and obeying safety limits. In this initiative we intend to demonstrate the capability of an ACM system to carry out reconfigurations at different levels of decision-making hierarchy based on available prognostics information for the Remaining Useful Life (RUL) of the failing component(s). By incorporating the likely future conditions of the system into the ACM routine, it is possible to assess the likelihood of accomplishing a given set of objectives and, if necessary, change the objectives to avoid catastrophic failures. A typical ACM implementation usually utilizes a hierarchical architecture that covers low-level redundancy management, mid-level fault accommodation strategies, and high-level adaptive mission re-planning modules. For a successful deployment of such a system various practical requirements must be met such as:

  • Reliability, Availability, Maintainability and Durability (RAM-D)
  • Safety
  • Scalability
  • Cost effectiveness
  • Open system architecture with modular design
  • Ability to integrate prognostic information into reconfiguration plan, and
  • Mission Success


Ideally an ACM should be incorporated at the design phase of any system when most comprehensive information is available regarding structural and functional composition of a system. Alternatively, exhaustive analysis techniques like FMECA (Failure Modes, Effects and Criticality Analysis) can yield required information for already existing systems. The design philosophy revolves around optimizing various objectives, such as mission success, safety, performance, etc., within the constraints of available resources, time, relative severity of various components needing attention, etc. While many of these quantities are not directly measurable, it is not entirely straightforward to formulate a multi-objective optimization problem. In this project we have attempted to show that if these quantities can be transformed into cost models based on physical dynamics and heuristic information an optimization problem can be formulated.

Chart of ACM Design

Ingredients for an ACM design

Analytically, the objective of the ACM system is to optimize the utility of the vehicle with impaired capability to accomplish an assigned mission. The ACM system can be formulated as an optimization problem at two levels:

Formula for j of m

Formula for j of r


U is a cost function that quantifies the usefulness of the vehicle to accomplish its mission. U is a function of the available prognostic information (Pr), the system's closed-loop performance (Pe), and the mission objectives (M and Mcom).

Pe is a function of fault mode Fm, future prediction Pr, as well as any re-structuring/reconfiguration R applied to the system and current mission objective M.

Fm is a vector of indicators (0 or 1) that characterizes the fault modes detected on the spacecraft.

R is a vector of indicators that characterizes all restructuring applied to the system.

Mcom describes the mission assigned to the spacecraft.

M allows the fault-tolerant control architecture, specifically the mission adaptation and resource management components, to modify the parameters of the assigned mission and redistribute the available resources based on vehicle's current performance Pe.

At high level, mission adaptation and resource redistribution (M) allows the control architecture to pursue relaxed mission objectives in order to achieve greater vehicle usefulness U. At lower level, the objective is to optimize vehicle performance Pe while satisfying the mission constraints, through restructuring and reconfiguration R. Practically, the above optimization problems have to be solved while adhering to various constraints including system dynamics and resource limitations.

First Gov logo
NASA Logo -