APGEN Scheduling: 15 Years of Experience in Planning Automation

Abstract

In this paper, we discuss the scheduling capability of APGEN (Activity Plan Generator), a multi-mission planning application that is part of the NASA AMMOS (Advanced MultiMission Operations System), and how APGEN scheduling evolved over its applications to specific Space Missions. Our analysis identifies two major reasons for the successful application of APGEN scheduling to real problems: an expressive DSL (Domain-Specific Language) for formulating scheduling algorithms, and a well-defined process for enlisting the help of auxiliary modeling tools in providing high-fidelity, system-level simulations of the combined spacecraft and ground support system. Nomenclature ACS= Attitude Control System AI = Artificial Intelligence AL = APGEN Language AMMOS = Advanced Multi-Mission Operations System APGEN = Activity Plan GENerator DI = Deep Impact DPT = Data Priority Table DSL = Domain-Specific Language DSN = Deep-Space Network FPT = Frame Priority Table INSIGHT = Interior Exploration using Seismic Investigations, Geodesy and Heat Transport MER = Mars Exploration Rover MRO = Mars Reconnaissance Orbiter MSL = Mars Science Laboratory PEF = Predicted Event File PEL = Power Equipment List S/C = Spacecraft SASF = Spacecraft Activity Sequence File SEQGEN = SEQuence GENerator TOL = Time-Ordered Listing

Europa Lander Concept: High fidelity system modeling informing flight system and concept of operations years before launch

“…The models are written into APGen Adaptation Files (.aaf), which are ingested by the program when it starts up, before activity instances begin being processed. For a more detailed discussion of the control flow during modeling, see Ref [10]. Each activity instance is also a tree in that activities can 'decompose' into child activities, allowing locally top-down generation of more detailed activities.…”

Towards a Cradle-to-Grave, Mission-Wide Simulation System

“…The central tool used to build and run simulations of the Europa Clipper Mission in its current stage is APGen, a modeling and simulation framework developed by NASA's Advanced Multi-Mission Operating System (AMMOS) organization [2], [3]. The focus of the original requirements on APGen was mission operations, and in particular the planning function within the Mission Operations System (MOS).…”

“…The basic scheduling algorithm used by the APGen modeling engine has been described in detail elsewhere [2], and we will therefore limit ourselves to a brief overview. The basic component of the APGen scheduling engine is a window finder which takes as input a "scheduling condition" and a minimum duration over which the scheduling condition is to be continuously true.…”

“…Table 4 below lists the specifics of the DSN model. 1 APGen models commands as activity types, but can create special output products (e. g. sequence files) specifically for activities that represent commands 2 The IMCE Ontology includes timelines but not the means to generate them (other than the equations of physics) 3 If the model is implemented externally, the modeling section relies on the API specified in the user-defined library to express the required behavior 4 An activity type that implements a goal should contain decomposition or modeling information that enforces the constraint imposed by the goal 5 Elaboration is in the process of being added to the IMCE Ontology Table 5 below highlights the components of the GNC subsystem. Although GNC implementation varies from one mission to the next, computational components of the model such as quaternion algebra and matrix manipulation are mission-independent, providing significant heritage and reuse opportunities.…”

Leveraging Multiple Artificial Intelligence Techniques to Improve the Responsiveness in Operations Planning: ASPEN for Orbital Express

“…ASPEN admits search in the infeasible space (in fact, threats and constraint violations are rolled up into a single generic entity called a conflict) allowing for faster response to off-nominal execution (Chien et al 2000;Chien et al 2005). In fact, ASPEN has been demonstrated for in situ rover technology (Castano et al 2007;Estlin et al 1999;Gaines et al 2008) P. Maldague, A. Ko, D. Page, and T. Starbird (Maldague et al 1997) have developed a schedule generation framework (APGEN) that automatically generates and validates plans used for commanding spacecraft but does not perform general planning and scheduling.…”