Modeling the Operation of Satellite Payloads for On-Board, Goal-Based Planning

摘要

The concept of goal-based operations is being explored at the Brazilian National Institute for Space Research (INPE), aiming its future missions, through the development of on-board planning software. After a first prototype for an on-board replanner created for a specific case study, we've started to follow a new path, developing on-board software to be reusable across different missions and aligning our work with the European standards currently in use by INPE. This work culminated in the Goal-based Enabling Software Architecture (GOESA), a multi-mission software capable of performing on-board, goal-based planning. The core component of GOESA is the Internal State Inference Service (ISIS), an ECSS-like on-board service that contains a model of the satellite's payload. ISIS provides features such as states inference, resources profiling, causal relationships and constraints management, which can be consumed by the ground segment and on-board applications. ISIS also manages the interface between the model and the real equipment, and lets the operations personnel to update its model in flight. GOESA receives goals (such as "get an image" at a given time) and asks its on-board planner, called LetMeDo, to query the ISIS model in order to explore what-if scenarios, trying to achieve them. The goals can be generated by the ground segment or another flight application. The output of the planning process can be a number of changes on the current operations plan, or even an entirely new plan. An ISIS model is concise, easy to understand by the engineering team and operations personnel, and light at a runtime perspective - after all, it shall run in space-qualified (and slower) processors. Although light, the modeling language comprises mechanisms that allow a behavioral description close to the real equipment operation. Up to now, we have exercised the on-board planning for two different Brazilian missions: EQUARS, a scientific mission to study the upper atmosphere, and the remote sensing satellite Amazonia-1. This paper describes the modeling and results gotten, including performance data, from the studies performed for the Amazonia-1 mission case study.