Simone Fratini

Mexar2: AI Solves Mission Planner Problems

Deep-space missions carry an ever larger set of different and complementary onboard payloads. Each payload generates data, and synthesizing it for optimized downlinking is one way to reduce the ratio of mission costs to science return. This is the main role of the Mars-Express scheduling architecture (Mexar2), an Al-based tool in daily use on the Mars-Express mission since February 2005. Mexar2 supports space mission planners continuously as they plan data downlinks from the spacecraft to Earth. The tool lets planners work at a higher abstraction level while it performs low-level, often-repetitive tasks. It also helps them produce a plan rapidly, explore alternative solutions, and choose the most robust plan for execution. Additionally, planners can analyze any problems over multiple days and identify payload overcommitments that cause resource bottlenecks and increase the risk of data losses. Mexar2 has significantly increased the data return over the whole Mars-Express mission duration. Its effectively become a work companion for mission planners at the European Space Agencys European Space Operations Center (ESOC) in Darmstadt, Germany.

MrSPOCK—STEPS IN DEVELOPING AN END-TO-END SPACE APPLICATION

This article elaborates around a recent effort to build a planning system that helps human mission planning in a space mission. Specifically, the article describes the steps that brought us to develop MrSPOCK, the MARS EXPRESS Science Plan Opportunities Coordination Kit, a tool for supporting long‐term planning in the MARS EXPRESS mission of the European Space Agency. In showing our effort for creating MrSPOCK, we will underscore the key ingredients for developing complete applications and how they are connected to a stable line of research on planning and scheduling with timelines.

Validation and verification issues in a timeline-based planning system

To foster effective use of artificial intelligence planning and scheduling (PS moreover, they employ resolution processes designed to optimize the solution with respect to non-trivial evaluation functions. Knowledge engineering environments aim at simplifying direct access to the technology for people other than the original system designers, while the integration of validation and verification (V&V) capabilities in such environments may potentially enhance the users’ trust in the technology. Somehow, V&V techniques may represent a complementary technology, with respect to P&S, that contributes to developing richer software environments to synthesize a new generation of robust problem-solving applications. The integration of V&V and P&S techniques in a knowledge engineering environment is the topic of this paper. In particular, it analyzes the use of state-of-the-art V&V technology to support knowledge engineering for a timeline-based planning system called MrSPOCK. The paper presents the application domain for which the automated solver has been developed, introduces the timeline-based planning ideas, and then describes the different possibilities to apply V&V to planning. Hence, it continues by describing the step of adding V&V functionalities around the specialized planner, MrSPOCK. New functionalities have been added to perform both model validation and plan verification. Lastly, a specific section describes the benefits as well as the performance of such functionalities.

TGA-based controllers for flexible plan execution

Plans synthesized by Temporal Planning and Scheduling systems may be temporally flexible hence they identify an envelope of possible solutions. Such flexibility can be exploited by an executive systems for robust on-line execution. Recent works have addressed aspects of plan execution using a quite general approach grounded on formal modeling and formal methods. The present work extends such an approach by presenting the formal synthesis of a plan controller associated to a flexible temporal plan. In particular, the controller synthesis exploits Timed Game Automata (TGA) for formal modeling and UPPAAL-TIGA as a model checker. After presenting a formal extension, the paper introduces a detailed experimental analysis on a real-world case study that demonstrates the viability of the approach. In particular, it is shown how the controller synthesis overhead is compatible with the performance expected from a short-horizon planner.

Flexible timeline-based plan verification

Flexible temporal planning is a general technique that has demonstrated wide applications possibilities in heterogeneous domains. A key problem for widening applicability of these techniques is the robust connection between plan generation and execution. This paper describes how a model-checking verification tool, based on UPPAAL-TIGA, is suitable for verifying flexible temporal plans. Moreover, we further investigate a particular perspective, i.e., the one of verifying dynamic controllability before actual plan execution.

Flexible Plan Verification: Feasibility Results

Timeline-based planning techniques have demonstrated wide application possibilities in heterogeneous real world domains. For a wider diffusion of this technology, a more thorough investigation of the connections with formal methods is needed. This paper is part of a research program aimed at studying the interconnections between timeline-based planning and standard techniques for formal validation and verification (V&V). In this line, an open issue consists of studying the link between plan generation and plan execution from the particular perspective of verifying temporal plans before their actual execution. The present work addresses the problem of verifying flexible temporal plans, i.e., those plans usually produced by least-commitment temporal planners. Such plans only impose minimal temporal constraints among the planned activities, hence are able to adapt to on-line environmental changes by trading some of the retained flexibility. This work shows how a model-checking verification tool based on Timed Game Automata (TGA) can be used to verify such plans. In particular, flexible plan verification is cast as a model-checking problem on those automata. The effectiveness of the proposed approach is shown by presenting a detailed experimental analysis with a real world domain which is used as a flexible benchmark.

The First Competition on Knowledge Engineering for Planning and Scheduling

We report on the staging of the first competition on knowledge engineering for AI planning and scheduling systems, held in Monterey, California, in colocation with the ICAPS 2005 conference. The background and motivation is discussed, together with the relationship of this new competition with the current international planning competition. We report on the new competitions format, its outcome, and the benefits we hope it will bring to the research area.

The RAXEM Tool on Mars Express - Uplink Planning Optimisation and Scheduling Using AI Constraint Resolution

This paper describes an AI-based application, referred to as RAXEM, which has been developed to support the Flight Control Team of Mars Express in the daily planning task of uplinking telecommands to the spacecraft. The tool is part of the mission improvement activities within the team to move from a manually-oriented to a more tool-assisted and automated approach. As such it not only provides an interactive process for solving the uplink problem but also contributes to reducing the workload of the engineers. A first AI based system, MEXAR2, has been in operational use at ESOC since the beginning of 2005. It was developed to perform the planning of the downlink of the telemetry data. The tool is based on AI constraint resolution techniques and was the first of its kind to be in operational use at ESOC. It has gained wide-spread acclaim in the mission planning and scheduling community. The operational experience and success of MEXAR2 and the similarity of the problem subsequently led to the definition for an application to generate a detailed uplink plan and schedule. Mars Express is a mission that is not operated in real-time but all commands are loaded into the Master Timeline Buffer (MTL) on the spacecraft from where they are executed at the given time-tag. The uplink plan depends on a number of constraints, i.e. the loading profile of the MTL, the available uplink opportunities (or uplink windows) depending on the ground station allocation and spacecraft orientation to Earth (which is determined by the science requirements), the occultations by Mars and moons and the availability of power that determines the activation periods of the spacecraft transmitter. The paper presents the operations and mission planning constraints that influenced the requirements for the RAXEM planning tool. It further focuses on the tool and its operational usage, the iterative-prototyping approach from requirements definition to software development and operational validation, and the optimization and benefits for the Flight Control Team compared to the manual approach adopted up to the introduction of the tool. The paper will also address the issues raised by introducing so-called “clever” tools to help human operators and compare the consequences depending on the target user group, the planners (as for the downlink planning tool, MEXAR2) or the spacecraft engineers (as for RAXEM). Furthermore, another implementation option for the uplink problem based on Operations Research algorithms developed as a student project will be included.

Rapid Prototyping of Planning & Scheduling Tools

The Advanced Planning and Scheduling Initiative, or APSI, is an ESA programme to design and implement an Artificial Intelligence (AI) software infrastructure for planning and scheduling that can generically support different types and classes of space mission operations. The goal of the APSI is twofold: (1)~creating a software framework to improve the cost-effectiveness and flexibility of mission planning support tool development; (2)~bridging the gap between AI planning and scheduling technology and the world of space mission planning. A key aspect of the success of this project is the presence of a flexible timeline representation module that allows to exploit alternatives in the modeling of mission features. This paper shows an example of such a flexibility by using a real problem in the space realm - the HERSCHEL Science Long Term Planning process.

A Planning and Scheduling Service for the ULISSE Platform

ULISSE is an EU project that aims at data valorization around the ISS experiments. The ULISSE software platform is endowed with a number of additional services to improve both data production and data analysis. This paper describes the Planning and Scheduling Service (PSS), a module developed to support functions of data production around the ISS activities and integrated in the ULISSE platform. In particular, the PSS is a software application developed within the Timeline Representation Framework and relies on a combination of different P&S algorithms in a loosely coupled way. Its current use to support Increment Planning activities for the Fluid Science Laboratory facility is shown and fully analyzed from design to application service delivery.