Space systems

The PST has matured experience and is particularly active in the field of AI Planning and Scheduling techniques for space applications. Space represents a particularly suitable applicability domain for AI techniques. At present time, the need of automatic tools that aim at coadiuvating the human operators in the complex task of planning the phases of a space mission in a robust and consistent manner, as well as introducing a certain extent of onboard autonomy [i.e. automatic post-contingency rescheduling], is strongly felt. In the form of joint research projects with the leading space agencies, our research has produced a variety of applications which make use of these techniques in the context of space missions. The PST has also been active in the European Network of Excellence in AI Planning (PLANET) as the Aerospace Applications Technical Coordination Unit (AA TCU).

When it comes to the application of AI techniques to space missions, a critical issue is the development of effective means of interaction between problem solving tools and their users. For this reason, although our research activity is concerned with the implementation of AI Planning and Scheduling systems (see the Research and Software areas on the side menu), it necessarily follows a dual guideline:

  • on one hand, the development of interactive support tools for decision-making and solution of complex problems, where the tasks involved would result too long or too hard for a human to complete successfully without a contribution from an automated and "intelligent" system;
  • on the other hand, the development of interactive support tools for the management of simpler situations, where the tasks involved would result too tedious, repetitive or time-consuming for a human to carry out unassisted.

    In both cases, the "appropriate" style of interaction would depend on the user perceiving the AI system as an autonomous and separate entity from "self", with a "personality" of its own: for that reason, our work takes into account specifications and solutions from the domains of agent-based systems, artificial personalities and human-computer interfaces.