Towards an Evaluation Framework for Autonomous Systems

Andrea Bombarda, Silvia Bonfanti, Martina De Sanctis, Angelo Gargantini, Patrizio Pelliccione, Elvinia Riccobene, Patrizia Scandurra

Abstract: Despite the active and proficuous research in autonomous and self-adaptive systems of the last years, an evaluation framework to assess abilities related to adaption and to provide guidance to make a system smarter is still missing. In this paper, we perform the first steps towards an evaluation framework for autonomous systems to (i) make an assessment of a system from the perspective of its capacity to adapt and learn over time to handle new and unexpected conditions, (ii) explore and identify the possible pathways of improvement of the smart abilities (e.g., decisional autonomy, adaptability, perception, interaction, etc.) of a system, (iii) make a re-assessment when the improvement has been performed.