Autonomous, self-adaptive, and smart systems
Autonomous and smart systems are more and more required to cooperate and collaborate each other and with human actors, and, in general, they need to operate in environments shared with humans.
This calls for innovative software engineering techniques. Specific techniques in which we focus are:
testing techniques, e.g., self-adaptive testing in the field, testing approaches for
BSNs, and approaches to test autonomous vehicles anomaly detection
approaches and
evaluation framework for autonomous systems.
We also studied, through empirical industrial studies, the motivations for applying self-adaptation in practice, the problems solved using self-adaptation and how these problems are solved, and the difficulties and risks that
industry faces in adopting self-adaptation.
In this digital society, humans feel unprotected and unsafe, constrained by technology; values like safety, fairness, ethics and privacy are at risk. There is the need of new methodology and tools for
engineering digital systems for humanity and trustworthy and trustable systems. From some years, we investigate techniques and methodologies aiming at better balancing of forces:
we will no longer need to accept the opaque behaviour of the software technologies, rather, we will provide effective means to mold these technologies into shapes that better reflect our human values
(see the Exosoul project).
Robotic Software engineering
We are pioneers in establishing the discipline of robotic software engineering for service robots. We investigated safety aspects and the readiness of using robots in real life scenarios, software variability in service robotics, runtime verification and field-based testing for ROS-based robotic systems, architectures for the coordination of heterogeneous robots, Model-Driven Engineering (MDE) for service robotics, synthesis of controllers of a planners under uncertainty.
We also offered solutions for democratizing robots in everyday-life scenarios. We offer means to make systems programmable and usable by domain experts not necessarily experts in computer science or robotics. We identified and formulated mission specification patterns and mission specification patterns for quantitative properties to enable the specification of mission in structured English and, then, an automatic translation in temporal logic. We also contributed a Domain Specific Language (DSL) for mission specification, called Promise.
Architecting complex system
We have consolidated experience in languages for providing architecture descriptions. We have consolidated competencies in architecture frameworks (ISO/IEC/IEEE 42010:2011) and reference architectures. We investigated the definition of an architecture framework for Volvo Cars and the definition of a viewpoint for cars as constituents of a System of Systems (SoS), as well as a compositional approach to creating architecture frameworks with an application to distributed AI systems.
Concerning ML and AI systems, we studied architectural challenges, best practices, and design decisions in architecting ML-enabled systems.
We investigated the role of architectures when companies are in the transition towards the adoption of agile methodologies, and we introduced the concept of
boundary objects and the use of boundary objects in agile systems engineering.
Recently, we introduced the concept of Architecture as Code (AaC), intended as an approach where software architecture is continuously defined, managed, and evolved through a machine-readable, version-controlled code-base.
