S3P-2019 Rationale

The 2019 IEEE SPS / EURASIP / ISIF Summer School on Signal Processing (S3P‐2019) / Signal Processing for Autonomous Systems (SP-AS) is the seventh edition of the technically co-sponsored by IEEE Signal Processing Society (SPS) via the Seasonal Schools in Signal Processing (S3P) initiative and by the SPS Italy Chapter. This Summer School can be seen also as the first Summer School related with the IEEE SPS Autonomous Systems Initiative.

The ASI initiative is a volunteer activity from several members of the IEEE SPS that is trying to highlight the signal processing related aspects of autonomous systems.

Over the last decade, researchers have been proposing and investigating computing systems with advanced levels of autonomy in order to manage the ever-increasing requirements in complexity. An autonomous system is an artificial system able to perform a certain number of tasks with a high degree of autonomy. Signal Processing plays a key role within autonomous systems by providing theories and techniques to design algorithms used for perception, control and learning within an autonomous system. Several models highlight this strict relationship and provide frameworks under which multifunctional SP methods can efficiently cooperate to allow an autonomous system to realize its tasks. For example, Cognitive Dynamic Systems (CDS) are a class of such models that can be well suited to tackle these challenges. CDS aim at building up rules of behavior over time through learning from continuous experiential interactions with the environment. By exploiting these rules, that can be designed starting from SP advanced techniques, CDS can deal with environmental dynamics and uncertainties that an autonomous system has to face. The automation of tasks can be leveraged where complex multilevel perception-action cycles are present as in infrastructure active surveillance, autonomous driving, unmanned aerial vehicles, cognitive radio, traffic management, robot-mediated industrial and smart home applications. Self-awareness and adaptability to internal and external non-stationary conditions are emerging requirements that make it necessary to autonomous systems to require setting of models using SP based machine learning techniques.

Many real-world systems, despite being initialized with offline learned models, frequently experience non-stationary conditions (i.e., unknown situations) due to uncertain and time changing interactions with the environment and users (including symbiotic human agents), failures or structural changes. Such situations can require an autonomous system to be provided of incremental and unsupervised learning capabilities, that again can be optimized using advanced SP techniques, to enable autonomous systems to cope with changing environments and operative conditions whether this is on land, underwater, in the air, underground, or in space.

This Summer School SP-AS is focused on providing an updated state of the art over most advanced signal processing theories and techniques that are relevant for developing autonomous systems. Lectures will be focused on novel signal processing algorithms and technologies for autonomous systems but also on in-depth reviewing of state-of-the-art autonomous systems

The main goal of the autonomous systems related research is to study and develop innovative concepts, algorithms, services based on more advanced signal processing techniques for autonomous systems along multiple fundamental and practical dimensions.

During this Summer School, lectures will be organized having in mind these different research axes and well-known high impact scientists will be invited for describing and analyzing autonomous systems concepts and novel approaches.

First, the researchers have to investigate and develop concepts for integrating multilevel Signal Processing techniques for fusion of heterogeneous sensors signals in an autonomous system and to study how to represent, maintain and exploit knowledge about internal states and environmental interactions with incremental and unsupervised learning techniques.

Second, to identify which signal processing techniques can be used in environments where humans and autonomous systems strictly interact in order to achieve common goals. Bio-inspired techniques will be also described that are used for analysis of natural autonomous systems as humans to address possible developments in the field of artificial autonomous systems. Collaborative symbiotic systems have special robustness and safety requirements and require a human-focused, task-dependent design of behaviors as well as timely and intuitive communication mechanisms.

Third, to investigate how signal processing can provide solutions in distributed networks of autonomous systems and how to generalize signal processing based learning and decision making/control methods from an individual node to a network/swarm of nodes.

Short hands-on tutorial guiding students through the process of building signal processing based autonomous systems. Summer school attendants will have the opportunity to learn and study innovative algorithms and systems for artificial interaction and cognition, to apply these concepts for building real working autonomous systems components and to cooperate with other students for designing self/aware related modules. The IEEE SPS Italy Chapter is also supporting in view of evaluating the setup of an annual Summer School event to be offered in appropriate Italian locations. A 2019 IEEE SPS Distinguished Lecturer will be invited to join this day event included in the Summer School as soon as the updated list will be available.