RELIABLE

Advanced robotics, smart sensors, and related control systems methodologies have the potential to be disruptive technologies that may generate significant societal benefits, but it may also produce serious consequences in case of failure, particularly in safety-critical applications. Safety is a critical requirement for a wide range of engineering systems, and engineering design of such systems is complex and involves many technical fields ranging from conceptual algorithmic level design with formal guaranties to hardware and software development, implementation, and operation. In this project, we target the methodological/conceptual side at the algorithmic/design level in the field of automation and control systems. The aim is to develop system methodological tools and algorithms that explicitly integrates in the formulation safety constraints and that are provably (mathematically) safe certified. To this end, a particular goal is to combine data driven approaches and machine learning techniques (known to be non-reliable, and/or extremely difficult to obtain formal guarantees) with recent optimization control based techniques capable of enforcing invariance in the context of control barrier functions (CBFs) and Control Lyapunov Functions (CLFs) in the presence of challenging restrictions and uncertainties. Another goal is to move from a single system to (possible large scale) safety critical networked systems involving multiple agents operating autonomously over networks in dynamic environments, where additional challenges arise due to the presence of a communication network. In this case, we will investigate the following crucial aspects: i) how to design control strategies to detect and/or isolate faulty agents and malicious attacks, ii) how to improve robustness under faulty agents and/or tempered input/output signals regarding controllability and observability properties of the overall system, and iii) how to improve resilience of consensus like processes, covering in this way a wide range of applications (distributed optimization, motion coordination tasks like flocking, leader following, among others). This project is mainly devoted to conceptual aspects, but to illustrate, demonstrate and also assure that it is driven by high-impact safety-critical application systems, RELIABLE will also focus on the following case studies: - Robotic vehicles in space, aerial and underwater scenarios: The scenarios considered here are mostly related to remote monitoring and exploration applications that require high performance control systems, and are critical in the sense that the consequences of failure can lead to an increase of operational costs, loss of resources (e.g., equipment), opportunities, and failure of the mission. An example to study is the development of path planners for hopping robots traversing rugged terrains of large celestial bodies (such as planets and moons) which has disconnected safe regions cluttered with forbidden zones. Other important topics include the development of localization, navigation and motion control strategies of single and multiple autonomous underwater vehicles and unmanned fixed-wing aerial vehicles with safety guarantees. Here, we plan to go beyond algorithm development, but to test them in digital twins (realistic simulators) and field experiments. - Mobile robotics in industry 4.0 scenarios: The aim is to contribute to the development of innovative manufacturing solutions with particular emphasis on robotic co-workers scenarios using mobile robotic vehicles. Key research points include the development of high performance active and reliable perception algorithms, reactive planning, navigation and control systems to enable mobile robots to operate autonomously in unstructured environments with effective human-robot collaboration with safety guarantees. This proposal brings together experts from the areas mentioned above involving a team of PhD students, pos-docs and several faculty members with the University of Porto and University of Coimbra. The merit of this research program is that it targets fundamental/conceptual research, but it is also well motivated by applications. The theoretical solutions envisioned will be strongly rooted in research work done by the team, where obtaining formal guarantees of safety, robustness, stability and performance is a key objective. At practical level, one key objective is to demonstrate and integrate some of the algorithms developed in software tools for command and control of single and multiple robotic systems, simulate and test within hardware in the loop, and validate through real field tests. https://home.isr.uc.pt/~rui/projects.html