IMPROVE - Nonlinear Control, Estimation and Fault-Detection Tools with Provably Guarantees for Mobile Robotic Systems

The aim of this project is to develop system theoretical tools and algorithms in framework of mobile robotic systems that explicitly integrate in the conceptual formulation not only the desired main task but also other key objectives (e.g., economic, performance, robustness, safety, system observability properties, communication behavior and interaction with other systems, etc.) in the presence of challenging restrictions and unstructured environments. The emphasis will be placed on the design of nonlinear and optimization based control and estimation procedures that are provably accurate by construction for single and multiple robotic systems including fault-detection and isolation strategies in order to obtain high performance robotic systems capable of meeting the end-user requirements. To assure that the research is driven by high-impact application areas, the project will focus on the following case studies: - Shop floor logistics and manipulation: The aim is to study and contribute to the development of innovative manufacturing solutions with particular emphasis on logistics and robotic co-workers scenarios using mobile manipulators. Key research points include the development of robust and high performance torque control strategies for robotic manipulators, 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. - Cooperation of air and marine autonomous robotic vehicles for ocean monitoring and sampling: The motivating scenario is the detection and tracking of some event/feature of particular interest in the ocean (e.g., oil spill pollution) where a network of heterogeneous robotic vehicles (air, surface, and underwater) that can interact autonomously with the environment and among themselves, works in cooperation to obtain measurements with adequate temporal and spatial resolutions. The same network can also adapt in real-time its behavior/geometrical configuration in response to environmental variables measured in-situ in order to improve performance and optimize the detection and measurement strategy. This proposal brings together experts from the areas mentioned above. The merit of this research program is that it targets fundamental research well motivated by applications. The theoretical solutions envisioned will be strongly rooted in research work done by the team. Obtaining formal proofs of robustness, stability, and performance of the control and estimation algorithms is a key objective. At practical level, one key objective is to demonstrate and integrate some of algorithms developed in the software tools for command and control of the robotic systems, simulate and test within hardware in the loop, and validate through field tests.