The project will focus on the use of AI techniques and of autonomous and cooperative robot swarms (ACRSs) for exploration & data mapping (EDM). In EDM task scenarios, robots act as mobile active sensors that cooperatively gather and fuse geospatial data for attributes of interest with the goal of constructing a map of attributes’ values. During a data mapping mission, which is usually time-constrained, the core challenge is how to plan and coordinate robots’ data sampling actions and navigation paths such that the amount of collectively gathered information is maximized, resulting in an accurate map estimator. Planning and coordination are expected to be distributed and to happen adaptively, in closed-loop with mission enrollment. The project will focus on map building in scenarios of environmental monitoring in marine ecosystems and of indoor surveillance. Examples of attributes of interest in these scenarios can include, water quality, distribution of 2ora and fauna, presence of intruders. Various approaches exist in the scienti1c literature for tackling EDM tasks with multi-robot systems but major challenges related to scalability, robustness, and adaptivity of the planning and control approaches are still present, limiting the actual deployment in the realworld. The research in the project will precisely tackle these challenges, aiming to develop novel and eZcient distributed solutions for planning, control, and communications in ACRSs. More speci1cally, the project will extend current work of the applicant by adding AI data-driven components to boost generality, robustness, and adaptivity of the solutions, and by including human supervisory control in the loop of operations of the swarm for implementing a shared control that can boost reliability and eZciency. The AI components will be based on a design 1rst using ergodic theory to generate reference plans and controls for the robots, and then using these references as imitation examples for supporting the learning of an action policy in a distributed reinforcement learning framework. Project’s main outputs will consist in novel algorithmic solutions for distributed planning, control, and human-swarm interaction for using ACRSs in EDM task scenarios. The developed solutions will be evaluated in realistic ROS simulations and will be ddemonstrated in realworld scenarios using a swarm of up to 6 robots. Collaboration with local stakeholders will be fostered to have concrete use cases to evaluate and apply the developed technologies. The project will develop advanced technologies and methodologies towards improving quality and eZciency of data mapping processes, which are a core component of any program of environmental analysis and governance, and are fully aligned with the goals of Understanding and protecting Qatar’s natural environment and of Environmental sustainability de1ned in the Qatar National Research Strategy. In particular, project will focus on marine environments, which are fundamental to preserve correct equilibrium of Qatar’s ecosystem. Moreover, developed solutions will also be applicable for inspection tasks in marine oil & gas platforms, which are at the core of Qatar’s economy.