The role of renewable energy communities, consisting of citizens, businesses and institutions that produce, consume, store, and share energy, is becoming increasingly prominent in the energy landscape and in electricity markets. In this context, the opportunity to use community shared batteries arises and, therefore, the necessity to adapt battery control strategies to the needs of the community is an open question that is currently investigated by the power and energy community.This study frames, simulates, and experimentally tests scheduling and control models assessing their feasibility for a community to self-dispatch, engage in energy arbitrage and maximize collective self-consumption as well as preserving battery lifespan. A two-stage optimal control is proposed for a community whose resources are interfaced by means of a LV radial network. The first scheduling layer calculates the optimal dispatch plan and battery trajectories to maximize profits over a future horizon, based on production and load day-ahead forecasts. The second real-time control adjusts battery set points to minimize hourly dispatch errors based on real-time situational awareness of the energy community grid and short-term forecasts of both loads and renewables’ generation.The performance assessment of the proposed two-stage control is evaluated through simulations (i.e., to study the sensitivity of key parameters) and then, experimentally validated in a real-scale microgrid characterized by a configuration close to the one of an actual energy community.