Planning of a multi-agent mobile robot-based adaptive charging network for enhancing power system resilience under extreme conditions

The rapid proliferation of electric vehicles (EVs) poses escalating challenges to voltage stability in modern distribution systems, especially under extreme conditions (e.g., sudden load surges, generator failures, or network disruptions). Conventional fixed charging stations (FCS) lack flexibility, failing to adapt to the spatiotemporally dynamic and heterogeneous nature of EV charging demand, thereby exacerbating grid resilience risks. To address this, we propose a dual-mode multi-agent mobile robot-based adaptive charging network (MRACN) that enhances power system resilience through real-time, intelligent coordination of mobile charging resources. Our dynamic scheduling strategy enables MRACN units to switch seamlessly between cost-optimized dispatch during normal operations and resilience-driven prioritization during emergencies, guided by a real-time voltage stability index (VSI)-derived resilience coefficient. The framework combines a power-traffic co-simulation environment modelling urban congestion, heterogeneous EV mobility (private, taxi, fleet), and spatiotemporal demand variations, and a multi-objective mixed-integer nonlinear programming (MINLP) model optimizing energy cost, customer satisfaction, and voltage stability under realistic constraints (traffic delays, mobility restrictions, grid safety margins). Simulations on a coupled IEEE 33-bus distribution system and 32-node transportation network demonstrate MRACN's superiority in enhancing voltage resilience, reducing operational costs, and improving EV service reliability across both normal and extreme grid scenarios. The results validate MRACN as a scalable, adaptive solution for resilient smart grids in EV-dominated environments.