Modern cities have to respond to the growing demands of more efficient and sustainable urban development, as well as an increased quality of life. In this context, the cities of the future will need the ability to gain insight about current urban conditions and react dynamically to them. According to this view, "smart cities" can be seen as cybernetic urban environments in which different agents (e.g., citizens) and actuators (e.g., robots) exploit the city-wide infrastructure as a medium to operate synergistically. Urban Swarms explores the feasibility of swarm robotics systems in urban environments. By using bio-inspired methods, a swarm of robots is able to handle important urban systems and infrastructures, improving their efficiency and autonomy. A diverse set of simulation experiments were designed and conducted using real-world GIS data. Results show that the proposed combination is able to outperform current approaches. Urban Swarms not only aims to show the efficiency of our proposed solution, but also to give insights about how to design and customize these systems. CityScope Volpe ABM model has been customized to integrate Swarm behavior using the Gama Platform as an open source project.