Unlike existing aerial swarm systems, we aim at developing algorithms which do not require global or relative positioning information concerning agents and their neighbors. This alleviates the need for sensors which require calibration, are expensive and heavy or unusable because of environmental constraints. Rather than positioning, MAVs rely only on simple sensors (magnetic compass, speed sensor and altitude sensor) and local communication with neighbors. Our endeavor is motivated by an application whereby Micro Air Vehicles (MAVs) must organize autonomously to establish a robust communication network between users located on ground. Such a system is aimed towards the rapid and easy deployment of communication networks in disaster areas.