SMART: Implementing a New Flight System Based on Multi-agent and Embedded on a Real-Time Platform

Current work in the field related to the drones mainly aimed at improving their autonomy related to the environment. The purpose of this paper focuses on the areas of improvement of autonomy, flexibility and adaptability of these aircraft during a flight. The developing of flying machines characterized by a pseudo-autonomy, better flexibility and adaptability to situations are mainly blocked by order constraints equipment, such as for architecture requires more complexity. To meet this need autonomous control architecture has been developed; it supports various aspects of the development process of the agent, from the design of the agent architecture, to the implementation on the hardware. An original architecture has been developed that allows the real-time control and manage different decisions by evaluating the path to follow and the speed of the UAV (unmanned aerial vehicle). It is deployed on an embedded system platform that provides good computing power for this kind of tasks in addition to managing communication with the ground operator.