Autonomous planning and replanning of a single Unmanned Aerial Vehicle: strategies and simulations

The aim of this paper is to define and elaborate the main features of an Unmanned Aerial System (UAS) in order to make it as autonomous as possible. In this context, the authors would propose and verify possible strategies for an autonomous replanning of a route in case out-of-nominal situations would happen. This activity has been performed within the framework of SMAT F2, an Italian project in which Politecnico di Torino is involved. SMAT (Sistema di Monitoraggio Avanzato del Territorio – Advanced Territory Monitoring System) is a project funded by Regione Piemonte and Fondo Europeo di Sviluppo Regionale (F.E.S.R.), now at its second phase, that studies and demonstrates an advanced monitoring system able to comply with planned task (e.g. traffic monitoring, pollution monitoring, plantations observation and measurements, etc.) and to prevent and monitor different types of emergency events (e.g. floods, fires, landslips, etc.). In the first part of the paper, a general methodology to perform an autonomous re-planning is proposed, with a particular attention to the techniques already proposed and studied by other research teams [1], [2]. This activity can be seen as a preliminary study to create an algorithm able to propose alternative routes in case replanning actions would be required. In this context, it was necessary to define how many and what kind of critical conditions the UAS has to recognize and correct autonomously. All these critical conditions have been separated in different scenarios. After having analysed in depth the requirements provided by the principal task partners, it was decided to limit the analysis to the following scenarios: Weather phenomena (No Fly Zone or NFZ) and Permanent No Fly Zone Avoidance; New Targets/New Waypoints from SSC; Vehicle Failure/Payload Failure that require to land; Vehicle Failure and Payload Failure that degrade the overall mission. With the aim of studying the replanning algorithms for each above-mentioned scenario, the earlier phase of this work leads with the definition of the logical processes that stand behind them. The most important purpose of this activity is to define the main requirements and all the logical processes required to allow the mission planning and re-planning. This phase of the work produced two different outputs. The first consists in a flow chart definition aimed to describe the logical processes of the mission re-planning. The other output is a top-level requirement definition. For this purpose, a typical System Engineering approach has been proposed and applied, exploiting its iterative and recursive characteristics. All these evaluations would be a guideline for the generation of a code able to autonomously propose in real-time replanned routes, like those proposed in literature [3], [4]. The second part of the work deals with the simulations of the previous identified scenarios using a specific software called STK (Systems Tool Kit, previously Satellite Tool Kit). This tool is able to determine not only the assets, the attitude, the dynamic position in the geographical space and in the time of the considered UAV, but also the existing relationships among the objects modelled including possible relationships (i.e. accesses) taking into account a certain number of simultaneous constraining conditions. In conclusion, the paper would like to highlight some peculiar aspects of an autonomous re- planning in order to enhance the efficiency and efficacy of an Unmanned Aerial System. This is considered as a fundamental step to develop and simulate replanning solutions for a fleet of UAVs.