Towards More Autonomous and Safer Unmanned Aerial Vehicle Systems

在过去的十年中，无人驾驶飞行器（uav）经历了真正的繁荣，它们在日常生活中的应用成倍增加：航空摄影，基础设施检查，农业用地监测，可能在不久的将来自动包裹递送。即使这些无人机具有一定的吸引力，它们的自主水平仍然有限，并且操作员的存在执行某些操作通常是必要的。对于许多应用来说，一个特别有趣的领域是完全自主的起飞和降落，特别是在高速行驶的移动车辆上。本着这种精神，我们的研究小组参加了2016年大疆挑战赛，使无人机以高达50公里/小时的速度自动降落在车辆上，这是迄今为止的记录。

Over the past decade, unmanned aerial vehicles (UAVs) have experienced a real boom with a multiplication of their applications in everyday life: aerial photography, infrastructure inspection, agricultural land monitoring and probably automated parcel delivery in the near future. Even if these UAVs have a certain appeal, their level of autonomy is still limited, and the presence of an operator to perform certain manoeuvres is often necessary. A particularly interesting area for many applications is then the completely autonomous take-off and landing, especially on mobile vehicles travelling at high speed. In this vein, our research group's participation in the DJI Challenge 2016 led to the autonomous landing of a drone on a vehicle at speeds of up to 50 km/h, which remains a record to date. The long-term objective of this research program is to make drones safer and more autonomous in order to diversify their applications. To this end, this discovery grant focuses in particular on the take-off and landing phases of drones, and their complete automation. Thanks to the experience acquired by my group, we wish to expand our knowledge in the field and improve the guidance, navigation and control algorithms currently available. For the short-term objectives, we will study the following three themes: (1) high-fidelity modelling of a drone moving at high speed and its behaviour in the wake of a mobile vehicle; (2) the use of bidirectional thrust to increase the agility of multicopters and make the most of their flight envelope; and (3) means for estimating the relative position of the drone in relation to the mobile vehicle, and the take-off, approach and landing algorithms that will use bidirectional thrust. Advances in these three themes will expand the capabilities of drones and improve their autonomy. The research conducted will have a significant impact on the scientific community and the UAV industry, which will be able to offer more efficient solutions or use these results in their own applications. At present, the training of highly qualified personnel in this field is essential. With a growing market, the sector needs more and more experts: this research program will train people with advanced theoretical knowledge and practical experience acquired in our laboratory.