Lead researcher:

Bálint Vanek, PhD, Research Advisor, Deputy Head of Research Laboratory, SZTAKI
Dániel Rohács, PhD, Associate Professor, Head of Department, BME

Background

  • An unmanned aerial vehicle (UAV) is an aircraft that does not have a human pilot on board, but still flies. A UAV is normally used within an unmanned aircraft system (UAS), which includes the mentioned UAV, a ground-based controller, and some means of communications between the two.
  • The UAV flights are carried out with various degrees of autonomy: either under remote control by some human operator, or autonomously relying on onboard computers. Aspects of the latter flight capabilities, as well as the vehicles fit out with such capabilities are the targets of the present research.
  • Historically, the UAVs were mostly used in military applications. Nowadays, however, their use rapidly extends into civilian application fields/airspace, as well. The civil applications include aerial photography, product deliveries, agriculture, policing and surveillance, infrastructure inspections, engineering to mention a few.

Research and development directions and activities in the field

  • Evaluation of UAV-based solutions to various high-level aerial tasks/missions, such as the detection of particulate matter in the air, an aerial survey of certain forest contaminations, the evaluation of some cultivated areas. These tasks and missions can be accomplished using either unaccompanied UAVs, or squads of cooperating UAVs.
  • Harmonizing and synchronizing the operation of various on-board sensors, multi-camera vision systems, and communication devices, as well as that of actuators.
  • Planning and execution of coordinated flights of cooperating aircraft based on cutting-edge technology that ensures safe aerial movements/manoeuvers of autonomous UAVs, furthermore facilitates their intra-squad cooperation and communication.
  • Developing methods and setting out technologies that support and ensure a high-level secure communication between land/road vehicles and aircraft that need to operate in a coordinated and safe manner in joint missions.
  • Analysing the methods of coordinated control applicable for such joint missions, with special regard to the arising safety, security, traffic management, and cartographical issues, tasks and problems.
  • Mapping, understanding, modelling and predicting the joint movement of systems that comprise vehicles manufactured using different technologies, and exhibit different dynamic behaviour.
  • For the effective operation of such heterogeneous systems, both some local low-level control for the individual vehicles, and a system-level – so-called supervisory – control responsible for the coordination among the vehicles are required. Furthermore, when designing such a hierarchical control, also the real-time issues and criteria must be considered, modelled and analysed.
  • Densely populated areas are frequent targets of UAV-based measurement, monitoring and data collection missions. Permitting UAVs to fly and operate over such areas involves various obvious and less obvious safety and security risks. These risks must be minimized, and if something still goes wrong with the UAV operation, then dangerous situations must be tackled to avoid serious consequences.
  • Focusing on the above aspects and possibilities, the researchers of the Consortium intend to combine approaches based on reliability theory with flight/route planning capabilities and with reconfigurable control. Also, the sense-and-avoid methods are expected to play a significant role in this respect.
  • Setting up an experimental infrastructure for autonomous UAVs equipped with remote sensing capabilities. To achieve a useful level of service with such UAVs, it is required that their autonomy extends to both the remote sensing operations, and to the observance of aerial separation rules with respect to nearby aircraft.
  • In the context of beyond-visual-line-of-sight (BVLOS) flights, the Consortium intends to set up an experimental infrastructure that supports autonomous BVLOS operation of UAVs, and enables secure high-bandwidth bi-directional communication, e.g., based on 5G wireless technology, between the UAVs, and between the UAVs and the ground station.
  • Validating of the autonomous operation of autonomous vehicles, be they land/road, or aerial vehicles, and guaranteeing of their safe operation are particularly challenging tasks within the present research. Much of the test and validation tasks must be carried out via simulation as the number of vehicle model-parameters and of the environmental variables is too large to experiment with during real drives/flights.
  • Generating flight situations for simulations and test flights in an automatic fashion. This task is motivated by the generally excepted view that tackling critical and emergency situations is fundamental to safe traffic and transport.
  • Among many other causes, faulty sensors, the decreasing performance of the actuators, as well as uncertainties, disturbances and unexpected situations (e.g., road closures on the ground, approaching damaged, illegal, or clearly hostile aircraft in the air) can result in critical and even emergency situations during the operation of autonomous vehicles.
  • Developing methods and technologies for automated collection and recognition of the above losses, disturbances and situations from relevant data signals, designing and testing adequate responses to these are of prime importance for traffic and transport safety and security.
  • Selecting and setting adequate formal performance specifications that conform to robustness requirements.
  • The probabilistic relaxation is frequent-ly employed to increase the efficiency of the control. In the designs relying on this technique, the aim is to find controllers that work as expected in a great majority of the probabilistic cases. The approach makes it possible to agreeably manage situations that cannot be computed in a timely fashion with the more customary control design methods due to the exceedingly high computation times, and/or their exceedingly high computing and storage resource requirements. On the other hand, the theoretical robustness analysis of such probabilistic control methods is theoretically very challenging. The researchers of the Consortium intend to address both the theoretical and practical aspects of the above design issues and problems.