Researcher (M/F/X) on AI for human-friendly drone control Departement of Mechanics project “HADRON”

Updated: 2 months ago
Job Type: FullTime
Deadline: 10 Jul 2022

In the framework of the study HADRON on the development of Artificial Intelligence approaches enabling human-friendly control for military drone operations, we are looking for a full-time researcher with a master degree in Applied Sciences / Engineering / Physics / Computer Science / Informatics and experience in the field of Robotics & Artificial Intelligence (AI).

The Royal Military Academy of Belgium (RMA) is a military institution responsible for the basic academic, military and physical training of future officers, and for the continuing advanced training of officers during their active career in the Belgian Defense department ( ). It is fully recognized as a university, fulfilling the same criteria as civilian universities. The Royal Military Academy is also conducting scientific research at university level for projects funded by the Belgian Defense department or external sources.
You work within the research cell ‘Robotics & Autonomous Systems’ ( ) of the department of Mechanical Engineering of the Royal Military Academy and in close collaboration with the Belgian company SkyeBase ( ). You conduct scientific research at university level on a project entitled ‘HADRON’ (Human-friendly Control & Artificial Intelligence for Military Drone Operations).
You work within a multi-national research team and in close collaboration with the Belgian Navy and LAND component and a commercial partner.

The use of unmanned aerial systems or drones, also within Belgian Defence, is exploding in recent years. This poses challenges related to the training of the personnel able to operate these tools. It is the vision that in the future, defence organisations will evolve to a situation where any soldier is a drone operator and the drone becomes just one of the many tools in the toolkit of the individual soldier. However, in order to arrive to such a situation, there are still multiple obstacles to be tackled. One major problem is that both the control and the data interpretation of current advanced military drones are much too complicated to be handled by personnel that has not received an extensive training, whilst it is impossible (or not feasible) to organize such a training for each and every soldier. A solution to this problem would be the augmented use of autonomous capabilities, both for the vehicle control and for the data interpretation aspect.
The global objective of this study is to render the use of military drones easier for the military operator. This is to be achieved through the development of artificial intelligence (AI) capabilities assisting the human operator in controlling the drone and in processing the sensor data, thereby alleviating the need for extensive operator training.

Specific objectives of this study are to:
• Develop AI tools, allowing for safe and human-friendly control and navigation of drones. It is specifically *not* the goal to develop totally autonomous capabilities, but to develop assistive tools & algorithms that enable:
o a novice pilot to operate a drone safely and effectively
o an expert pilot to operate multiple drones from a distance.
• Develop AI tools, assisting drone operators with semi-automated real-time data processing. The main goal here is to take the cognitive load on the human operator away from assessing large data streams towards assessing anomalies that have been detected automatically. Based on advice of COMOPSLAND and COMOPSNAV, the focus of this study will be on reconnaissance and surveillance operations, in order to maximize the impact.
(image: see Doc attached)
Main Tasks
• Develop the operational and technical user requirements for the project, in close collaboration with the Belgian Navy and LAND component.
• Perform a literature and state-of-the-art study and develop use cases for the validation of the tools to be developed, based on the input of the Belgian Defence clients.
• Develop novel multi-modal control interfaces, allowing military personnel without extensive training to control drones for their standard operations.
• Develop a novel control interface, allowing expert military drone operators to control not only one but multiple drones at the same time
• Develop algorithms to select automatically the most relevant parts of the data gathered by drones and to transform this data into information by automatic data interpretation using artificial intelligence approaches.
• Develop training tools enabling Belgian Defence drone operators to work with the developments.

• It is possible to perform a PhD. in the framework of the research study. This is facultative and up to the choice of the applicant. Choosing to pursue a PhD. will add to the task list to follow courses imposed by the doctoral school and write & defend a PhD. thesis about the study.

View or Apply

Similar Positions