Early Stage Research projects

We are in the process of opening the 15 positions. When a position is open, a link to the job vacancy will be added.

Do you want to know more? Download the official call for applicants!

Platform independent 6D localization for UAV in outdoor and GNSS environment

The localization problem represents a main limitation on both outdoor and indoor environment, when floating base system need to perform precise tasks. Objectives: - develop a platform independent generalized 6D localization module for UAVs based on multimodal sensor fusion, - develop a state estimation scheme based on data fusion coming from onboard sensors, visual sensors and Ultra Wide Band information (UWB), - create a fast-deployable positioning scheme being able to increase the current localization accuracies. Note: the position is currently on hold

Stabilization and control of aerial manipulators in contact with the environment for on-site measurements

From a theoretical point of view, stability theorems exhibit only local validity and estimators of external disturbances may not be enough standalone, while they must be tailored on the design of the aerial manipulator itself. A fast reaction to unexpected situations should be guaranteed for reliable measurements.


- design and control an aerial manipulator for interaction with the environment for on-site measurements;

- study aerodynamic hurdles caused by the proximity of the floating platform with the surrounding environment, preventing even simple perching operations;

- develop estimators of unmodelled aerodynamic effects and external disturbances and test them in real life experiments.

Host: Universita’ di Napoli Federico II

Go to opening

Full-body Control of an Interactive Aerial Robot (closed)

Force estimation and position control in floating-base, sensor-less systems require the precise definition of the system model and trajectory planning.


- develop an accurate model of the full system that includes the dynamics of the arm, forces from the UAV motors and aerodynamic effects among other influences;

- use the model for full body control of the system where the arm can predict and compensate for the movements of the UAV

- combine the model with sensor measurements and motor torques for high accuracy estimation of the current forces the system is being subjected to.

Host: ETH Zurich

The position is closed

Aerodynamics-aware cognitive control of aerial manipulation of flexible objects (closed)

The project focuses on the design and control of aerial manipulators for outdoor applications in inspection and maintenance. Aerial robots with arms that interact physically manipulating objects and applying forces to the environment need to fly very close to them. Aerial robots have to generate dynamic airflow with their actuators (i.e., rotors) to get lift and manoeuvrability, and this airflow interacts with the robot frame, the arms and any object or surface around, and also with the wind and turbulence that are present outdoors. In case these are non-rigid or deformable, as for example in aerial robots with arms with elastic or variable stiffness joints, or when manipulating or flying close to vegetation, trees, cables, hoses or flexible panels, the elastic deformation can be coupled with the air movement and the robot dynamics, and directly affects the control and stability of the aerial robots.


The project will develop a new resilient cognitive control framework including machine learning techniques to adapt the robot to the previous situations. CFD analysis, tunnel/testbed and on-site experiments will be done to model the complex interactions and validate the control framework and its application range.

Host: University of Seville

The position is closed

High Force Tool Manipulation with Aerial Robots

Floating base systems have limitations applying meaningful forces on the environment due to the lack of base constraints.


- develop frame independent methods for dynamic modelling of floating-based systems physically interacting with the environment;

- develop mechanisms that allow exploiting the system dynamic for high force application (e.g. impacts);

- develop control strategies to ensure stable physical interaction near actuator’s saturation;

- validate results in real application in outdoor test facilities.

Host: DTU

Go to opening

Soft aerial robots physically interacting with humans and objects in the environment (closed)

The project focuses on the the development of an innovative aerial robotic system, which embeds soft-robotics technology for its use in multiple environments and tasks. It has been widely demonstrated that the use of Unmanned Aerial Vehicles (UAVs) has large benefits in many applications, due to their capability of accessing locations which are usually dangerous for human operators. Nowadays, much of the research done in this area has been focused on the automation of applications using UAVs neglecting human interaction. However, it has been recently highlighted that theses systems can be more efficient if they work in cooperation with human beings. However, aerial vehicles might possess an inherent risk as their propulsion system can harm the operators. Particularly, in intelligent transportations tasks (such as humanitarian aids), UAVs are promising. However, last-mile delivery usually implies the interaction with a person, which, as previously mentioned, might be at risk. For this reason, embedding soft-robotics technologies in these platforms will make them perfectly suitable to human-robot interaction making them not only faster and efficient, but also compliant and safe.The efforts will be focused on validating an UAV prototype that includes soft robotics technologies for allowing secure and safe co-working operations in a controlled environment.


This project aims to enhance the compatibility of this systems in environments shared with human operators performing a revolutionary change in the paradigm of aerial robotics. The use of this family of materials will transform the aerospace sector for aerial robots.

Host: University of Seville

The position is closed

Human-drone interaction performance assessment in aerial infrastructure inspection (closed)

There is a lack of understanding of how autonomous systems and human can interact together, to achieve human-like performances at work.


- to evaluate the impact of AR (ESR7), mapping (ESR8) and defect detection (ESR10) on the performance of an O&M remote operation (WP3), covering both drone operational aspects and data analysis;

- define the ontology of infrastructure inspection, modelling of operational goals, key points of interest, robot capabilities, payload capabilities, tasks, constraints, global plan, etc.;

- define the optimal inspection for a given scenario (iv) to analyse the operator cognitive workload;

- develop the operational infrastructure and the model for the evaluation and comparison against nominal performances.


Awards PhD: DTU

The position is closed

Bio-inspired Learning approaches for Anomaly Detection and Condition Monitoring with Aerial Robots (closed)

The detection of novelty in visual classification and the search for novelty is an aspect that is of paramount importance in asset damage exploration.


- enable the aerial manipulators to detect anomalies and novelties by using state-of-the-art machine learning methods to classify as adequately as possible known faults;

- develop methods to provide “novelty indicators” independently of specific classes;

- develop bio-inspired machine learning methods and CNN for anomaly detection and condition monitoring with multimodal sensing technologies.

Host: DTU

The position is closed

Immersive displays maintaining ultra-realism and 3D visual cues (closed)

Class requirements require the remote system to guarantee a level of realism that implies the development of new immersive technologies and algorithms.


- develop a visualization platform for immersive visualization in remote aerial manipulator control;

- optimize the amount of information given the limited sensing and imaging resources at the manipulator side;

- design a display system capable to seamlessly deliver visual cues in an ergonomically appropriate manner;

- determine the ways how potentially non-visual information can be localized and integrated into the visual feedback.

Host: Tampere University

The position is closed

Augmented Reality for enhancing semi-autonomous Remote Aerial Manipulation

The interface between the human operator and the autonomous system needs to ensure that only the most relevant information are shared, intuitively.


- handle the seamless integration/fusion of the robotic modeling, path planning, obstacle avoidance, low level flying control and the environmental perception;

- develop algorithms for autonomous robot to user suggestions;

- develop methods for user to robot intuitive command.

Host: Lulea Technical University Go to opening

Uncertainty Aware Mapping and Control for Aerial Robots (closed)

User-driven navigation in cluttered environment requires that autonomous system to provide information to the user regarding the operational risks.


- develop a system for uncertainty aware high-resolution mapping;

- transform the map into a form that can be used by (a) the onboard controller to allow it to prevent collisions, (b) by the AR interface to provide high-level user information to allow the operator to gain awareness of contacts during operation;

- develop means for detecting contact locations of the arm by mounting additional sensors or active components on the arm.

Host: ETH Zurich

The position is closed

Haptic control for remote teleoperation of cooperative aerial manipulators

Inspection operation may use several platforms to cooperatively accomplish a task, all requiring to be maneuvered simultaneously by the remote operator


- develop a control framework for single user to multi-aerial manipulator mission control;

- assess the effect of time delays on multi-robot collaboration physically interacting with each other and the environment;

- develop tele-manipulation control strategies using multi-modal user information for distributed system coordination by a remote user.

Host: University of Naples

Go to opening

Remote Control of Aerial Robots for Tool Manipulation

Current approach to tele-manipulation suffer from transparency issues and instability when applying forces on the environment


- develop passive approaches to tele-operation of floating base systems in physical interaction with the environment, for performing remote tool manipulation;

- develop energy-based metrics based on information flow between the human operator and the aerial tool.

Host: DTU

Go to opening

Advanced guidance and precise control of aerial robots for industrial inspections (closed)

The teleoperation framework needs proper information and additional aiding, to increase the perceptual capabilities of the aerial system for tele-manipulation


- design advanced stabilization and guidance techniques that allows an easy, robust and stable way to guide the contact device;

- generate commands to the aerial robot when it is closed to the locations to be inspected, in a natural and easy-to-learn way;

- develop a relative localization system based on on-board sensor and/or minimal infrastructure on the ground with minimum deployment time.


Awards PhD: University of Seville The position is closed

Drone based Robust Control Schemes against varying time delays

Time delays in teleoperation are a major source of failure for real-time control, with implications in aerial systems in remote user-environment interaction


- develop advanced robustness against time-delay in remote aerial manipulation due to a) the drone to base communication latency, b) the AR computation latency, and c) the operation and control time delays;

- propose novel control schemes, based on principles from the area of networked controlled system based on the measured time varying round trip delay;

- develop a reconfigurable control structure that satisfies the overall stability of the flying closed loop system and the performance of the flying control scheme, e.g. ability for dexterous maneuvers.

Host: LTU

Go to opening