The Faculty of Aerospace Engineering of Delft University of Technology, the Netherlands, announces a vacancy for one Ph.D. student within the project:
Neuromorphic Perception for Drones
Description of the project:
Neuromorphic sensing and processing form a highly promising technology for creating autonomous small drones, due to the potential for high-speed perception at a low energy cost. The goal of this research project is to develop neuromorphic vision algorithms for complex perceptual tasks that are relevant to drones, such as optical flow and ego-motion estimation. Specifically, we will investigate self-supervised learning algorithms for spiking neural networks (SNNs) with the goal of implementing them onboard of our drone in the context of a control task. In the research we will expand upon our previous work on unsupervised learning of optical flow with a spiking neural network , self-supervised learning (e.g., [2,3]), and transferring learned SNNs to controlling a real drone in flight .
The project will be carried out within the Control and Simulation section of the Faculty of Aerospace Engineering. A large component will consist of developing self-supervised neuromorphic algorithms applied to event-based camera data, typically performed with the help of a scripting language like Python. However, during the development of the algorithms, the final application on real onboard neuromorphic hardware has to be taken into account, limiting the allowed complexity of the networks. Successful algorithms will then be ported to a real drone, which will be a team effort. The supervisors will be prof. dr. Guido de Croon and dr. Julien Dupeyroux from the Micro Air Vehicle Laboratory.
What do we ask?
We are looking for a candidate with an MSc degree in an area such as Artificial Intelligence, Computer Science, Robotics, Aerospace Engineering, or a similar field. The candidate is expected to be passionate about developing novel AI algorithms with an application on real robotic hardware (“AI at the edge”). Programming experience is required in Python/MATLAB and/or C/C++.
Previous experience with deep learning algorithms / autonomous robots is an asset. Please note that the Ph.D. candidate will work within a team with ample experience in design and prototyping of tiny drones, autopilots (in particular Paparazzi) and micro-electronics as well as with bio-inspired artificial intelligence and control. The candidate must have strong analytical skills and must be able to work at the intersection of several research domains (artificial intelligence, control, robotics, aerospace). A very good command of the English language is required, as well as excellent communication skills.
What do we offer?
TU Delft offers PhD-candidates a 4-year contract, with an official go/no go progress assessment after one year. Salary and benefits are in accordance with the Collective Labour Agreement for Dutch Universities, increasing from € 2395 per month in the first year to € 3061 in the fourth year. As a PhD candidate you will be enrolled in the TU Delft Graduate School. The TU Delft Graduate School provides an inspiring research environment with an excellent team of supervisors, academic staff and a mentor. The Doctoral Education Programme is aimed at developing your transferable, discipline-related and research skills.
The TU Delft offers a customizable compensation package, discounts on health insurance and sport memberships, and a monthly work costs contribution. Flexible work schedules can be arranged. For international applicants we offer the Coming to Delft Service and Partner Career Advice to assist you with your relocation.
How to apply?
All applications must be done online here.
 Paredes-Vallés, F., Scheper, K. Y., & de Croon, G. C. H. E. (2019). Unsupervised learning of a hierarchical spiking neural network for optical flow estimation: From events to global motion perception. IEEE transactions on pattern analysis and machine intelligence, 42(8), 2051-2064.
 de Croon, G. C. H. E., De Wagter, C., & Seidl, T. (2021). Enhancing optical-flow-based control by learning visual appearance cues for flying robots. Nature Machine Intelligence, 3(1), 33-41.
 Ho, H. W., De Wagter, C., Remes, B. D. W., & de Croon, G. C. H. E. (2015). Optical flow for self-supervised learning of obstacle appearance. In 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (pp. 3098-3104). IEEE.
 Dupeyroux, J., Hagenaars, J., Paredes-Vallés, F., & de Croon, G. C. H. E. (accepted at ICRA 2021). Neuromorphic control for optic-flow-based landings of MAVs using the Loihi processor.