Robust Trajectory Generation for Event-Based Drones (M/F)

Job Description

Missions

The primary mission of the Postdoctoral researcher is to design and develop a robust collision avoidance chain for autonomous drones equipped with event-based cameras. The ultimate objective is to bring the algorithmic solution to a mature stage, ensuring robustness, reactivity, and regulatory compliance for a TRL 7 demonstration in quasi-operational conditions, such as urban delivery or automated inspection

Activity

- Algorithmic Foundations: Design optimization-based trajectory generation algorithms (spanning both offline and online settings) that strictly account for the drone's actuation and perception constraints.
- Metric Definition & Robustness: Test and define observability metrics to optimize the accuracy and speed of obstacle detection. Evaluate robustness metrics—specifically focusing on statistical lower bounds rather than simple expected values—to generate motion plans that remain valid despite uncertainties in the drone and camera models. Incorporate air traffic rules and safety requirements into the mathematical optimization framework.
- Optimization and Learning: Optimize the perception and trajectory generation pipeline using advanced learning approaches, such as Knowledge Distillation, to enable real-time computation on lightweight embedded resources.
- Prototyping: Test the reactive trajectory generation algorithms within a simulation environment, progressively relaxing simplifying assumptions (such as perfect state estimation or zero measurement noise) to ensure resilience.
- Experimental Validation: Port the algorithms to real hardware (a drone with onboard event-based cameras) for experimental validation, starting from simplified conditions and culminating in flight testing under realistic outdoor conditions.

Your Profil

Skills

- Education: Ph.D. degree in Robotics, Control, Optimization, Machine Learning, Computer Vision for Robotics, or related fields.
- Technical Expertise: Strong background in optimization-based and/or learning-based motion generation for robots (particularly aerial drones).
- Development: Proven experience and proficiency in C++ and Python.
- Specific Knowledge: Prior knowledge of event-based vision (neuromorphic engineering) is highly desirable.
- Soft Skills: Scientific curiosity, high degree of autonomy, and the ability to work independently within a collaborative project environment.

Your Work Environment

The PostDoc research will naturally fit into our several activities over the last years about the general problems of shared control of robots and control/estimation for multi-robot systems.

The facilities in the team include a vicon-instrumented indoor room for experiments with multiple quadrotors, four manipulator arms (two 6-dof industrial manipulators, and two 7-dof torque-controlled manipulators), and state-of-the-art 6-dof haptic devices. Exploiation of the Immersia VR facility (on campus) for running human/multi-robot experiments will also be possible (and encouraged).
Finally, two research engineers (technical staff) will assist the PostDoc for all what concerns hardware/software development and maintenance of the robotic platforms.

Constraints and risks

N/A

Compensation and benefits

Compensation

between 2900€ and 4260€ gross depending on the candidate's experience

Annual leave and RTT

44 jours

Remote Working practice and compensation

Pratique et indemnisation du TT

Transport

Prise en charge à 75% du coût et forfait mobilité durable jusqu’à 300€

About the offer

About the CNRS

The CNRS is a major player in fundamental research on a global scale. The CNRS is the only French organization active in all scientific fields. Its unique position as a multi-specialist allows it to bring together different disciplines to address the most important challenges of the contemporary world, in connection with the actors of change.

CNRS

The research professions