PhD Researcher · L2S CentraleSupélec / CNRS / Université Paris-Saclay
Ilyes
Jaouedi
Camera-radar fusion via optimal transport — developing geometry-aware sensor fusion architectures for ADAS, at L2S in collaboration with FORVIA.
Paris, France
PhD Candidate
Since Oct. 2024
Laboratory
L2S · CentraleSupélec / CNRS / Université Paris-Saclay
Industry
FORVIA Collaboration
Certified
GCP Professional ML Engineer
Who I am
About Me
I am a PhD candidate at the Laboratory of Signals and Systems (L2S), CentraleSupélec / CNRS / Université Paris-Saclay. My research sits at the intersection of computer vision, radar signal processing, and mathematical optimization.
Specifically, I work on fusing camera and radar modalities using optimal transport theory — a principled mathematical framework for comparing and aligning probability distributions — to build robust, geometry-aware perception systems for Advanced Driver Assistance Systems (ADAS). This PhD is conducted in industrial collaboration with FORVIA.
I hold an engineering degree from Ecole Polytechnique de Tunisie.
PhD Candidate
2024 – PresentCentraleSupélec — L2S / CNRS / Université Paris-Saclay
Industrial Partner
PhD CollaborationFORVIA — Camera-Radar Fusion for ADAS
Multidisciplinary Engineering Degree
2021 – 2024Ecole Polytechnique de Tunisie
Preparatory Cycle — Math & Physics
2019 – 2021IPEIM Monastir
Based in
Paris, France
PhD Research
Research
My PhD investigates the fusion of camera and 4D radar for robust autonomous driving perception. The central challenge — and contribution — is developing mathematically principled methods to align heterogeneous sensor representations, bridging the optical and RF domains using optimal transport theory.
Pipeline
Optimal Transport
Cross-modal alignment
3D Object Detection
Uncertainty-aware
ADAS Systems
All-weather perception
Affiliation
Laboratory of Signals and Systems (L2S) ↗CentraleSupélec · CNRS · Université Paris-Saclay
Signal & Statistics Group (GME)
Supervisors
Gilles Chardon
José Picheral
Industrial Partner
FORVIA ↗FAURECIA · HELLA — Tier-1 Automotive Supplier
4D Radar · Camera Fusion · ADAS Perception
Keywords
Career
Experience
PhD Candidate — Sensor Fusion for ADAS
L2S, CentraleSupélec / CNRS / Université Paris-Saclay ↗Oct 2024 – Present
Gif-sur-Yvette, France
Researching camera-radar fusion via optimal transport theory for Advanced Driver Assistance Systems. Developing OT-based cross-modal alignment methods for 4D radar and camera data. Industrial collaboration with FORVIA.
Computer Vision Researcher — Internship
CAOR, Mines Paris – PSL
2024
Paris, France
Developed a model for extracting and reconstructing vehicle trajectories from traffic camera video streams fused with LiDAR data. Worked on 3D reconstruction and tracking pipelines for urban traffic analysis.
Computer Vision Engineer — Internship
Enova Robotics
2023
Sousse, Tunisia
Implemented LiDAR-camera sensor fusion models using PV-RCNN and BiProDet architectures for 3D object detection. Benchmarked and fine-tuned multi-modal perception pipelines on custom robotic datasets.
Portfolio
Projects
Acoustics-Camera Fusion for 3D Source Localization
Sensor array and camera fusion for 3D acoustic source localization in real-world environments. Published at IEEE ICASSP 2026 in collaboration with L2S, CentraleSupélec.
High-Resolution Radar Imaging
High-resolution radar point cloud generation from raw automotive radar ADC data. Evaluated on the RADIal real-world driving dataset.
Radar-Camera Fusion for Autonomous Driving
Fusion of high-resolution radar imaging with camera-derived visual priors for robust 3D perception in autonomous driving scenarios.
RadarFlows — Synthetic Radar Generation
Generative model for realistic synthetic radar map synthesis conditioned on camera observations and 3D scene detections. Designed as a data augmentation pipeline for automotive radar perception.
PowerAnything — Visual Radar Power Prediction
Predicts dense radar received-power maps from RGB camera frames, providing a learned visual prior for cross-modal radar-camera fusion pipelines.
Expertise
Skills & Technologies
Signal Processing & Radar
Computer Vision & Deep Learning
Mathematical Foundations
Frameworks & Libraries
Languages & Tools
Cloud & AI Platforms
Credentials
Certifications
Google Cloud
Professional Machine Learning Engineer
Validates the ability to design, build, and productionize ML models to solve business challenges using Google Cloud technologies and established ML practices.
Verify on CredlyGet in touch
Contact
I am open to academic collaborations, research discussions, and relevant engineering opportunities. Feel free to reach out.