Rui Song
Research Fellow
University of California, Los Angeles
Visiting Researcher
University of Cambridge
About Me
I am a research fellow at UCLA Mobility Lab, working with Prof. Jiaqi Ma. I am also a visiting researcher at Cambridge CV4DT Lab, working with Prof. Olaf Wysocki. I obtained my Ph.D. degree from Technical University of Munich (TUM) with highest distinction (summa cum laude), under the supervision of Prof. Alois Knoll and Prof. Andreas Festag, with invaluable mentorship from Prof. Holger Caesar. I was a senior researcher at Fraunhofer IVI, and led/co-led to 4 research projects with a combined budget over €15 million. Before that, I was a technical product owner for autonomous driving at ZF R&D Center and IPG Automotive.
My research interests include deep learning, computer vision, cooperative intelligence and their applications in autonomous driving, transportation or robotics.
News
🔥 Our OccuFly was selected as an oral presentation at CVPR!
🔥 Two workshops were accepted at ECCV. DriveX@ECCV2026 & TwinWorld.
🔥 Two papers were accepted at CVPR. OccuFly & SpaceDrive. Congrats to Markus & Peizheng!
Our workshops were accepted both at CVPR 2026 and IEEE IV 2026. Find more details here: DriveX@CVPR2026 & DriveX@IV2026.
I joined the University of Cambridge as a visiting researcher and look forward to collaborating with outstanding researchers there.
I moved to Los Angeles and started my postdoctoral research at UCLA, exploring the intersection of spatial AI and autonomous systems.
One paper was accepted at NeurIPS. IPFormer.
Two papers were accepted at ICCV. Coda-4DGS & TrafficLoc.
I was recognized as an outstanding reviewer for CVPR 2025.
One paper was accepted at IROS.
One workshop was accepted at ICCV 2025. Details: 2nd DriveX.
Two papers were accepted at IEEE IV.
I successfully defended my PhD at TUM and graduated with highest distinction (summa cum laude).
One workshop was accepted at CVPR 2025. Details: 1st DriveX.
One workshop was accepted at IEEE IV 2025. Details: DLIVA.
I was recognized as an outstanding reviewer for ACM MM 2024.
One paper was accepted at ECCV Workshop, and we won 3rd place in the ECCV Dataset Distillation Challenge.
One paper was accepted at IEEE IV.
Two papers were accepted at CVPR: CoHFF & TUMTraf-V2X.
One paper was accepted at Knowledge-based Systems (IF=8.8).
One paper was accepted at ICRA-CoPerception.
I open-sourced a PyTorch repository for federated learning (FL), integrating multiple FL algorithms, models, and datasets.
One paper was accepted at CVPR (Highlight).
We released SensorTraffix, a SENSORIS-based software product for infrastructure communication, as part of Fraunhofer IVI assets. Contact.
I had a live interview at the opening event for our 5G Innovation Lab in Ingolstadt.
One paper was accepted at IEEE VTC.
I officially started my PhD at TUM Informatics-VI department.
One paper was accepted at FISITA World Congress (Outstanding Paper Award).
I joined Fraunhofer IVI as a researcher and served as lead in the InFra research project.
Selected Publications
* equal contribution; † corresponding author. Full list at Google Scholar.
ConFixGS: Learning to Fix Feedforward 3D Gaussian Splatting with Confidence-Aware Diffusion Priors in Driving Scenes
Under review, 2026
MDrive: Benchmarking Closed-Loop Cooperative Driving for End-to-End Multi-agent Systems
Under review, 2026 · Hosted as an official challenge at CVPR 2026
TIC-VLA: A Think-in-Control Vision-Language-Action Model for Robot Navigation in Dynamic Environments
ICML 2026
Project Page | Paper | Code
SpaceDrive: Infusing Spatial Awareness into VLM-based Autonomous Driving
CVPR 2026
Project Page | Paper | Code
IPFormer: Visual 3D Panoptic Scene Completion with Context-Adaptive Instance Proposals
NeurIPS 2025
Project Page | Paper | Code
CoDa-4DGS: Dynamic Gaussian Splatting with Context and Deformation Awareness for Autonomous Driving
ICCV 2025
Project Page | Paper | Code
TrafficLoc: Localizing Traffic Surveillance Cameras in 3D Scenes
ICCV 2025
Project Page | Paper | Code
Collaborative Semantic Occupancy Prediction with Hybrid Feature Fusion in Connected Automated Vehicles
CVPR 2024
Project Page | Paper | Code
TUMTraf V2X Cooperative Perception Dataset
CVPR 2024
Project Page | Paper | Data
Invited Research Talks
Scene Models as a Foundation for Autonomous Driving
University of Cambridge, UK
From Real Data to Generative Scene Modeling for Autonomous Driving
University of California, Merced, USA
Scene Reconstruction and Synthesis for Autonomous Driving
University of California, Riverside, USA
Academic Activities
Workshop Organizer
4th Workshop on Foundation Models for V2X-Based Cooperative Autonomous Driving (DriveX@CVPR2026).
3rd Workshop on Cooperative Intelligence for Embodied AI (MEIS).
Workshop on Embodied and Safe-Assured Robotic Systems (E-SARS).
Workshop on Data-Driven Intelligent Vehicle Applications (DDIVA).
Tutorial Organizer
Program Committee Member
1st Workshop on Cooperative Intelligence for Embodied AI (MEIS).
Reviewer
CVPR, ICCV, ECCV, NeurIPS, ACM MM, AISTATS, IJCNN, IROS, IEEE IV, ITSC, TRB.
TPAMI, TNNLS, TCSVT, RA-L, TR-C, Information Fusion, Neurocomputing.
Outstanding Reviewer Award: CVPR 2025, ACM MM 2024.
Research Projects
2021-2024: 5G Innovation Concept in Ingolstadt (5GoIng), funded by the German Federal Ministry for Digital and Transport (BMDV)
2021-2024: Ingolstadt Innovation Labotory (In2Lab), funded by Bavarian Ministry of Economic Affairs (StMWi), Regional Development and Energy
2021-2023: Artificial Intelligence in the Transportation System in Ingolstadt (KIVI), funded by the German Federal Ministry of Education and Research (BMBF)
2021-2023: Intelligent Infrastructure (Infra), funded by Bavarian Ministry of Economic Affairs (StMWi), Regional Development and Energy.
2018-2021: SetLevel4to5, funded by the German Federal Ministy for Economic Affairs and Energy (BMWi)
2017-2018: Intelligent Maneuver Automation – cooperative hazard avoidance in realtime (IMAGinE) funded by the German Federal Ministy for Economic Affairs and Energy (BMWi)
2017-2018: PEGASUS funded by the German the Federal Ministry for Economic Affairs and Energy (BMWi)
Hobbies
Snowboard: Certificated as level 3 instructor by DSLV (German Ski/Snowboard Instructor Association).
Marathon: PB 3h55; 5KM: PB 18.
© Copyright 2025 Rui Song