ADLab of Shanghai AI Lab
This is the official website of ADLab (ADG at Shanghai AI Laboratory).
As a research team affiliated with the Shanghai Artificial Intelligence Laboratory, ADLab is at the forefront of autonomous driving research. Our current endeavor involves developing a next-generation autonomous driving system by integrating human knowledge and common sense reasoning. We believe that this knowledge-driven paradigm will elevate the reliability and generalizability of autonomous driving systems to unprecedented levels.
Our dynamic team consists of approximately 20 talented faculty members and over 20 interns with huge potential (including full-time research interns and joint Ph.D. students).
Join our ranks and be a part of shaping the future of this pioneering technology.
news
selected publications
-
arXivOASim: an Open and Adaptive Simulator based on Neural Rendering for Autonomous DrivingarXiv preprint arXiv:2402.03830, 2024
-
arXivLimSim++: A Closed-Loop Platform for Deploying Multimodal LLMs in Autonomous DrivingarXiv preprint arXiv:2402.01246, 2024
-
ICLRDiLu: A Knowledge-Driven Approach to Autonomous Driving with Large Language ModelsIn The Twelfth International Conference on Learning Representations, 2024
-
ICLRReSimAD: Zero-Shot 3D Domain Transfer for Autonomous Driving with Source Reconstruction and Target SimulationIn The Twelfth International Conference on Learning Representations, 2024
-
arXivTowards knowledge-driven autonomous drivingarXiv preprint arXiv:2312.04316, 2023
-
arXivStreetsurf: Extending multi-view implicit surface reconstruction to street viewsarXiv preprint arXiv:2306.04988, 2023
-
arXivDrive like a human: Rethinking autonomous driving with large language modelsarXiv preprint arXiv:2307.07162, 2023
-
NeurIPSRangePerception: Taming LiDAR Range View for Efficient and Accurate 3D Object DetectionAdvances in Neural Information Processing Systems, 2023
-
NeurIPSAD-PT: Autonomous Driving Pre-Training with Large-scale Point Cloud DatasetAdvances in Neural Information Processing Systems, 2023
-
ITSCLimSim: A Long-term Interactive Multi-scenario Traffic SimulatorIEEE 25th International Conference on Intelligent Transportation Systems (ITSC), 2023
-
ICCVDetZero: Rethinking Offboard 3D Object Detection with Long-term Sequential Point CloudsIn Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023