YUKAI MA

I am a visiting researcher at the Department of Computer Science, University of California, Los Angeles, advised by Bolei Zhou. I am studying for PhD degree in Zhejiang University since 2021, where I was a member of the APRIL Lab under the guidance of Dr. Yong Liu. I am interning in ADLab at Shanghai AI Laboratory from 2023 to 2025. Prior to this, I got my bachelor’s degree from Jianxing campus of Zhejiang University of Technology in 2021 under the guidance of Prof. Li Yu.

Research Interests

The following are my current research interests:

Autonomous Driving
Emobodied AI
Mobile Robots

News

Feb 20, 2026	🎉 Paper AURA: Multimodal Shared Autonomy for Real-World Urban Navigation is accepted by CVPR 2026 !
Jan 6, 2026	🎉 Paper From Seeing to Experiencing: Scaling Navigation Foundation Models with Reinforcement Learning is accepted by ICLR 2026 !
Dec 20, 2025	🎉 Paper Learning Sidewalk Autopilot from Multi-Scale Imitation with Corrective Behavior Expansion is accepted by ICRA 2026 !
Nov 2, 2025	🎉 Paper LeapVAD: A Leap in Autonomous Driving via Cognitive Perception and Dual-Process Thinking is accepted by TNNLS 2025 !
Aug 2, 2025	🎉 Paper DriveArena: A Closed-loop Generative Simulation Platform for Autonomous Driving is accepted by ICCV 2025 !

Selected Publications

2026

Arxiv

Monocular 3D Occupancy Perception for Robots on Sidewalks via Hybrid 2D-3D Learning

Yukai Ma, Joe Lin, Liu Liu, and 5 more authors

2026

Bib Demo Code

@misc{ma2026monocular3doccupancyperception,
  title = {Monocular 3D Occupancy Perception for Robots on Sidewalks via Hybrid 2D-3D Learning},
  author = {Ma, Yukai and Lin, Joe and Liu, Liu and He, Honglin and Ricketts, Lulu and Squicciarini, Brad and Liu, Yong and Zhou, Bolei},
  year = {2026},
  eprint = {2606.19122},
  archiveprefix = {arXiv},
  primaryclass = {cs.RO},
  demo = {https://vail-ucla.github.io/walkocc/},
}

IROS

SparseWorld: Enhancing End-to-End Autonomous Driving via World Models with Sparse Scene Representation

Ruoyu Wang, Jingke Wang, Yukai Ma, and 5 more authors

2026

arXiv Bib Demo Code

@misc{wang2026sparseworldenhancingendtoendautonomous,
  title = {SparseWorld: Enhancing End-to-End Autonomous Driving via World Models with Sparse Scene Representation},
  author = {Wang, Ruoyu and Wang, Jingke and Ma, Yukai and Huang, Yuehao and Lei, Shuangming and Xu, Guanglin and Ye, Aixue and Liu, Yong},
  year = {2026},
  eprint = {2605.24354},
  archiveprefix = {arXiv},
  primaryclass = {cs.CV},
  demo = {https://wryzju.github.io/SparseWorld/},
}

Arxiv

VEOcc: Voxel-Centric Online Semantic Occupancy Prediction For Embodied Scene Understanding

Ruoyu Wang, Yong Liu, Sheng Tao, and 2 more authors

arXiv preprint arXiv:2605.25059, 2026

Bib Demo Code

@article{wang2026veocc,
  title = {VEOcc: Voxel-Centric Online Semantic Occupancy Prediction For Embodied Scene Understanding},
  author = {Wang, Ruoyu and Liu, Yong and Tao, Sheng and Lin, Yuhang and Ma, Yukai},
  journal = {arXiv preprint arXiv:2605.25059},
  year = {2026},
  demo = {https://wryzju.github.io/VEOcc/},
}

CVPR

AURA: Multimodal Shared Autonomy for Real-World Urban Navigation

Yukai Ma, Honglin He, Selina Song, and 2 more authors

2026

Bib Demo Code

@misc{ma2026auramultimodalsharedautonomy,
  title = {AURA: Multimodal Shared Autonomy for Real-World Urban Navigation},
  author = {Ma, Yukai and He, Honglin and Song, Selina and Wu, Wayne and Zhou, Bolei},
  year = {2026},
  eprint = {2604.01659},
  archiveprefix = {arXiv},
  primaryclass = {cs.RO},
  demo = {https://vail-ucla.github.io/aura/},
}

CVPR

Drive-Cascade: Autoregressive Occupancy to LiDAR and Video Synthesis

Shuangming Lei, Yuehao Huang, Yao Yi, and 8 more authors

In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2026

Bib Demo Code

@inproceedings{lei2026drive,
  title = {Drive-Cascade: Autoregressive Occupancy to LiDAR and Video Synthesis},
  author = {Lei, Shuangming and Huang, Yuehao and Yi, Yao and Xie, Yijia and Wang, Jingke and Wang, Ruoyu and Lv, Jiajun and Xu, Guanglin and Ye, AiXue and Liu, Bingbing and others},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages = {4552--4561},
  year = {2026},
  demo = {https://summersray.github.io/Drive-Cascade/},
}

ICRA

Learning Sidewalk Autopilot from Multi-Scale Imitation with Corrective Behavior Expansion

Honglin He, Yukai Ma, Brad Squicciarini, and 2 more authors

2026

Bib Demo Code

@misc{he2026learningsidewalkautopilotmultiscale,
  title = {Learning Sidewalk Autopilot from Multi-Scale Imitation with Corrective Behavior Expansion},
  author = {He, Honglin and Ma, Yukai and Squicciarini, Brad and Wu, Wayne and Zhou, Bolei},
  year = {2026},
  eprint = {2603.22527},
  archiveprefix = {arXiv},
  primaryclass = {cs.RO},
  demo = {https://vail-ucla.github.io/MIMIC},
}

ICLR

From Seeing to Experiencing: Scaling Navigation Foundation Models with Reinforcement Learning

Honglin He*, Yukai Ma*, Wayne Wu, and 1 more author

Arxiv, 2026

Bib Demo Code

@article{he2025seeingexperiencingscalingnavigation,
  title = {From Seeing to Experiencing: Scaling Navigation Foundation Models with Reinforcement Learning},
  author = {He*, Honglin and Ma*, Yukai and Wu, Wayne and Zhou, Bolei},
  journal = {Arxiv},
  year = {2026},
  publisher = {Arxiv},
  demo = {https://metadriverse.github.io/s2e/},
}

2025

ACMMM

CogDDN: A Cognitive Demand-Driven Navigation with Decision Optimization and Dual-Process Thinking

Yuehao Huang, Liang Liu, Shuangming Lei, and 7 more authors

In Proceedings of the 33rd ACM International Conference on Multimedia, 2025

arXiv Bib Demo

@inproceedings{huang2025cogddn,
  title = {CogDDN: A Cognitive Demand-Driven Navigation with Decision Optimization and Dual-Process Thinking},
  author = {Huang, Yuehao and Liu, Liang and Lei, Shuangming and Ma, Yukai and Su, Hao and Mei, Jianbiao and Zhao, Pengxiang and Gu, Yaqing and Liu, Yong and Lv, Jiajun},
  booktitle = {Proceedings of the 33rd ACM International Conference on Multimedia},
  pages = {5237--5246},
  year = {2025},
  demo = {https://yuehaohuang.github.io/CogDDN/},
}

IROS

L2COcc: Lightweight Camera-Centric Semantic Scene Completion via Distillation of LiDAR Model

Ruoyu Wang, Yukai Ma, Yi Yao, and 5 more authors

2025

arXiv Bib Demo Code

@article{wang2025l2cocc,
  title = {L2COcc: Lightweight Camera-Centric Semantic Scene Completion via Distillation of LiDAR Model},
  author = {Wang, Ruoyu and Ma, Yukai and Yao, Yi and Tao, Sheng and Li, Haoang and Zhu, Zongzhi and Liu, Yong and Zuo, Xingxing},
  booktitle = {Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  year = {2025},
  demo = {https://studyingfufu.github.io/L2COcc/},
}

AAAI

Driving in the occupancy world: Vision-centric 4d occupancy forecasting and planning via world models for autonomous driving

Yu Yang, Jianbiao Mei, Yukai Ma, and 5 more authors

In Proceedings of the AAAI Conference on Artificial Intelligence, 2025

arXiv Bib Demo Code

@inproceedings{yang2025driving,
  title = {Driving in the occupancy world: Vision-centric 4d occupancy forecasting and planning via world models for autonomous driving},
  author = {Yang, Yu and Mei, Jianbiao and Ma, Yukai and Du, Siliang and Chen, Wenqing and Qian, Yijie and Feng, Yuxiang and Liu, Yong},
  booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence},
  volume = {39},
  number = {9},
  pages = {9327--9335},
  year = {2025},
  demo = {https://drive-occworld.github.io/},
}

TNNLS

LeapVAD: A Leap in Autonomous Driving via Cognitive Perception and Dual-Process Thinking

Yukai Ma, Tiantian Wei, Naiting Zhong, and 6 more authors

arXiv preprint arXiv:2501.08168, 2025

Bib Demo Code

@article{ma2025leapvad,
  title = {LeapVAD: A Leap in Autonomous Driving via Cognitive Perception and Dual-Process Thinking},
  author = {Ma, Yukai and Wei, Tiantian and Zhong, Naiting and Mei, Jianbiao and Hu, Tao and Wen, Licheng and Yang, Xuemeng and Shi, Botian and Liu, Yong},
  journal = {arXiv preprint arXiv:2501.08168},
  year = {2025},
  demo = {https://pjlab-adg.github.io/LeapVAD/},
}

2024

RAL

LiCROcc: Teach radar for accurate semantic occupancy prediction using lidar and camera

Yukai Ma, Jianbiao Mei, Xuemeng Yang, and 6 more authors

IEEE Robotics and Automation Letters, 2024

Bib Demo Code

@article{ma2024licrocc,
  title = {LiCROcc: Teach radar for accurate semantic occupancy prediction using lidar and camera},
  author = {Ma, Yukai and Mei, Jianbiao and Yang, Xuemeng and Wen, Licheng and Xu, Weihua and Zhang, Jiangning and Zuo, Xingxing and Shi, Botian and Liu, Yong},
  journal = {IEEE Robotics and Automation Letters},
  year = {2024},
  publisher = {IEEE},
  demo = {https://hr-zju.github.io/LiCROcc/},
}

Preprint
DriveArena: A Closed-loop Generative Simulation Platform for Autonomous Driving

Xuemeng Yang*, Licheng Wen*, Yukai Ma*, and 11 more authors

arXiv preprint arXiv:2408.00415, 2024

Abs arXiv Bib Demo Code

This paper presetns DriveArena, the first high-fidelity closed-loop simulation system designed for driving agents navigating in real scenarios. DriveArena features a flexible, modular architecture, allowing for the seamless interchange of its core components: Traffic Manager, a traffic simulator capable of generating realistic traf- fic flow on any worldwide street map, and World Dreamer, a high-fidelity conditional generative model with infinite autoregression. This powerful synergy empowers any driving agent capable of processing real-world images to navigate in DriveArena simulated environment. The agent perceives its surroundings through images generated by World Dreamer and output trajectories; then these trajectories are fed into Traffic Manager, achieving realistic interactions with other vehicles and producing a new scene lay- out. Finally, the latest scene layout is relayed back into World Dreamer, perpetuating the simulation cycle. This iterative process fosters closed-loop exploration within a highly realistic environment, providing a valuable platform for developing and evaluating driving agents across diverse and challenging scenarios. DriveArena signifies a substantial leap forward in leveraging generative image data for the driving simulatior, opening insights for closed-loop autonomous driving.
@article{yang2024drivearena, title = {DriveArena: A Closed-loop Generative Simulation Platform for Autonomous Driving}, author = {Yang*, Xuemeng and Wen*, Licheng and Ma*, Yukai and Mei*, Jianbiao and Li*, Xin and Wei*, Tiantian and Lei, Wenjie and Fu, Daocheng and Cai, Pinlong and Dou, Min and Shi, Botian and He, Liang and Liu, Yong and Qiao, Yu}, journal = {arXiv preprint arXiv:2408.00415}, year = {2024}, demo = {https://pjlab-adg.github.io/DriveArena/}, }
NeurIPS
Continuously Learning, Adapting, and Improving: A Dual-Process Approach to Autonomous Driving

Jianbiao Mei*, Yukai Ma*, Xuemeng Yang, and 8 more authors

Advances in Neural Information Processing Systems (NeurIPS), 2024

Abs arXiv Bib Demo Code

Autonomous driving has advanced significantly due to sensors, machine learning, and artificial intelligence improvements. However, prevailing methods struggle with intricate scenarios and causal relationships, hindering adaptability and interpretability in varied environments. To address the above problems, we introduce LeapAD, a novel paradigm for autonomous driving inspired by the human cognitive process. Specifically, LeapAD emulates human attention by selecting critical objects relevant to driving decisions, simplifying environmental interpretation, and mitigating decision-making complexities. Additionally, LeapAD incorporates an innovative dual-process decision-making module, which consists of an Analytic Process (System-II) for thorough analysis and reasoning, along with a Heuristic Process (System-I) for swift and empirical processing. The Analytic Process leverages its logical reasoning to accumulate linguistic driving experience, which is then transferred to the Heuristic Process by supervised fine-tuning. Through reflection mechanisms and a growing memory bank, LeapAD continuously improves itself from past mistakes in a closed-loop environment. Closed-loop testing in CARLA shows that LeapAD outperforms all methods relying solely on camera input, requiring 1-2 orders of magnitude less labeled data. Experiments also demonstrate that as the memory bank expands, the Heuristic Process with only 1.8B parameters can inherit the knowledge from a GPT-4 powered Analytic Process and achieve continuous performance improvement.
@article{mei2024continuously, title = {Continuously Learning, Adapting, and Improving: A Dual-Process Approach to Autonomous Driving}, author = {Mei*, Jianbiao and Ma*, Yukai and Yang, Xuemeng and Wen, Licheng and Cai, Xinyu and Li, Xin and Fu, Daocheng and Zhang, Bo and Cai, Pinlong and Dou, Min and others}, journal = {Advances in Neural Information Processing Systems (NeurIPS)}, year = {2024}, demo = {https://leapad-2024.github.io/LeapAD/}, }

TIP

Camera-Based 3D Semantic Scene Completion With Sparse Guidance Network

Jianbiao Mei, Yu Yang, Mengmeng Wang, and 5 more authors

IEEE Transactions on Image Processing, 2024

Bib Code

@article{10694710,
  author = {Mei, Jianbiao and Yang, Yu and Wang, Mengmeng and Zhu, Junyu and Ra, Jongwon and Ma, Yukai and Li, Laijian and Liu, Yong},
  journal = {IEEE Transactions on Image Processing},
  title = {Camera-Based 3D Semantic Scene Completion With Sparse Guidance Network},
  year = {2024},
  volume = {33},
  number = {},
  pages = {5468-5481},
  keywords = {Three-dimensional displays;Semantics;Geometry;Feature extraction;Solid modeling;Proposals;Cameras;Convergence;Autonomous vehicles;Visualization;Semantic scene completion;sparse guidance network;hybrid guidance;voxel aggregation},
  doi = {10.1109/TIP.2024.3461989},
}

ICRA

RadarCam-Depth: Radar-Camera Fusion for Depth Estimation with Learned Metric Scale

Han Li*, Yukai Ma*, Yaqing Gu, and 3 more authors

In 2024 IEEE International Conference on Robotics and Automation (ICRA), 2024

Bib Code

@inproceedings{10610929,
  author = {Li*, Han and Ma*, Yukai and Gu, Yaqing and Hu, Kewei and Liu, Yong and Zuo, Xingxing},
  booktitle = {2024 IEEE International Conference on Robotics and Automation (ICRA)},
  title = {RadarCam-Depth: Radar-Camera Fusion for Depth Estimation with Learned Metric Scale},
  year = {2024},
  volume = {},
  number = {},
  pages = {10665-10672},
  keywords = {Point cloud compression;Image coding;Accuracy;Three-dimensional displays;Robot vision systems;Estimation;Radar},
  doi = {10.1109/ICRA57147.2024.10610929},
}

TITS

RIDERS: Radar-Infrared Depth Estimation for Robust Sensing

Han Li*, Yukai Ma*, Yuehao Huang, and 4 more authors

IEEE Transactions on Intelligent Transportation Systems, 2024

Bib Code

@article{10623522,
  author = {Li*, Han and Ma*, Yukai and Huang, Yuehao and Gu, Yaqing and Xu, Weihua and Liu, Yong and Zuo, Xingxing},
  journal = {IEEE Transactions on Intelligent Transportation Systems},
  title = {RIDERS: Radar-Infrared Depth Estimation for Robust Sensing},
  year = {2024},
  volume = {25},
  number = {11},
  pages = {18764-18778},
  keywords = {Radar;Radar imaging;Estimation;Cameras;Measurement;Laser radar;Accuracy;Autonomous driving;Infrared imaging;Multisensor systems;Depth estimation;radar perception;infrared camera;multi-sensor fusion},
  doi = {10.1109/TITS.2024.3432996},
}

IROS

Monocular Event-Inertial Odometry with Adaptive decay-based Time Surface and Polarity-aware Tracking

Kai Tang, Xiaolei Lang, Yukai Ma, and 4 more authors

In 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2024

Bib

@inproceedings{10802605,
  author = {Tang, Kai and Lang, Xiaolei and Ma, Yukai and Huang, Yuehao and Li, Laijian and Liu, Yong and Lv, Jiajun},
  booktitle = {2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  title = {Monocular Event-Inertial Odometry with Adaptive decay-based Time Surface and Polarity-aware Tracking},
  year = {2024},
  volume = {},
  number = {},
  pages = {12544-12551},
  keywords = {Power demand;Tracking;Dynamics;Cameras;Feature extraction;Robustness;Surface texture;High dynamic range;Odometry;Intelligent robots},
  doi = {10.1109/IROS58592.2024.10802605},
}

JFR

FMCW Radar on LiDAR map localization in structural urban environments

Yukai Ma*, Han Li*, Xiangrui Zhao, and 4 more authors

Journal of Field Robotics, 2024

Bib Code

@article{ma2024fmcw,
  title = {FMCW Radar on LiDAR map localization in structural urban environments},
  author = {Ma*, Yukai and Li*, Han and Zhao, Xiangrui and Gu, Yaqing and Lang, Xiaolei and Li, Laijian and Liu, Yong},
  journal = {Journal of Field Robotics},
  volume = {41},
  number = {3},
  pages = {699--717},
  year = {2024},
  publisher = {Wiley Online Library},
}

ICRA

A Coarse-to-Fine Place Recognition Approach using Attention-guided Descriptors and Overlap Estimation

Chencan Fu, Lin Li, Jianbiao Mei, and 4 more authors

In 2024 IEEE International Conference on Robotics and Automation (ICRA), 2024

Bib Code

@inproceedings{fu2024coarse,
  title = {A Coarse-to-Fine Place Recognition Approach using Attention-guided Descriptors and Overlap Estimation},
  author = {Fu, Chencan and Li, Lin and Mei, Jianbiao and Ma, Yukai and Peng, Linpeng and Zhao, Xiangrui and Liu, Yong},
  booktitle = {2024 IEEE International Conference on Robotics and Automation (ICRA)},
  pages = {8493--8499},
  year = {2024},
  organization = {IEEE},
}

2023

Coco-LIC: Continuous-Time Tightly-Coupled LiDAR-Inertial-Camera Odometry Using Non-Uniform B-Spline

Xiaolei Lang, Chao Chen, Kai Tang, and 4 more authors

IEEE Robotics and Automation Letters, 2023

Bib

@article{10251629,
  author = {Lang, Xiaolei and Chen, Chao and Tang, Kai and Ma, Yukai and Lv, Jiajun and Liu, Yong and Zuo, Xingxing},
  journal = {IEEE Robotics and Automation Letters},
  title = {Coco-LIC: Continuous-Time Tightly-Coupled LiDAR-Inertial-Camera Odometry Using Non-Uniform B-Spline},
  year = {2023},
  volume = {8},
  number = {11},
  pages = {7074-7081},
  keywords = {Trajectory;Splines (mathematics);Laser radar;Visualization;Odometry;Cameras;Sensors;LiDAR-inertial-camera SLAM;localization;sensor fusion;state estimation},
  doi = {10.1109/LRA.2023.3315542},
}

RAL

Geo-Localization With Transformer-Based 2D-3D Match Network

Laijian Li*, Yukai Ma*, Kai Tang, and 5 more authors

IEEE Robotics and Automation Letters, 2023

Bib Code

@article{10168166,
  author = {Li*, Laijian and Ma*, Yukai and Tang, Kai and Zhao, Xiangrui and Chen, Chao and Huang, Jianxin and Mei, Jianbiao and Liu, Yong},
  journal = {IEEE Robotics and Automation Letters},
  title = {Geo-Localization With Transformer-Based 2D-3D Match Network},
  year = {2023},
  volume = {8},
  number = {8},
  pages = {4855-4862},
  keywords = {Laser radar;Point cloud compression;Feature extraction;Three-dimensional displays;Satellites;Location awareness;Global Positioning System;Geo-localization;2D-3D match;SLAM},
  doi = {10.1109/LRA.2023.3290526},
}

ICRA

RoLM: Radar on LiDAR Map Localization

Yukai Ma, Xiangrui Zhao, Han Li, and 3 more authors

In 2023 IEEE International Conference on Robotics and Automation (ICRA), 2023

Bib

@inproceedings{10161203,
  author = {Ma, Yukai and Zhao, Xiangrui and Li, Han and Gu, Yaqing and Lang, Xiaolei and Liu, Yong},
  booktitle = {2023 IEEE International Conference on Robotics and Automation (ICRA)},
  title = {RoLM: Radar on LiDAR Map Localization},
  year = {2023},
  volume = {},
  number = {},
  pages = {3976-3982},
  keywords = {Location awareness;Laser radar;Automation;Autonomous systems;Robot sensing systems;Cameras;Robustness},
  doi = {10.1109/ICRA48891.2023.10161203},
}

2022

Ctrl-VIO: Continuous-Time Visual-Inertial Odometry for Rolling Shutter Cameras

Xiaolei Lang, Jiajun Lv, Jianxin Huang, and 3 more authors

IEEE Robotics and Automation Letters, 2022

Bib

@article{9869309,
  author = {Lang, Xiaolei and Lv, Jiajun and Huang, Jianxin and Ma, Yukai and Liu, Yong and Zuo, Xingxing},
  journal = {IEEE Robotics and Automation Letters},
  title = {Ctrl-VIO: Continuous-Time Visual-Inertial Odometry for Rolling Shutter Cameras},
  year = {2022},
  volume = {7},
  number = {4},
  pages = {11537-11544},
  keywords = {Cameras;Delays;Calibration;Splines (mathematics);Trajectory;Visualization;Robot vision systems;Localization;sensor fusion;visual-inertial SLAM},
  doi = {10.1109/LRA.2022.3202349},
}