Chen (Albert) Feng

I am a Ph.D. student in Dept of Electronic and Computer Engineering (ECE) at The Hong Kong University of Science and Technology, supervised by Prof. Shaojie Shen. Prior to HKUST, I obtained a B.Eng from Harbin Institute of Technology and majored in Mechatronics Engineering (ME) in 2021 in the State Key Laboratory of Robotics and System.

I was a research intern from May. 2021 to Sept. 2021 at Megvii Research of Megvii (Beijing, China). In my undergraduate study, I was an team member of computer vision group and mechanics group in Harbin Institute of Technology Competition Robotics Team (HITCRT), named I Hiter (Harbin, China).

Email / Google Scholar / Github

Education

Ph.D. Student | Electronic and Computer Engineering (ECE), HKUST
Time: 2021 - Present. Supervisor: Prof. Shaojie Shen

B.Eng | Mechatronics Engineering (ME), HIT
Time: 2017 - 2021.

Research

Now I'm Ph.D. student of HKUST Aerial Robotics Group supervised by Prof. Shaojie Shen. I'm interested in robotics and deep learning. Before HKUST, My research worked on lane detection and trajectory prediction in autonomous driving. In the present stage, I focus on autonomous aerial reconstruction/exploration, coverage path planning, and 3D scene understanding collborating with Boyu Zhou.

Publications

(^# for the corresponding author, ^* for equal contributions)

	Star-Searcher: A Complete and Efficient Aerial System for Autonomous Target Search in Complex Unknown Environments Yiming Luo, Zixuan Zhuang, Neng Pan, Chen Feng, Shaojie Shen, Fei Gao, Hui Cheng, and Boyu Zhou^#. [Under Review] Submitted to IEEE Robotics and Automation Letters (RA-L). code / video We propose Star-Searcher, a complete and efficient aerial system for autonomous target search in complex unknown environments. Our aerial system incorporates specialized sensor suites, mapping, and planning modules, all geared towards improving task efficiency and completeness.
	FC-Planner: A Skeleton-guided Planning Framework for Fast Aerial Coverage of Complex 3D Scenes Chen Feng, Haojia Li, Jinqi Jiang, Xinyi Chen, Shaojie Shen, and Boyu Zhou^#. [Under Review] Submitted to IEEE International Conference on Robotics and Automation (ICRA), 2024. Yokohama, Japan. paper / code / video We propose FC-Planner, a skeleton-guided planning framework tailored for fast coverage of large and complex 3D scenes, which results in the generation of high-quality coverage paths and high computational efficiency.
	MASSTAR: A Multi-Modal Large-Scale Scene Dataset with a Versatile Toolchain for Surface Prediction and Completion Jinqi Jiang^, Guiyong Zheng^, Chen Feng^, Shaojie Shen, and Boyu Zhou^#. [Under Review]* Submitted to IEEE International Conference on Robotics and Automation (ICRA), 2024. Yokohama, Japan. code / video We propose MASSTAR, a multi-modal large-scale scene dataset composed of over a thousand collected scene-level 3D data, which could be used for training and testing different learning methods. Additionally, a versatile and highly automatic toolchain is developed to generate a multi-modal dataset only from 3D model sets.
	MacFormer: Map-Agent Coupled Transformer for Real-time and Robust Trajectory Prediction Chen Feng, Hangning Zhou, Huadong Lin, Zhigang Zhang, Ziyao Xu, Chi Zhang, Boyu Zhou^#, and Shaojie Shen. IEEE Robotics and Automation Letters (RA-L), 2023. Presented at ICRA 2024, Yokohama, Japan. paper / video We propose MacFormer, an one-stage Map-Agent Coupled Transformer for real-time and robust trajectory prediction that explicitly incorporates map constraints into the network achieving state-of-the-art performance with significantly lower inference latency and fewer parameters.
	AutoTrans: A Complete Planning and Control Framework for Autonomous UAV Paylaod Transportation Haojia Li, Haokun Wang, Chen Feng, Fei Gao^#, Boyu Zhou^#, and Shaojie Shen. IEEE Robotics and Automation Letters (RA-L), 2023. Presented at ICRA 2024, Yokohama, Japan. paper / code / video We propose AutoTrans, a systematic solution for fully autonomous aerial payload transportation that includes a real-time planning solution to generate smooth trajectories and an adaptive NMPC with a hierarchical disturbance compensation strategy to overcome unknown external perturbations as well as inaccurate model parameters.
	PredRecon: A Prediction-boosted Planning Framework for Fast and High-quality Autonomous Aerial Reconstruction Chen Feng, Haojia Li, Fei Gao, Boyu Zhou^#, and Shaojie Shen. IEEE International Conference on Robotics and Automation (ICRA), 2023. London, UK. paper / code / video / poster We propose PredRecon, a prediction-boosted planning framework that can efficiently reconstruct high-quality 3D models for the target areas in unknown environments with a single flight.
	TENET: Transformer Encoding Network for Effective Temporal Flow on Motion Prediction Yuting Wang^, Hangning Zhou^,#, Zhigang Zhang^, Chen Feng, Huadong Lin, Chaofei Gao, Yizhi Tang, Zhenting Zhao, Shiyu Zhang, Jie Guo, Xuefeng Wang, Ziyao Xu, and Chi Zhang. IEEE / CVF Computer Vision and Pattern Recognition Conference Workshop on Autonomous Driving (CVPRW), 2022. New Orleans, USA. paper We propose TENET* to enhance the trajectory temporal encoding via Temporal Flow Header. Besides, an efficient K-means ensemble method is used. Using our Transformer network and ensemble method, we win the first place of Argoverse 2 Motion Forecasting Challenge with the state-of-the-art brier-minFDE score of 1.90.