CV

Curriculum vitae

Profile

Peiyuan Zhai — PhD Researcher, Autonomous Perception and Sensor Fusion

Delft, The Netherlands · hantyou.zhai@gmail.com

PhD researcher in Electrical Engineering at Delft University of Technology (expected February 2027), building probabilistic perception methods for autonomous vehicles. I develop occupancy grid mapping algorithms that fuse automotive radar, LiDAR, and camera detections using sparse Bayesian learning, and evaluate them on the nuScenes and RADIATE datasets.

Highlights

• Patent family: co-inventor of “Occupancy grid mapping system and method”, published in 2025 as US20250336088A1, EP4641250A1, and CN120846354A.
• Publications: 3 IEEE Sensors Journal articles and 4 conference papers (IEEE SENSORS, ICASSP) on occupancy mapping, sensor fusion, and distributed learning.
• Mentoring: daily supervisor of an MSc thesis that led to a journal article and a conference paper.
• Teaching: co-presented a tutorial on occupancy maps for autonomous driving and robotics at IEEE SENSORS 2026.

Skills

• Perception: occupancy grid mapping, radar and LiDAR processing, multi-sensor fusion, camera-aided priors, dynamic-scene mapping
• Methods: sparse Bayesian learning, Bayesian inference, convex and non-convex optimization, Gaussian processes, distributed estimation (ADMM)
• Tools: Python, C/C++, MATLAB, Linux, Git, LaTeX, FPGA
• Languages: English (fluent), Chinese (native), Dutch (beginner)

Experience

• PhD Researcher — Autonomous Perception, Signal Processing Systems Group, Delft University of Technology — 2023-02 to 2027-02 (expected). Advisor: Dr. Geethu Joseph.
  • Camera-aided occupancy mapping: incorporated machine-learning-based visual detections into the prior of a sparse Bayesian learning mapper (IEEE Sensors Journal 2026; patent applications in the US, Europe, and China).
  • Radar–LiDAR fusion: designed a strategy to softly fuse both sensors for robust mapping (IEEE SENSORS 2024; IEEE Sensors Journal 2025).
  • Dynamic mapping: exploited radar measurements across time frames for dynamic occupancy grid mapping (IEEE SENSORS 2026).
  • Daily supervisor of the MSc thesis “Computationally-Efficient Sparsity-Aware Occupancy Grid Mapping for Automotive Driving” (F. Harraway, graduated September 2025).
• MSc Thesis Researcher — Distributed Gaussian Process Learning, Circuits and Systems Group, Delft University of Technology — 2021-12 to 2022-08. Under Dr. Raj Thilak Rajan. Developed fully-distributed asynchronous algorithms for Gaussian process hyperparameter optimization in multi-agent systems without a fusion center. Published at ICASSP 2023.
• BEng Thesis Researcher — Video Background Modeling, Multi-media Processing Group, Xi’an Jiaotong University — 2019-11 to 2020-06. Developed a selective-update Gaussian Mixture Model for background modeling and designed its implementation on Digital Signal Processor platforms.

Education

• PhD, Electrical Engineering, Delft University of Technology, Delft, The Netherlands — 2023-02 to 2027-02 (expected).
• MSc, Electrical Engineering, Delft University of Technology, Delft, The Netherlands — 2020-09 to 2022-08. GPA 8.32/10. Thesis: “Distributed Gaussian Process for Multi-agent Systems” (8.5/10).
• BEng, Information Engineering, Xi’an Jiaotong University, Xi’an, China — 2016-09 to 2020-07. GPA 86.44/100.

The full publication list is on the Publications page. The PDF below has full details.

Interests

Cycling, cooking, photography, and badminton.