ZHANG Kaiyan

Education

The University of Hong Kong (HKU) Sept.2023-Now
Master of Science in Artificial Intelligence

Harbin Institute of Technology (HIT) Sept.2017-Jun.2021
Bachelor of Electrical Engineering and Automation

Jul.2021-Aug.2023

Algorithm Development

Conducted research on 3D occupancy prediction
Developed the state estimation based on linear Kalman filter of multi-sensor fusion algorithm in ADAS (Advanced Driver Assistance System), now mass-produced on BYD Han’s cars
Developed the parking space detection algorithm in APA (Auto Parking Assist) system, introduced Associative Embedding and increased the slot recall rate by 8%
Developed the middleware for Horizon Journey 2 and J3 chips, such as image transmission, analysis of code performance etc.

Project Management

Led a collaboration with data annotation services providers, managing transaction over ¥2 million. Responsibilities included researching and selecting partners, and authoring data annotation standards adopted by BYD
Responsible for communicating and coordinating with upstream and downstream business departments on technical requirements

Zhang K. CyliFormer: Cylindrical Transformer for Vision-based 3D Semantic Occupancy Prediction (In Progress)
Zhang K, Feng, X. A Vehicle Classification Method Based on Deep Learning and Data Augmentation. In International Conference on Optics and Machine Vision (ICOMV 2023), Proc. SPIE (Vol.12634, p.1263408). (EI)
Zhang K, Feng, X. A Vehicle Classification Method Based on Improved ResNet. In 2022 2nd Conference on High Performance Computing and Communication Engineering (HPCCE 2022), Proc. SPIE (Vol.12605, p.1260506). (EI)

A Multi-Sensor Fusion-Based Parking Detection Method (Application Number: 202211055726.6)
An Obstacle Detection Method Based on Monocular Camera (Application Number: 202310326277.2)
An Obstacle Detection Method Based on Monocular Camera and Millimeter Wave Radar (Application Number: 202310478355.0)
A Lane Detection Method Based on Monocular Camera (patent pending)
An End-to-end Assisted Driving Method Based on Monocular Camera (patent pending)
An Automatic Pavement Recognition Method Based on Monocular Camera (patent pending)

Jan.2021-Jun.2021

Graduation Project: Designed mathematical model of the motor by MATLAB, implemented the PID algorithm for precise control, and developed embedded software code to achieve optimal performance

Sept.2018-May.2020

As a member of the electronic control group, completed the control codes of robots
As a project manager, planned and coordinated the project progress of mechanical, electronic control, and visual algorithm groups

National Second-Class Award in RoboMaster University Championship Aug.2019
Provincial First-Class Award in North Division RoboMaster University Championship Jun.2019
Social Work Scholarship Jun.2020

Python, C++, C, LaTex, Linux, Shell, MATLAB, Simulink
Proficient in algorithm research and application
Proficient in autonomous driving data workflows, encompassing collection, processing, and annotation
Experienced in embedded terminal deployment