About Me
-
I am passionate about developing cutting-edge algorithms that drive the future of transportation. With 5 years of experience in this field, I have been involved in various aspects of autonomous driving, including perception, decision-making, and control systems. I have gained expertise in developing and optimizing algorithms for object detection, and classification using sensor data such as LiDAR and cameras. I have also worked on developing advanced deep learning models and reinforcement learning algorithms for prediction and decision-making modules. One of my key strengths is to bridge the gap between research and practical implementation. I enjoy translating state-of-the-art algorithms into robust and efficient solutions that can be deployed in real-world autonomous vehicles. I am skilled in programming languages such as Python, C++, and have experience with popular deep learning frameworks like PyTorch.
-
Collaboration and continuous learning are important aspects of my work. I thrive in dynamic and multidisciplinary teams, where I can contribute my technical expertise and learn from others. I am always eager to explore new technologies and stay updated with the latest advancements in the field of autonomous driving. I am excited about the opportunity to contribute my skills and knowledge to this world. I believe that my strong background in autonomous driving algorithms, combined with my passion for innovation, make me a valuable asset to any team working in this exciting field.
Experience
ECARX Technology
Senior Algorithm Engineer
March 2022 - Present
Developed advanced algorithms for autonomous driving decision-making, perception, and data processing.
- Prediction Algorithm Module (March 2022 - October 2022)
- Contributed as a core member in the development of the vehicle prediction module.
- Developed the vehicle prediction module which takes inputs from obstacle fusion, lane fusion, ego vehicle localization, static object perception (high-definition maps), and previous planned trajectory.
- Implemented obstacle prioritization, intent prediction, and trajectory prediction in the module.
- Responsible for the development of the vehicle-side code, including pre- and post-processing of the models.
- Handled data parsing, filtering, packaging, result validation, sorting, and deletion.
- Implemented the integration of different model results based on different scenarios.
- Parking Obstacle Perception Algorithm Module (Mass Production Project) (October 2022 - Present)
- Core development member responsible for the 2D and 3D object detection model finetuning.
- Adapted the related models for on-vehicle inference and aligned the inference accuracy with the PC-side.
- Implemented post-processing for the 3D model.
- Supported the visualization with DHU module for the central control screen and obstacle avoidance with the PNC module.
- Data Closed-Loop Module (October 2022 - Present)
- Module primarily serving the parking perception mass production project.
- Led the module as the main responsible person.
- Defined data acquisition requirements and tracked the establishment of the data acquisition platform and ground truth system.
- Implemented the end-to-end automation of the data logging, filtering, automatic annotation, labeling, acceptance, and storage processes.
- Leveraged the Speedup ground truth system to generate 3D object detection bounding boxes with centimeter-level accuracy.
- Utilized GroundingDINO and SAM for text-to-2D box and segmentation outputs for large-scale and fast annotation data generation.
State Key Laboratory of Computer Science, Institute of Software, Chinese Academy of Sciences
Algorithm Engineer
March 2021 - February 2022
Worked in the field of autonomous driving decision-making.
- Autonomous driving decision-making
- Conducted research on training vehicles to drive on highways and make correct decisions using deep reinforcement learning methods.
- Traditional decision-making algorithms are typically based on rules and manual feature design, which often exhibit a conservative behavior to ensure safety. Utilized DQN (Deep Q-Network), one of the classic deep reinforcement learning algorithms, to make decisions based on the surrounding environment and ego vehicle information, aiming to maximize the expected reward.
- Data engineering was primarily performed using open-source datasets and simulation environments.
- Achieved an accuracy rate of over 91% in simulated testing, demonstrating the model’s ability to make safe and reasonable decisions in high-speed scenarios.
- Teaching assistant for the “Reinforcement Learning” course at the University of Chinese Academy of Sciences
- Assisted in the development of course materials and experiments.
- Covered topics such as Markov decision processes, Monte Carlo methods, temporal difference learning, eligibility traces, etc.
- Designed unmanned vehicle simulation scenarios using the gym framework for implementing reinforcement learning methods like SARSA, Q-Learning, etc.
- Collaboration with Professor Zhang Lijun
Waytous Intelligent Co., Ltd.
Algorithm Engineer
January 2020 - August 2020
Played a key role in the development and optimization of algorithms for path planning and control in autonomous systems.
- Worked on tasks related to the planning module.
- Added the Jump Point Search algorithm to the algorithm library for global path planning in a freespace scenario in a certain mining area. In scenes with a high number of surrounding obstacles, this algorithm significantly improved search speed while still outputting the optimal path, compared to traditional A* algorithm.
- Adapted the commonly used TEB (Timed Elastic Band) algorithm from ROS (Robot Operating System) to the trajectory optimization library for local trajectory planning.
- Worked on data processing tasks, including constructing trajectory datasets for trajectory clustering, prediction, and other analysis tasks.
- Utilized ETL (Extract, Transform, Load) techniques to process environment perception results from cameras and LiDAR (Light Detection and Ranging), as well as the positioning information from integrated inertial navigation systems.
- Used the zarr library, which supports multiple data formats, to unify the data format and achieve fast data loading.
- Uploaded the dataset to S3 for easy access and future algorithm iteration.
Projects
Open Source Project
December 2020 - February 2021
Development of open-source tutorials related to ROS (Robot Operating System)
- Developed tutorials covering various aspects of ROS, including communication mechanisms such as topics and services.
- Explored core functionality packages, such as the parameter server and TF (Transform) coordinate transformation.
- Analyzed the hardware and software components of robot platforms, including modeling language URDF (Unified Robot Description Format), controllers, and the Gazebo simulator.
- Provided detailed explanations of the ROS Navigation Stack, including nav_core, global_planner, local_planner, move_base, cost_map, AMCL (Adaptive Monte Carlo Localization), and more.
- Published tutorials and corresponding code on platforms like CSDN and GitHub.
Visual Odometry
May 2019 - August 2019
Work related to unmanned aerial vehicle (UAV) visual odometry
- Developed a visual odometry system for UAV localization.
- Constructed local feature maps using keyframes and matched ORB feature points between each frame and the local map.
- Used the PnP (Perspective-n-Point) algorithm to estimate the global pose.
- Designed specifically for scenarios where only short-term motion needs to be considered, such as UAV control, with lower computational requirements compared to full-scale VSLAM systems while meeting the accuracy requirements of the application.
Formula Student Driverless Competition
August 2018 - December 2018
Responsible for the planning and control of the Beijing University of Aeronautics and Astronautics (BUAA) AERO autonomous race car
- Led the planning and control division for the BUAA AERO autonomous race car in the China Formula Student Driverless Competition.
- Managed tasks such as straight-line acceleration, figure-eight maneuvering, and high-speed trajectory following, with high requirements for speed and steering control.
- Utilized PID control algorithms that have been widely validated in engineering applications due to their practicality.
- Given the competition’s focus on high dynamic response, the control module primarily used proportional and differential terms for accurate trajectory tracking, with less emphasis on steady-state error.
Education
Sun Yat-Sen University
2020.08-2020.12
Ph.D. in Computer Science and Technology
- Pursued a Ph.D. in Computer Science and Technology at Sun Yat-sen University, renowned for its academic excellence and research contributions.
- Engaged in cutting-edge research projects related to SLAM and Perception algorithms.
- Developed an in-depth understanding of advanced algorithms and their applications in the field of autonomous systems.
- Although the program was discontinued due to personal circumstances, the short duration allowed for the acquisition of valuable knowledge and research skills.
Beihang University
2017.09-2020.01
M.Sc. in Control Science and Engineering
- Successfully completed a rigorous Master’s program in Control Science and Engineering at Beihang University, a renowned institution known for its excellence in engineering and technical education.
- Engaged in cutting-edge research projects focused on autonomous systems, particularly in the field of planning and control with its applications.
- Actively contributed to the academic community by publishing six scholarly papers in conferences and journals, showcasing expertise in areas such as control algorithms, motion planning, and autonomous navigation.
- Collaborated with esteemed professors and fellow researchers, participating in research discussions, experimental design, and data analysis.
Northeastern University
2011.09-2015.07
B.Sc. in Automation
- Earned a Bachelor’s degree in Automation from Northeastern University.
- Acquired a comprehensive understanding of automation principles and techniques.
- Developed a solid foundation in programming, mathematics, and engineering fundamentals.
- Successfully completed projects involving the design and implementation of control systems.
Skills
- Proficient in C++, Python, ROS, Linux, PyTorch, SQL, Shell, Git, Docker, MongoDB, TensorRT.
- Familiar with autonomous driving algorithms, including reinforcement learning algorithms for prediction and decision-making modules, 2D detection, classification, and segmentation for perception vision modules, and 3D object detection.
- Experienced in deploying algorithm models on different platforms, such as NVIDIA Xavier, Orin, BST A1000, and ECARX Antora1000.
- Knowledgeable in sensors including cameras, lidars, INS (Inertial Navigation Systems), including basic principles, performance evaluation, comparative selection, spatial arrangement, time synchronization, and spatial calibration.
- Proficient in the data closed-loop of autonomous driving, including data acquisition systems, ground truth systems, automated annotation, algorithm evaluation, shadow mode, data feedback, data mining, and data anonymization.
- Hands-on experience in production projects.
- English proficiency: CET 6; IELTS 6.
A Little More About Me
I am an adventurous and free-spirited individual who finds joy in exploring new horizons. Passionate about rock music, I immerse myself in its powerful melodies and captivating rhythms. As an avid band enthusiast, I not only appreciate the artistry behind it but also enjoy actively participating in creating music.
Traveling on a shoestring budget is my preferred way to wander the world. I embrace the thrill of discovering hidden gems, immersing myself in diverse cultures, and forging unforgettable memories along the way.
Being a cinephile, I find solace and inspiration in the realm of cinema. From independent films to blockbuster hits, I am captivated by the art of storytelling and the magic of visual experiences.
My spirit is rebellious and unrestrained, always seeking freedom in expressing myself authentically. I cherish the value of individuality and embrace a lifestyle that encourages self-discovery and personal growth.