Welcome to my page!
I'm Jeffery (Jingyuan), incoming Algorithm Engineer/Software Developer from China, currently living in London, United Kingdom.
I am deeply passionate about robotics and possess a strong foundation in robot-related knowledge. My coding profiency extends to various aspects of robotics application, including Robotics Software Development, development in Linux environment and ROS (Robot Operating System), which are essential tools in my field.
Besides, I am fervent about machine learning, machine vision, and SLAM (Simultaneous Localization and Mapping). These cutting-edge fields represent the forefront of technological innovation in intelligent systems, and I am committed to exploring and contributing to these domains.
🎁 Projects | 📚 Descriptions |
Robotics Arm / Allegro Hand Simulation with Visuo-tactile Sensing | This project involves the simulation of a robotic arm integrated with an Allegro Hand, focusing on enhancing its capabilities through visuo-tactile sensing. Tactile sensing is enabled with DIGIT sensor and Tacto interface. Leveraging ROS (Robot Operating System) and Pybullet (bullet) engine for real-time simulation, the project showcases the implementation of advanced algorithms for tactile sensing and visual feedback, enabling the robotic hand to perform delicate tasks with precision. |
Panda Robot Autonomous Pick and Place Challenge | I developed a fully autonomous system for a Panda robotic arm to perform pick-and-place tasks with high accuracy and efficiency. Using ROS for system integration, I implemented computer vision techniques to recognize and locate objects, and machine learning algorithms for adaptive grasp planning. This project highlights my skills in robotics automation, software development for robotic control, and the practical application of machine learning in solving complex tasks in dynamic environments. |
Visual Machine Learning for Intelligent Vehicle | This project delves into the core aspects of autonomous driving technologies, focusing on replicating human driving behavior at Level 1 autonomy, enhancing lane detection capabilities, and performing advanced road semantic segmentation. Central to this endeavor is the application of Convolutional Neural Networks (CNNs) and Fully Convolutional Networks (FCNs) for training robust machine learning models capable of end-to-end predictions and driving video inference. Machine vision pipelines and supporting tools to streamline the entire inference process were also developed. |
YOLO Object-Detection Automated Depolyment Framework | This projects develops two pipelins for automating YOLO object detection inference. The first pipeline is to compare the performance of YOLO inference capabilities on Python and C++ends. The second pipeline concentrates on evaluating the performance across various YOLO models, analyzing their accuracy, precision and speed. |
Quadcopter drone simulation | This project entailed creating a simulation of a quadcopter drone in MATLAB, focusing on flight dynamics, control systems, and autonomous navigation simulations. |