Jingxue Jiang

Engineer • Roboticist

I build reliable and scalable software

solutions for robots.

a picture of myself

I'm Jiang, a recent graduate from Queen's University with a B.A.Sc in Computer Engineering. I'm passionate about robotics and intelligent systems.

Over the past few years, I've had the privilege of contributing to many impactful robotics projects, gaining expertise in ways I never thought possible. Yet, I've come to realize that the value lies not in the achievements themselves, but in the people who have supported me along the way - parents who supported me, mentors who have guided me, and peers who have collaborated with me. Through these relationships, I've learned to approach relationships with humility, persist in facing obstacles with grit, and prioritize others’ needs attentively. I'm grateful for the knowledge gained, but these relationships and experiences have meant more to me than any academic or professional achievement.

This website showcases a curated selection of my project portfolio, encompassing both collaborative endeavours and personal initiatives, so welcome!

Please don't hesitate to reach out to me if you'd like to discuss anything in this portfolio further.

Email: 17jj33@queensu.ca

+1 (343) 580-1359

Longitudinal and Lateral Feedback Control of Self-Driving Car

a blue car a silver intel server
Role: Controls Lead
Duration: 14 Months
Applied skills: Feedforward Control Feedback Control Stanley Steering Control Feedback Linearization Kinematic Bicycle Model Dynamic Bicycle Model Controller Tuning
Technologies used: Linux C++ Python ROS2 NovAtel GNSS/GPS

Buttons
Learn more

Open Source Kalman Filter Nonlinear Variants Library

a gif a math figure
Role: Sole Developer
Duration: -
Applied skills: Kalman Filtering Non-linear Filtering State Estimation Vehicle Motion Models Linerization via Taylor Series Euler's Method Fourth Order Runge Kutta EKF UKF Error State Kalman Filter
Technologies used: Python SymPy Multi-Robot Cooperative Localization and Mapping Dataset (UofT)

This Python library implements six major nonlinear variants of the Kalman filter targeted at robotics and autonomous system state estimation. It also implements motion simulations based on several kinematic vehicle motion models, including differential drive, bicycle, and Ackermann kinematics.

Localization performance was evaluated using the Multi-Robot Cooperative Localization and Mapping Dataset from the Autonomous Space Robotics Lab (ASRL) at the University of Toronto Institute for Aerospace Studies.

Novel Fontier-Based Wheeled Robot Autonomous Exploration in Unknown Environment

a robot vacuum with lidar sensor a gif
Role: Sole Developer
Duration: 6 Months
Applied skills: Path Planning A* Search Graph Search Finite State Machine Design Computer Graphics Algoritms Non-linear Optimization
Technologies used: Linux C++ ROS2 Ceres Solver SLAM Toolbox

This planner allows robots to probe the farthest reaches of unknown indoor spaces autonomously, finding their way and creating a map without human intervention. The robot builds the map in real-time, and while it explores, it figures out where it wants to go next.

As the project advanced, I recognized an opportunity to apply this planner to floor coverage planning problems. This insight led to the development of a new project—Systematic Floor Coverage of Unknown Environments for Cleaning Robots, which leverages this frontier-based exploration algorithm to achieve comprehensive floor coverage.

Buttons
Learn more

Systematic Floor Coverage Planning of Unknown Environments for Cleaning Robots

a black robot vacuum a gif
Role: Sole Developer
Duration: 4 Months
Applied skills: Path Planning Coverage Planning Boustrophedon Cell Decomposition Frontier Exploration Dubin's Optimal Path Planner Regulated Pure Pursuit Controller
Technologies used: Linux C++ ROS2 NVIDIA Jetson RPLidar

Buttons
Learn more

Real-Time Semantic Segmentation of Drivable Surface for Robot Navigation

a gif A person testing a computer vision algorithm
Role: Team member (4)
Duration: 4 Months
Applied skills: Semantic Segmentation Real-Time Inference Separable Convolution Neural Network Transfer Learning
Technologies used: Linux C++ Python PyTorch CUDA Computer Vision Annotation Tool(CVAT) DeepLabV3+ ZED RGB-D Camera

Willy the Home Robot

a blue 3d-printed robot the back of a blue 3d-printed robot
Role: Sole Developer
Duration: 8 Months
Applied skills: Proportional–Integral Control Integrator Windup Protection Cascade Control Differential-Drive Kinematics Real-Time Opearating System 3D Modelling 3D Printing CAD
Technologies used: Linux C/C++ NVIDIA Jetson STM32 Microcontroller STM32CubeIDE FreeRTOS Embedded ROS1 SolidWorks Ultimaker Cura

Buttons
Learn more

Model Predictive Control for Trajectory Tracking of Autonomous Vehicles with Genetic Algorithm-assisted Parameter Optimization

a gif a bell shaped curve
Role: Sole Developer
Duration: 4 Months
Applied skills: Model Predictive Control Trajectroy Tracking Polynomial Trajectory Generation Quadratic Programming Unicycle Kinematics Genetic Optimization
Technologies used: MATLAB Robotics System Toolbox Optimization Toolbox Quadprog

Experience

  • GM Logo
    Company
    Sanctuary AI
    Location
    Vancouver, British Columbia, Canada
    Title
    Software Engineer - Humanoid Real-time Systems
    Type
    Full-time
    Date
    Present
    Description
  • GM Logo
    Company
    General Motors & Queen's University
    Location
    Kingston, Ontario, Canada; Michigan, USA
    Title
    Controls Lead - Autonomous Vehicle
    Type
    Contract
    Date
    2022 - 2023
    Description
    Architected, implemented and deployed real-time motion control software for autonomous Chevrolet Bolt EV, enabling autonomous intersection navigation through precise trajectory tracking

© 2024 Copyright Jingxue Jiang. This website was designed and built by Jingxue Jiang.