Liyang Wang

Senior Supervisor R&D Engineer

COMAC Beijing Aircraft Technology Research Institute

Biography

Liyang Wang is a senior supervisor R&D engineer at COMAC Beijing Aircraft Technology Research Institute AI lab. His research interests include robotics, reinforcement learning, deep learning, path planning, motion planning, trajectory optimization, and controller design. His current main research project is flight Four-Dimensional trajectory planning under multiple constraints and optimization objectives. At the same time, he leads the Flight Management System (FMS) group, which develops navigation and guidance module.

Interests

Robotics
Autonomous Driving
Reinforcement Learning
Deep Learning
Planning and Control

Education

PhD in Mechanical and Aerospace Engineering, 2019
Rutgers, The State University of New Jersey
MEng in Electronic Communication Engineering, 2015
University of Chinese Academy of Sciences
BSc in Electronic Engineering, 2012
Beihang University

Skills

Technical

ROS

Python

Pytorch

C++/C

MATLAB

Hobbies

Piano

Swimming

Rabbits

Fishing

Cooking

Experience

Senior Supervisor R&D Engineer

COMAC Beijing Aircraft Technology Research Institute

August 2022 – Present Beijing, China

Responsibilities include:

Supervising team members for Navigation and Guidance module development within FMS (flight management system)
Conducting research for 4D trajectory planning under multiple constraints and optimization objectives
Conducting research for wheeled robot autonomous driving to support aircraft taxiing
Tracking SOTA AI algorithms and architectures

Senior Software Engineer

Baidu USA

December 2019 – June 2022 Sunnyvale, CA, USA

Responsibilities include:

Developing excavator autonomous driving
Conducting research for motion planning and trajectory optimization
Modelling and control of hydraulic system
Developing robot hardware platform
Delivering excavator autonomous solutions to customers

Ph.D Candidate

Rutgers University, Mechanical and Aerospace Engineering

September 2015 – October 2019 Piscataway, NJ, USA

Perform practical and theoretical research in state estimation, sensor fusion, path planning, image processing, machine learning, and control with robotic applications.

Projects

Aircraft Trajectory Planning with Multi-Objectives

This project aims to research and develop a four-dimensional trajectory planning method for aircraft. The trajectory’s constraint conditions include avoidance of no-fly zones and terrain obstacles, collision avoidance with surrounding moving aircraft, trajectory smoothness, and constraints on aircraft capabilities include speed and acceleration. The optimization goals for the trajectory include minimizing the trajectory duration and maximizing comfort.

Wheeled Robot Movement Modeling and Control

This project focuses on the modeling and control of wheeled robots, including three sub-projects I have completed at different times. They are autonomous taxiing of aircraft on the runway, crawler chassis locomotion for excavators, and a robotic guide dog for vision impaired people.

Excavator Autonomous Driving

This project involves a wide range of technologies. My research and works on this has three directions, excavator arm motion planning, trajectory optimization, and control, crawler excavator base motion planning and control, and task-level planning involving decision-making.

Tic Tac Toe

This is a personal project. Aim to learn and apply Tabular Reinforcement Learning on the game of Tic Tac Toe.

Poker (Texas hold’em)

A personal project about developing AI player and playing poker against each other or human player.

Selected Publications

Liyang Wang, Rui Zeng

September, 2023 ICRA 2024

Velocity Planning with Multi-Objectives in Displacement-Time Graphs Using Deep Reinforcement Learning

This paper presents a robust velocity planning method for robotics and autonomous vehicles using deep reinforcement learning, offering scene-independent, efficient, and comfortable performance.

Liangjun Zhang, Jinxin Zhao, Pinxin Long, Liyang Wang, Lingfeng Qian, Feixiang Lu, Xibin Song, Dinesh Manocha

June, 2021 Science Robotics

An autonomous excavator system for material loading tasks

This paper introduces an Autonomous Excavator System (AES) designed for material loading tasks in challenging environments. Combining advanced perception and planning algorithms, AES demonstrates autonomous operation with efficiency comparable to human operators and robust performance in complex scenarios, achieving 24-hour continuous operation without human intervention.

Liyang Wang, Xiaoli Bai

September, 2018 Journal of Intelligent & Robotic Systems

Quadrotor Autonomous Approaching and Landing on a Vessel Deck

This paper presents a monocular vision-based solution for autonomous quadrotor UAV landing on a vessel deck under challenging sea conditions, using on-board sensors only. Experimental results with a Parrot AR.Drone platform confirm the solution’s accuracy and robustness.