Yutong Zhu

Yutong Zhu

M.Sc. student in Control Engineering

Northwestern Polytechnical University

Biography

I am Yutong Zhu, a M.Sc. student in the faculty of Astronautics, Northwestern Polytechnical University. I strive to analyse multi-robot systems through interdisciplinary research on game theory, autonomous decision-making, reachability, and reinforcement learning.

Interests

  • Autonomous decision-making
  • Multi-robot systems
  • Game theory
  • Aerial robotics

Education

  • M.Sc. in Control Engineering, 2022-2025

    Northwestern Polytechnical University

  • B.Sc. in Electrical Engineering, 2018-2022

    Nanjing Normal University

News

  • July 24, 2024. Our paper “FlexNet: an open-formation configuration for cooperative herding in pursuit-evasion games with field of view interactions” has been accepted by IEEE Conference on Decision and Control (CDC). Congratulations! 🎉🎉🎉
  • April 23, 2024. Our paper “A hybrid optimization algorithm for multi-agent dynamic planning with guaranteed convergence in probability” is accepted by Neurocomputing. Congratulations! 😃
  • April 15, 2024. I have successfully obtained the Practice and Innovation Funds for Graduate Students of NPU as the first fund in my life! 🎉🎉🎉
  • December 29, 2023. I have successfully won the Distinguished Graduate Student Award! 🚀
  • November 19, 2023. We have successfully presented our first paper “An Improved Hybrid Grey Wolf Optimizer for Multi-Agent Trajectory Planning in Complex Environment” on the China Automation Congress. 😆

Research Projects

Manoeuvrability Pursuit-Evasion Games for Robots

This project aims at dealing with problems including escape strategy, local optima, and uncertainty for multi-robot high-dimensional data.

Probabilistic Reachability for Hybrid Dynamics System

The probabilistic reachability problem, which involves the computation of probabilistic reachable sets, is studied for stochastic reach-avoid games.

FlexNet for Cooperative Herding

This project performs a novel cooperative herding strategy with an open-formation configuration called “FlexNet” in pursuit-evasion games under limited field of view.

Trajectory Planning with Optimization Algorithm

This project aims to solve the problem of multi-agent path planning in complex environment using optimization algorithm.

List of Publications

FlexNet: an open-formation configuration for cooperative herding in pursuit-evasion games with field of view interactions

We propose an open formation (FlexNet) of pursuers to achieve effective capture of adversarial evaders.
PDF Project
FlexNet: an open-formation configuration for cooperative herding in pursuit-evasion games with field of view interactions

DCFCI: a probabilistic game strategy for stochastic hybrid systems with reachability

A deterministic and stochastic dynamics model for the reach-avoid games is decoupled into a non-cooperative probabilistic game and a cooperative probabilistic game.
PDF Project
DCFCI: a probabilistic game strategy for stochastic hybrid systems with reachability

A finite-time evasion strategy in pursuit-evasion games using multi-agent reinforcement learning

Simulation and experiment results on escape tasks for ground and aerial robots demonstrate the effectiveness and robustness of our method.
PDF Project Video
A finite-time evasion strategy in pursuit-evasion games using multi-agent reinforcement learning

Probabilistic Reachability in Stochasitic Multiplayer Reach-Avoid Games

Our work provides a first attempt to address such a problem in stochastic environment and has shown some preliminary results.
PDF Project
Probabilistic Reachability in Stochasitic Multiplayer Reach-Avoid Games

A game-theoretic cooperative path planning strategy using hybrid heuristic optimization algorithm

The proposed LCD-SCA is employed to identify the optimal strategy of the Stackelberg-Nash game, with the introduction of obstacle avoidance potential function.
PDF Project
A game-theoretic cooperative path planning strategy using hybrid heuristic optimization algorithm

A hybrid optimization algorithm for multi-agent dynamic planning with guaranteed convergence in probability

RSCDWOA-APF is shown to achieve global optimal and smooth trajectory. The proof is amazing.
PDF Project DOI
A hybrid optimization algorithm for multi-agent dynamic planning with guaranteed convergence in probability

A collision-free pursuit-evasion framework for indirect herding and formation control of non-cooperative UAVs

We aim to design a general framework for the pursuit-evasion game of multiple UAVs.
PDF Project
A collision-free pursuit-evasion framework for indirect herding and formation control of non-cooperative UAVs

An Improved Hybrid Grey Wolf Optimizer for Multi-Agent Trajectory Planning in Complex Environment

An improved GWO algorithm based on GWO is proposed by the idea of linear differential decrement and dynamic exponential weighted average.
PDF Project DOI
An Improved Hybrid Grey Wolf Optimizer for Multi-Agent Trajectory Planning in Complex Environment