Peiyao Zhang

I'm a PhD candidate in Mechanical Engineering at Johns Hopkins University (JHU), advised by Prof. Marin Kobilarov and Prof. Iulian Iordachita. My current research focuses on robot-assisted eye surgery. I'm interested in achieving surgical robot autonomy using computer vision and deep learning methods.

Previously, I obtained my second Master's degree in Mechanical Engineering at JHU, advised by Marin Kobilarov. I obtained both my Bachelor's and first Master's degree in Naval Architecture and Ocean Engineering at Harbin Engineering University.

Email: pzhang24@jhu.edu

Scholar  |  LinkedIn  |  Github

News

[Jan 2025] - One paper was accepted by ICRA 2025.
[Apr 2024] - One paper was accepted by EMBC 2024.
[Feb 2024] - One paper was accepted by RA-L.
[Jan 2024] - One paper was accepted by ICRA 2024.
[Jan 2024] - One paper was accepted by RA-L.
[Oct 2023] - Our poster received the 1st Place Best Poster Award at DMMR Workshop at IROS 2023. [Poster][Photo]
[Apr 2023] - One abstract was accepted by ICRA 2023 Workshop: New Evolutions in Surgical Robotics. [Poster]
[Jan 2023] - Two papers were accepted by ICRA 2023.
[Oct 2022] - One paper was accepted by BIBM 2022.
[Jan 2022] - One paper was accepted by BioMed 2022.
[July 2021] - One paper was accepted by CDC 2021.
[Oct 2020] - One paper was accepted by CoRL 2020.

Selected Publications

In this study, we present a feasible workflow for Retinal Vein Cannulation (RVC) using 100μm beveled metal needle in ex vivo pig eyes.

This work combines the Steady Hand Eye Robot (SHER) with intraoperative Optical Coherence Tomography (iOCT) to develop a workflow for autonomous needle navigation for subretinal injection.

We demonstrate a real-time autonomous system and workflow for subretinal injection. This is enabled by the global view provided by microscope images and dynamically-aligned B-scan images that track the needle axis for real-time depth feedback.

This work develops a strategy for autonomous needle navigation in retinal microsurgery aiming to achieve precise manipulation, reduced end-to-end surgery time, and enhanced safety. This is accomplished through real-time geometry estimation and chance-constrained Model Predictive Control (MPC) resulting in high positional accuracy while keeping scleral forces within a safe level.

We present a framework that combines real-time eye geometry estimation and chance-constrained optimal control to bound the probability for tissue damage during autonomous robotic navigation. A neural network is trained to predict the relative location of 3-D points on the retina with respect to the current tool-tip position through expert demonstrations.

In this work, we demonstrate that autonomous navigation is possible by learning to predict goal positions on the retina from user-specified image pixel locations. Our work has been tested in simulated eyes as well as eye phantom models where eye movement, different textures, refractive view, and various lighting conditions are encountered.

Website template adapted from Jon Barron.