<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>AutoRobot | Yi Hu</title><link>https://yihu551.github.io/tag/autorobot.html</link><atom:link href="https://yihu551.github.io/tag/autorobot/index.xml" rel="self" type="application/rss+xml"/><description>AutoRobot</description><generator>Wowchemy (https://wowchemy.com)</generator><language>en-us</language><lastBuildDate>Thu, 02 Oct 2025 00:00:00 +0000</lastBuildDate><image><url>https://yihu551.github.io/media/icon_y_hu6062576646913215646.png</url><title>AutoRobot</title><link>https://yihu551.github.io/tag/autorobot.html</link></image><item><title>Fast Inference Diffusion Policy for Robot Ball Catching Task</title><link>https://yihu551.github.io/news/catchball.html</link><pubDate>Thu, 02 Oct 2025 00:00:00 +0000</pubDate><guid>https://yihu551.github.io/news/catchball.html</guid><description>
&lt;video autoplay loop >
&lt;source src="../../news/catchball/catchball.mp4" type="video/mp4">
&lt;/video>
&lt;!-- Research associate on the MicroSep project since September 2025. -->
&lt;p>Developed a diffusion-based imitation learning framework with probabilistic dynamic movement primitives that distills multi-step denoising into a single-step inference process, enabling fast action generation and validated on a robotic ball-catching task.&lt;/p></description></item><item><title>Finetune Vision-Language-Action Model</title><link>https://yihu551.github.io/news/vla_validation.html</link><pubDate>Wed, 01 Oct 2025 00:00:00 +0000</pubDate><guid>https://yihu551.github.io/news/vla_validation.html</guid><description>
&lt;video autoplay loop >
&lt;source src="../../news/vla_validation/pi05_realworld.mp4" type="video/mp4">
&lt;/video>
&lt;!-- Research associate on the MicroSep project since September 2025. -->
&lt;p>Finetune vision-language-model Pi0.5 on real-world robot dataset. Robot Task: grasp the salt box, grasp the black bag&lt;/p></description></item><item><title>Pi0.5 with Spatial Module</title><link>https://yihu551.github.io/news/vla_libero.html</link><pubDate>Tue, 30 Sep 2025 00:00:00 +0000</pubDate><guid>https://yihu551.github.io/news/vla_libero.html</guid><description>
&lt;video autoplay loop >
&lt;source src="../../news/vla_libero/pi05_libero.mp4" type="video/mp4">
&lt;/video>
&lt;p>Vision-language-action model (Pi0.5) integrated with spatial module, designed with causal attention and mixture-of-expert mechanism. Testing the model in Libero Spatial environment.&lt;/p></description></item><item><title>Transfer Human Daily Skills to Surgical Robot Tasks</title><link>https://yihu551.github.io/news/adlsurg.html</link><pubDate>Sat, 27 Sep 2025 00:00:00 +0000</pubDate><guid>https://yihu551.github.io/news/adlsurg.html</guid><description>
&lt;video autoplay loop >
&lt;source src="../../news/adlsurg/ADL.mp4" type="video/mp4">
&lt;/video>
&lt;p>Instead of using specific surgica task dataset, we utilized human daily dataset in learning robotic surgical tasks.&lt;/p>
&lt;p>Patient Side Manipulator Picking Gauze
&lt;video autoplay loop >
&lt;source src="../../news/adlsurg/PSMGauze.mp4" type="video/mp4">
&lt;/video>&lt;/p>
&lt;p>Endoscopic Camera Manipulator Localizing Gauze
&lt;video autoplay loop >
&lt;source src="../../news/adlsurg/ECMStatic.mp4" type="video/mp4">
&lt;/video>&lt;/p>
&lt;p>Endoscopic Camera Manipulator Tracking Surgical Tool
&lt;video autoplay loop >
&lt;source src="../../news/adlsurg/ECMActive.mp4" type="video/mp4">
&lt;/video>&lt;/p></description></item><item><title>Autonomous Endoscopic Camera Manipulator Scanning</title><link>https://yihu551.github.io/news/ecmrandmm.html</link><pubDate>Fri, 26 Sep 2025 00:00:00 +0000</pubDate><guid>https://yihu551.github.io/news/ecmrandmm.html</guid><description>
&lt;video autoplay loop >
&lt;source src="../../news/ecmrandmm/ECMRANDMM.mp4" type="video/mp4">
&lt;/video>
&lt;p>We proposed a label-free adaptive Gaussian sample consensus framework for imitation learning from a noisy demonstration set.&lt;/p>
&lt;p>Please find the detail from &lt;a href="https://www.ece.ualberta.ca/~tbs/pmwiki/pdf/TMRB-Hu-2024.pdf" target="_blank" rel="noopener">Label-Free Adaptive Gaussian Sample Consensus Framework for Learning From Perfect and Imperfect Demonstrations&lt;/a>&lt;/p></description></item><item><title>Learning Pseudo Surgical Skill Training Task</title><link>https://yihu551.github.io/news/ringrailtask.html</link><pubDate>Thu, 25 Sep 2025 00:00:00 +0000</pubDate><guid>https://yihu551.github.io/news/ringrailtask.html</guid><description>
&lt;video autoplay loop >
&lt;source src="../../news/ringrailtask/RingRail.mp4" type="video/mp4">
&lt;/video>
&lt;p>Ring-and-rail task is an experimental platform for training and testing the surgical skill that manipulating the ring get through the wire without any touches. We utilize autunomous ring-and-rail tasks using a noisy demonstration set.&lt;/p>
&lt;p>Please find the detail from &lt;a href="https://www.sciencedirect.com/science/article/pii/S174680942500998X" target="_blank" rel="noopener">Learning from imperfect demonstrations in a surgical training task&lt;/a>&lt;/p></description></item><item><title>Autonomous Ultrasound Scanning</title><link>https://yihu551.github.io/news/ustask.html</link><pubDate>Wed, 24 Sep 2025 00:00:00 +0000</pubDate><guid>https://yihu551.github.io/news/ustask.html</guid><description>
&lt;video autoplay loop >
&lt;source src="../../news/ustask/US.mp4" type="video/mp4">
&lt;/video>
&lt;p>Autonomous ultrasound scanning for breast seroma using imitation learning&lt;/p>
&lt;p>Please find the detail from &lt;a href="https://www.ece.ualberta.ca/~tbs/pmwiki/pdf/Hu-IROS-2023.pdf" target="_blank" rel="noopener">Autonomous ultrasound scanning towards standard plane using interval interaction probabilistic movement primitives&lt;/a>&lt;/p></description></item></channel></rss>