Max Robotics — Global Service Robot Directory + US Certification

About

Chunxiao Miao is a robotics researcher specializing in vision-based grasping and manipulation, with a focus on enabling robots to interact with novel and texture-less objects. Their core contributions lie in advancing pixel-level grasp detection, moving beyond traditional discrete rectangle-sampling methods that are both time-consuming and prone to missing optimal grasps. Miao’s work introduces deep learning architectures like the Encoder-Decoder-Inception Network (EDINet) to achieve fine-grained, pixel-wise grasp synthesis directly from RGB-D images, significantly improving accuracy and efficiency for unknown objects. They have also pioneered optical-flow-based visual servoing for eye-in-hand robotic control, eliminating the need for system calibration, and developed robust 3D template matching techniques using ICP-enhanced LINEMOD for detecting and grasping texture-less objects. Additionally, Miao has addressed real-world challenges with adaptive noise-filtering structures for monocular vision positioning in noisy workshop environments. With over 29 citations across their top papers, Miao’s research bridges deep learning, computer vision, and control theory, offering practical solutions for industrial automation and service robotics. Their work is particularly notable for moving grasp detection from discrete sampling to continuous, pixel-level reasoning, a paradigm shift that enhances robotic dexterity in unstructured settings.