[10-Aug @10:30am] Webinar on Tactile Intelligence for Contact-Rich Robotic Manipulation

[10-Aug @10:30am] Webinar on Tactile Intelligence for Contact-Rich Robotic Manipulation Abstract:A robot may identify where to grasp, insert, or press an object, yet still fail in the final millimetres, when contact becomes the dominant source of uncertainty and feedback. Once contact begins, critical physical variables, including force, friction, compliance, slip, deformation, and contact mode, are often hidden from vision and change continuously as the robot acts. Contact-rich manipulation is therefore a problem of sensing, inference, prediction, and adaptation under physical uncertainty.This talk presents my research on tactile intelligence for contact-rich robotic manipulation. I will begin by explaining why physical contact remains a fundamental challenge for robot intelligence, particularly when interactions are partially observable, deformable, dynamic, or safety-critical. I will then introduce a research framework that connects four capabilities: designing tactile bodies that reveal task-relevant information; interpreting contact through physically grounded representations; predicting the consequences of interaction; and using those predictions to guide adaptive action.The technical focus of the talk is programmable tactile intelligence. Rather than treating a tactile sensor as a fixed peripheral device, my work investigates how its morphology, material properties, and integration with the robot can be configured according to task requirements. I will present our latest research on scalable tactile manufacturing, inverse material design, and tactile-driven learning and control. Together, these contributions establish a progression from multimodal sensing to the adaptive control of dexterous and precise interactions, enabling robots to respond to uncertainty and recover from physical disturbances. The broader objective is not simply to build better tactile sensors or task-specific policies, but to establish a unified approach in which embodiment, perception, prediction, and control are designed jointly.The long-term vision is to develop physical AGI: robotic systems that use interaction to build grounded models of the physical world; select informative and safe actions; adapt to changing conditions; diagnose and recover from unexpected events; and transfer knowledge across tasks, embodiments, scales, and application domains.Speaker’s Biography:Prof. Dandan ZHANG is an Assistant Professor at Imperial College London, with appointments in Artificial Intelligence and Machine Learning through the Imperial-X initiative and in Biomedical Robotics within the Department of Bioengineering. She leads the Multi-Scale Embodied Intelligence Lab.Her research explores how robots can perceive, understand and manipulate the physical world across multiple scales through trustworthy, adaptive and physically grounded intelligence. Her work spans tactile sensing, multimodal perception, dexterous and precise contact-rich manipulation, human-robot shared control and digital twins, with applications ranging from biomedical micromanipulation and robot-assisted surgery to assistive and home-care robotics. She also advances the concept of embodied cross-domain intelligence, in which physical, computational, biological and behavioural forms of intelligence are integrated and co-evolve across scales, embodiments and application domains.In 2025, Prof. Zhang was named an AI Visionary: Outstanding Pioneer in AI for Healthcare and received the IEEE/RSJ IROS New Generation Star Award. She has also received the UK Best PhD in Robotics Award, the Amazon PhD Prize and best-paper awards from IEEE ICRA, ICAC and MARSS, alongside several finalist and nomination recognitions, including recognition as a finalist for the IROS Best Student Paper Award and a nomination for the 2026 MARSS Best Conference Paper Award.Prof. Zhang has authored more than 80 journal articles, conference papers, book chapters and preprints. Her work has appeared in leading venues including Science Robotics, Chemical Society Reviews, Communicat Date/Time: 10/08/2026 10:30-11:30 Venue: Online via zoom Organized by: Department of Electrical and Electronic Engineering

Uploaded at 2026/08/09