Ph.D. Preliminary: Xiaoqian Mu

Xiaoqian Mu
Wednesday, October 14, 2020 - 9:00am
Event Type: 

Dexterous Robotic Cutting  

Skills of cutting natural foods are important for robots looking to play a bigger role in kitchen assistance. Issues in dexterous robotic cutting include knife holding and control, force modeling, two-hand coordination, etc.  In the first step of our investigation, we simplify this difficult task by mounting a kitchen knife rigidly on a robotic arm and by ignoring object deformations during a cutting action. Performance of cutting is measured by both speed and smoothness. Inspired by the smooth cutting skills of the human hand, we divide the cutting action into three sequential phases: pressing, touching, and slicing. In the pressing phase, the knife moves towards the cutting board until contact establishment. During this period, position control is applied to follow some prescribed knife trajectory. In the touching phase, impedance control takes charge to soften the impact between the knife and the cutting board. Due to the knife’s curved edge, the object has yet to be completely split into two pieces. In the slicing phase, the knife separates the object by moving the knife-board contact across the object’s bottom, while maintaining a desired level of force on the board. This knife-board contact force is extracted from the robotic arm’s force sensor reading via modeling based on fracture mechanics. Slicing is carried out under hybrid position/force control.  Experiments over several types of fruits and vegetables have been conducted to demonstrate smooth knife movements highly resembling those performed by the human hand.

Committee: Yan-bin Jia (major professor), Jin Tian, Alexander T. Stoytchev, James H. Oliver, and David Fernandez-Baca


Video-conference links are located in Inside Atansoff, our bi-monthly department e-newsletter. If you cannot locate the link, please email