Địa điểm: phòng 803 nhà B1, Đại học Bách khoa HN
Title: Learning and Executing Everyday Task from Instruction Manual and Human Demonstration
Abstract:
Robots have been widely used in industrial applications where they are often pre-programmed in a well-defined and controlled environment. Cognitive robots are becoming more and more skilled in performing human-scale manipulation tasks. However, performing everyday tasks at home demands much knowledge a robot needs to have. The main challenges facing the robots are (1) what actions the robot needs to perform in the task; (2) how to perform each action; (3) perceiving objects (identification, pose, location) for manipulation actions. This presentation focuses on a method for learning and executing everyday manipulation tasks which involve in operating objects or home appliances such as dispensing water from water thermos pot, making a cup of coffee by coffee maker, cooking some food by microwave oven, etc. Firstly, the tasks are planned by the sequences of actions which are automatically acquired from instruction manual. Secondly, the knowledge about how to perform the action is obtained by learning from hand movement in human demonstration using Dynamic Movement Primitive model for generating movement trajectories adapted to new changes. Thirdly, the object 3D information is used for recognizing and estimating 6-DOF pose in order to manipulate that object. To validate the proposed method, an example scenario with task ‘dispensing water’ from a water thermos pot was implemented for the 7-DOF robot arm Schunk LWA3.
Short Bio:
Pham Ngoc Hung got his PhD from Shibaura Institute of Technology, Japan in 2017. He received Master and Bachelor Degrees from Hanoi University of Science and Technology in 2010 and 2008, respectively. His research interests are robotics, intelligent robot systems, embedded system.