Grasp Planning for Interactive Object Manipulation

Abstract

Abstract — Nowadays, one important research area of robotics is the design of machines with capabilities to interact with humans or other robots. Service robots fall in this category of machines and they need a set of functions to accomplish their tasks. Future autonomous robots must be capable to work and take decisions in an automatic way. Such a system must be capable to handle most common objects in a dynamic human environment. The first problem to solve in every manipulation task is how to grasp the object under certain constraints. This paper presents a planner to grasp unknown arbitrary objects for interactive manipulation tasks.