17.2: An Introduction to Autonomous Manipulation

Last updated
Save as PDF

Page ID: 14880

\( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \) \( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)\(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\) \(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\)\(\newcommand{\AA}{\unicode[.8,0]{x212B}}\)

Although robotic manipulation is a much less mature field than autonomous mobile robots, teaching its basics, such as those treated in this book, is slightly easier, mainly due to the fact that concepts like uncertainty and non-holonomy are mostly absent. Robotic manipulation is also well suited for a practicebased curriculum due to the wide array of cheap, multi-DOF robotic arms. These, of course, quickly reach their limitations to demonstrate advanced topics such as dynamics or force control, which are beyond the scope of this book.

17.2.1. Overview

A manipulation-driven curriculum can be motivated by a “grand challenge” task such as robotic agriculture, robotic construction or assisted living, all of which have a manipulation problem at their core. Although a class project is likely to be limited to a toy-example, taking advantage of modern motion-planning frameworks and visualization tools, e.g. ROS/Moveit!, makes it easy to put the class into an industry-relevant framework and expose the students to state of the art platforms in simulation.

17.2.2. Materials

Possible class project range from “robot gardening” or “robots building robots”, for which setups can easily be created. These include real or plastic cherry tomato or strawberry plants and robotic construction kits such as Modular Robotics “Cubelets”, which easily snap together and have the advantage to form structures that are robots themselves, adding additional motivation. The robot arm, such as the open-source, 7-DOF CLAM arm, can be mounted on a portable structure that contains fixed a set of fixed (3D) cameras. In order to allow a large number of students to get familiar with the necessary software and hardware, the instructor can provide a virtual machine with a preinstalled Linux environment and simulation tools. In particular, using the “Robot Operating Systems” (ROS) allows recording so-called “bag”-files of sensor values, including entire sequences of joint recordings and RGB-D video. This allows the students to work on a large part of the homeworks and project preparation from a computer lab or from home, maximizing availability of real hardware.

17.2.3. Content

The first two weeks of this curriculum can be mostly identical to that described in Section E.1.3. If a message passing system such as ROS is used, a good exercise is to record a histogram of message passing times in order to get familiar with the software.

In Chapter 3, the focus is instead on manipulating arms, including the Denavit-Hartenberg scheme and numerical methods for inverse kinematics. In turn, the topics “Forward Kinematics of a Differential Wheels Robot”, and “Inverse Kinematics of Mobile Robots” do not necessarily need to be included. Forward and inverse kinematics can be easily turned into lab sessions using Matlab/Mathematica, simulation or a real robot platform. If the class uses a more complex or industrial robot arm, an alternative path is to record joint trajectories in a ROS bag and letting the students explore this data, e.g., sketching.