23.11.2012 Views

Master Thesis - Hochschule Bonn-Rhein-Sieg

Master Thesis - Hochschule Bonn-Rhein-Sieg

Master Thesis - Hochschule Bonn-Rhein-Sieg

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

2. Overview on human-robot <strong>Master</strong> <strong>Thesis</strong> Björn Ostermann page 14 of 126<br />

2.3 Overview of the project at the BGIA<br />

Prior to this master thesis two project reports have been created [77] [78], which have laid the<br />

foundation of this thesis and their results are included in this report.<br />

This chapter summarizes the complete framework project, consisting of both reports and this thesis.<br />

The overall goal of the complete project is to develop a collaborative workspace, consisting of an<br />

industrial jointed arm robot, a sensor system and a control system that allows a human operator to<br />

work safely in close proximity of the robot.<br />

In the R&D 1 report: “Development of a prototype of an autonomous motion control for assisting<br />

industrial robots” [77] several approaches to this topic were evaluated (see chapter 2.2) and a solution<br />

to the topic was developed.<br />

At the beginning of the R&D 2 project the developed theoretical approach to the above mentioned<br />

solution was discussed on the point of feasibility. An optimal tradeoff between functionality and<br />

feasibility was found, which is briefly described in chapter 2.3.1, along with the different algorithms in<br />

chapter 2.3.2 that are needed for the completion of the project in this master thesis.<br />

The hard- and software, discussed in the R&D 1 report, on which the project is implemented, was<br />

selected in the R&D 2. A description of this hard- and software, as well as its assembly in this master<br />

thesis, can be found in chapter 3.<br />

The main attention of this report is on the designed program that shows how the robot can be<br />

controlled to evade dynamic objects in its working place. Additionally, in chapter 6.4, the performance<br />

of the system is evaluated and compared to an optimal system.<br />

2.3.1 Path planning approach<br />

As described in the previous R&D 1 report [77], there are several modes of safe human-robot<br />

collaboration that can be achieved. The easiest is to separate the human and the robot completely, thus<br />

allowing no cooperation at all. The most difficult is to allow the human and the robot touching each<br />

other, while still executing tasks. To collaborate with a human, robots have to sense their complete<br />

environment and act accordingly, because safety, mainly of the human operator, is always the biggest<br />

issue.<br />

The approach described in the R&D 1 report has been designed to allow a high degree of collaboration<br />

between human and robot, enabling the human to touch and interact with the robot at all times. This<br />

approach not only puts high demands on the required hardware, using for example at least two high<br />

tech 3D cameras, but also needs a lot of computational effort, which in turn puts high demands on the<br />

processing capabilities of the logical unit.<br />

Based on the knowledge, acquired in the first report, a practical approach was found. This approach is<br />

a tradeoff between the functionality of the previously developed algorithm and the simplicity of a<br />

fence. It was develop in the R&D 2 report [78] and can best be described as a flexible fence. Human

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!