Mo Logo [Home] [Lexikon] [Aufgaben] [Tests] [Kurse] [Begleitmaterial] [Hinweise] [Mitwirkende] [Publikationen]

Mathematik-Online-Lexikon:

Beispiel: Robotorarm


A B C D E F G H I J K L M N O P Q R S T U V W X Y Z Übersicht

Die folgende Abbildung zeigt einen zweidimensionalen, zweiarmigen Roboterarm.

\begin{picture}(6.5,10)\linethickness{1.0mm}
\put(2,1){\line(0,1){8}}
\put(2,...
...,1){\oval(2,2)[tr]}
\put(2.3,8.4){$\beta$}
\put(2.3,1.2){$\alpha$}
\end{picture}

Für einen beliebigen erreichbaren Zielpunkt $ P=\left( p_1,p_2 \right)$ sollen die Gelenkwinkel $ \alpha$ und $ \beta$ bestimmt werden. Dazu werden die Koordinaten $ p_i$ als Funktion der Winkel und der Armlängen $ a$ und $ b$ ausgedrückt.

Man erhält zunächst die Koordinaten des Punktes $ Q = (q_1, q_2)$ als

$\displaystyle q_1$ $\displaystyle = a \cos \alpha$    
$\displaystyle q_2$ $\displaystyle = a \sin \alpha$    

und dann

$\displaystyle p_1$ $\displaystyle = q_1 + b \cos (\beta - (\pi - \alpha)) = q_1 + b \cos(\alpha + \beta - \pi)$    
  $\displaystyle = a \cos \alpha - b \cos(\alpha + \beta),$    
$\displaystyle p_2$ $\displaystyle = a \sin \alpha - b \sin(\alpha + \beta).$    

Dieses zweidimensionale lineare Gleichungssystem kann mit dem Newton-Verfahren gelöst werden. Man setzt dazu $ \gamma = \alpha + \beta$ und schreibt das System in der Standardform

$\displaystyle \begin{pmatrix}
f_1\\
f_2
\end{pmatrix} =
\begin{pmatrix}
a \co...
...- b \sin \gamma - p_2
\end{pmatrix} =
\begin{pmatrix}
0\\
0
\end{pmatrix}\,.
$

Mit der Jacobi-Matrix

$\displaystyle f^\prime (\alpha,\gamma)
= \begin{pmatrix}
-a \sin \alpha & b \sin \gamma \\
a \cos \alpha & -b \cos \gamma
\end{pmatrix}$

läuft ein Verfahrensschritt $ (\alpha,\gamma) \to (\alpha\,^\prime,\gamma\,^\prime)$ folgendermaßen ab:

Man löst das lineare Gleichungssystem

$\displaystyle \begin{pmatrix}
-a \sin \alpha & b \sin \gamma \\
a \cos \alpha ...
...s \alpha -b \cos \gamma -p_1\\
a \sin \alpha -b \sin \gamma -p_2
\end{pmatrix}$

und setzt

$\displaystyle \begin{pmatrix}
\alpha^\prime\\
\gamma^\prime
\end{pmatrix}=
\be...
...\end{pmatrix}-
\begin{pmatrix}
\Delta \alpha\\
\Delta \gamma
\end{pmatrix}\,.
$

Der Schritt ist für alle $ ( \alpha , \gamma )$ mit

$\displaystyle \det f^\prime (\alpha, \gamma) = ab ( \sin \alpha \cos \gamma -\c...
...in \gamma) \neq 0 \quad \Longleftrightarrow \quad \tan \alpha \neq \tan \gamma
$

durchführbar, d.h. $ \beta = \gamma - \alpha \neq 0 ,\, \pi$ .
(Autoren: Hager/Wohlmuth)

[Verweise]

  automatisch erstellt am 9.  6. 2006