www.matheraum.de
Das Matheforum.
Das Matheforum des MatheRaum.

Für Schüler, Studenten, Lehrer, Mathematik-Interessierte.
Hallo Gast!einloggen | registrieren ]
Startseite · Forum · Wissen · Kurse · Mitglieder · Team · Impressum
Forenbaum
^ Forenbaum
Status Mathe
  Status Schulmathe
    Status Primarstufe
    Status Mathe Klassen 5-7
    Status Mathe Klassen 8-10
    Status Oberstufenmathe
    Status Mathe-Wettbewerbe
    Status Sonstiges
  Status Hochschulmathe
    Status Uni-Analysis
    Status Uni-Lin. Algebra
    Status Algebra+Zahlentheo.
    Status Diskrete Mathematik
    Status Fachdidaktik
    Status Finanz+Versicherung
    Status Logik+Mengenlehre
    Status Numerik
    Status Uni-Stochastik
    Status Topologie+Geometrie
    Status Uni-Sonstiges
  Status Mathe-Vorkurse
    Status Organisatorisches
    Status Schule
    Status Universität
  Status Mathe-Software
    Status Derive
    Status DynaGeo
    Status FunkyPlot
    Status GeoGebra
    Status LaTeX
    Status Maple
    Status MathCad
    Status Mathematica
    Status Matlab
    Status Maxima
    Status MuPad
    Status Taschenrechner

Gezeigt werden alle Foren bis zur Tiefe 2

Navigation
 Startseite...
 Neuerdings beta neu
 Forum...
 vorwissen...
 vorkurse...
 Werkzeuge...
 Nachhilfevermittlung beta...
 Online-Spiele beta
 Suchen
 Verein...
 Impressum
Das Projekt
Server und Internetanbindung werden durch Spenden finanziert.
Organisiert wird das Projekt von unserem Koordinatorenteam.
Hunderte Mitglieder helfen ehrenamtlich in unseren moderierten Foren.
Anbieter der Seite ist der gemeinnützige Verein "Vorhilfe.de e.V.".
Partnerseiten
Mathe-Seiten:Weitere Fächer:

Open Source FunktionenplotterFunkyPlot: Kostenloser und quelloffener Funktionenplotter für Linux und andere Betriebssysteme
StartseiteMatheForenUni-Lineare AlgebraRotationsmatrix Herleitung
Foren für weitere Studienfächer findest Du auf www.vorhilfe.de z.B. Astronomie • Medizin • Elektrotechnik • Maschinenbau • Bauingenieurwesen • Jura • Psychologie • Geowissenschaften
Forum "Uni-Lineare Algebra" - Rotationsmatrix Herleitung
Rotationsmatrix Herleitung < Lineare Algebra < Hochschule < Mathe < Vorhilfe
Ansicht: [ geschachtelt ] | ^ Forum "Uni-Lineare Algebra"  | ^^ Alle Foren  | ^ Forenbaum  | Materialien

Rotationsmatrix Herleitung: Frage
Status: (Frage) beantwortet Status 
Datum: 18:29 Di 23.11.2004
Autor: Vielfrager

Ich habe diese Frage in keinem Forum auf anderen Internetseiten gestellt.

Ich möchte & muss die Herleitung einer Rotationsmatrix für eine Drehung im dreidimensionalen Vektorraum verstehen. Meine Frage im Klartext also:

- Wie gelange ich von den mir bekannten Rotationsmatrizen für Rotationen des Vektors [mm]\begin{pmatrix} x & y & z & 1 \end{pmatrix}[/mm]  um die X-,Y- und Z-Achse [mm]R_x, R_y[/mm] und [mm]R_z[/mm]...

[mm] R_x = \begin{pmatrix} 1 & 0 & 0 & 0 \\ 0 & \cos \alpha & \sin \alpha & 0 \\ 0 & -\sin \alpha & \cos \alpha & 0 \\ 0 & 0 & 0 & 1 \end{pmatrix} [/mm]

[mm] R_y = \begin{pmatrix} \cos \alpha & 0 & -\sin \alpha & 0 \\ 0 & 1 & 0 & 0 \\ \sin \alpha & 0 & \cos \alpha & 0 \\ 0 & 0 & 0 & 1 \end{pmatrix} [/mm]

[mm] R_z = \begin{pmatrix} \cos \alpha & \sin \alpha & 0 & 0 \\ -\sin \alpha & \cos \alpha & 0 & 0 \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & 1 \end{pmatrix} [/mm]

...zur folgenden Matrixgleichung für die Drehung um eine beliebige Achse:

[mm] (1-\cos \alpha) \begin{pmatrix} x^2 & xy & xz & 0 \\ xy & y^2 & yz & 0 \\ xz & yz & z^2 & 0 \\ 0 & 0 & 0 & 1 \end{pmatrix} +\cos \alpha \begin{pmatrix} 1 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & 1 \end{pmatrix} -\sin \alpha \begin{pmatrix} 0 & -z & y & 0 \\ z & 0 & -x & 0 \\ -y & x & 0 & 0 \\ 0 & 0 & 0 & 1 \end{pmatrix} [/mm]

wobei [mm]x,y[/mm] und [mm]z[/mm] die jeweilige Komponente der normierten Rotationsachse und [mm]\alpha[/mm] den Drehwinkel kennzeichnet.

        
Bezug
Rotationsmatrix Herleitung: Antwort
Status: (Antwort) fertig Status 
Datum: 16:41 Sa 27.11.2004
Autor: Stefan

Hallo Daniel!

Wie man aus den Standarddrehungen direkt auf die Darstellung kommt, weiß ich nicht, aber ich kann dir das Ganze ein wenig plausibilisieren. Die vierte Dimension lasse ich weg, denn ich sehe jetzt nicht, warum du die eingeführt hast (damit passiert ja gar nichts).

Für einen nomrierten Vektor $(x,y,z) [mm] \in \IR^3$ [/mm] und einen Winkel [mm] $\alpha$ [/mm] definiert man eine $(3 [mm] \times [/mm] 3)$-Matrix $D$ durch die Abbildungsvorschrift:

(*) [mm] $D(\alpha)w:= [/mm] w + [mm] (\sin(\alpha)) \cdot [(x,y,z)^T \times [/mm] w] + (1- [mm] \cos(\alpha)) \cdot [(x,y,z)^T \times ((x,y,z)^T \times [/mm] w)]$.

Wegen [mm] $(x,y,z)^T \times (x,y,z)^T [/mm] =0$ folgt:

[mm] $[D(\alpha)]((x,y,z)^T) [/mm] = [mm] (x,y,z)^T$, [/mm]

d.h. [mm] $\IR [/mm] (x,y,z)$ ist eine Fixgerade von [mm] $D(\alpha)$. [/mm]

Weiterhin gilt natürlich

[mm] $D(\alpha) [/mm] = [mm] D(\alpha [/mm] + [mm] 2\pi)$. [/mm]

Mit der Grassmann-Identität

($a [mm] \times( [/mm] b [mm] \times [/mm] c) [mm] =\langle [/mm] a,c [mm] \rangle [/mm] b - [mm] \langle [/mm] a,b [mm] \rangle [/mm] c$)

kann man (*) auch so schreiben:

[mm] $D(\alpha)w= [/mm] w + [mm] (\sin(\alpha)) \cdot [(x,y,z)^T \times [/mm] w] + (1- [mm] \cos(\alpha)) \cdot ((x,y,z)^T \times ((x,y,z)^T \times [/mm] w))$

$= w + [mm] (\sin(\alpha)) \cdot [(x,y,z)^T \times [/mm] w] + (1- [mm] \cos(\alpha)) \cdot [/mm] [ [mm] \langle (x,y,z)^T, [/mm] w [mm] \rangle [/mm] (x,y,z)T - [mm] \underbrace{\langle (x,y,z)^T,(x,y,z)^T \rangle}_{=\, 1} [/mm] w]$

$= w + [mm] \sin(\alpha) \cdot [(x,y,z)^T \times [/mm] w] + (1- [mm] \cos(\alpha)) \cdot \langle (x,y,z)^T, [/mm] w [mm] \rangle (x,y,z)^T [/mm] - w + [mm] \cos(\alpha) [/mm] w$

$= [mm] \sin(\alpha) \cdot [(x,y,z)^T \times [/mm] w] + (1- [mm] \cos(\alpha)) \cdot \langle (x,y,z)^T, [/mm] w [mm] \rangle (x,y,z)^T [/mm]  + [mm] \cos(\alpha) [/mm] w$.

Überlege dir durch Einsetzen der Einheitsvektoren [mm] $e_i$ [/mm] jetzt mal, wie die Matrixdarstellung von [mm] $D(\alpha)$ [/mm] aussieht, dann siehst du, dass man genau deine Darstellung erhält.

Jetzt muss man sich noch davon überzeugen, dass [mm] $D(\alpha)$ [/mm] tatsächlich eine Drehung, also aus $O(3)$, ist.

Dazu zeigt man (das ist leicht):

1) [mm] $D(\alpha)\, D(\beta) [/mm] = [mm] D(\alpha [/mm] + [mm] \beta)$, [/mm]
2) [mm] $[D(\alpha)]^T [/mm] = [mm] D(-\alpha)$. [/mm]

Aus 1) folgt zunächst wegen

[mm] $D(\alpha)\, D(-\alpha) [/mm] = D(0) = [mm] E_3$ [/mm]

die Beziehung

[mm] $[D(\alpha)]^{-1} [/mm] = [mm] D(-\alpha)$ [/mm]

und daraus zusammen mit 2)

[mm] $[D(\alpha)]^T [/mm] = [mm] [D(\alpha)]^{-1}$. [/mm]

Demnach ist [mm] $D(\alpha)$ [/mm] in der Tat orthogonal, also eine Drehung, und der Vektor [mm] $(x,y,z)^T$ [/mm] wird (siehe oben!) festgehalten.

Die Matrix leistet also alles, was wir uns von ihr wünschen!! [sunny]

Liebe Grüße
Stefan


Bezug
Ansicht: [ geschachtelt ] | ^ Forum "Uni-Lineare Algebra"  | ^^ Alle Foren  | ^ Forenbaum  | Materialien


^ Seitenanfang ^
www.matheforum.net
[ Startseite | Forum | Wissen | Kurse | Mitglieder | Team | Impressum ]