Transformationen
This commit is contained in:
33
Vorbereitungen_Fabian/Tests.ipynb
Normal file
33
Vorbereitungen_Fabian/Tests.ipynb
Normal file
@@ -0,0 +1,33 @@
|
||||
{
|
||||
"cells": [
|
||||
{
|
||||
"metadata": {},
|
||||
"cell_type": "code",
|
||||
"outputs": [],
|
||||
"execution_count": null,
|
||||
"source": "",
|
||||
"id": "32199e6b9a0ba953"
|
||||
}
|
||||
],
|
||||
"metadata": {
|
||||
"kernelspec": {
|
||||
"display_name": "Python 3",
|
||||
"language": "python",
|
||||
"name": "python3"
|
||||
},
|
||||
"language_info": {
|
||||
"codemirror_mode": {
|
||||
"name": "ipython",
|
||||
"version": 2
|
||||
},
|
||||
"file_extension": ".py",
|
||||
"mimetype": "text/x-python",
|
||||
"name": "python",
|
||||
"nbconvert_exporter": "python",
|
||||
"pygments_lexer": "ipython2",
|
||||
"version": "2.7.6"
|
||||
}
|
||||
},
|
||||
"nbformat": 4,
|
||||
"nbformat_minor": 5
|
||||
}
|
||||
@@ -78,18 +78,32 @@ for punkt in liste_Punkte:
|
||||
|
||||
|
||||
|
||||
dX, dY, dZ, m, q0, q1, q2, q3, xp1, yp1, zp1 = sp.symbols('dX dY dZ m q0 q1 q2 q3 xp1 yp1 zp1')
|
||||
dX, dY, dZ, m, q0, q1, q2, q3, xp1, yp1, zp1, xp2, yp2, zp2, xp3, yp3, zp3, xp4, yp4, zp4, xp5, yp5, zp5 = sp.symbols('dX dY dZ m q0 q1 q2 q3 xp1 yp1 zp1 xp2 yp2 zp2 xp3 yp3 zp3 xp4 yp4 zp4 xp5 yp5 zp5')
|
||||
|
||||
#print(Translation[0])
|
||||
|
||||
zahlen = {dX: Translation[0], dY: Translation[1], dZ: Translation[2], m: m0, q0: q0_wert, q1: q1_wert, q2: q2_wert, q3: q3_wert, xp1: p1[0], yp1: p1[1], zp1: p1[2]}
|
||||
zahlen = {dX: Translation[0], dY: Translation[1], dZ: Translation[2], m: m0, q0: q0_wert, q1: q1_wert, q2: q2_wert, q3: q3_wert, xp1: p1[0], yp1: p1[1], zp1: p1[2], xp2: p2[0], yp2: p2[1], zp2: p2[2], xp3: p3[0], yp3: p3[1], zp3: p3[2], xp4: p4[0], yp4: p4[1], zp4: p4[2], xp5: p5[0], yp5: p5[1], zp5: p5[2]}
|
||||
|
||||
#print(zahlen[zp1])
|
||||
|
||||
f = sp.Matrix(
|
||||
[[dX + m * (xp1 * (1 - 2 * (q2**2 + q3**2)) + yp1 * (2 * (q1 * q2 - q0 * q3)) + zp1 * (2 * (q0 * q2 - q1 * q3)))],
|
||||
[[dX + m * (xp1 * (1 - 2 * (q2**2 + q3**2)) + yp1 * (2 * (q1 * q2 - q0 * q3)) + zp1 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp1 * (2 * (q1 * q2 + q0 * q3)) + yp1 * (1 - 2 * (q1**2 + q3**2)) + zp1 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp1 * (2 * (q1 * q3 - q0 * q2)) + yp1 * (2 * (q0 * q1 + q2 * q3)) + zp1 * (1 - 2 * (q1**2 + q2**2)))]
|
||||
[dZ + m * (xp1 * (2 * (q1 * q3 - q0 * q2)) + yp1 * (2 * (q0 * q1 + q2 * q3)) + zp1 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
[dX + m * (xp2 * (1 - 2 * (q2**2 + q3**2)) + yp2 * (2 * (q1 * q2 - q0 * q3)) + zp2 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp2 * (2 * (q1 * q2 + q0 * q3)) + yp2 * (1 - 2 * (q1**2 + q3**2)) + zp2 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp2 * (2 * (q1 * q3 - q0 * q2)) + yp2 * (2 * (q0 * q1 + q2 * q3)) + zp2 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
[dX + m * (xp3 * (1 - 2 * (q2**2 + q3**2)) + yp3 * (2 * (q1 * q2 - q0 * q3)) + zp3 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp3 * (2 * (q1 * q2 + q0 * q3)) + yp3 * (1 - 2 * (q1**2 + q3**2)) + zp3 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp3 * (2 * (q1 * q3 - q0 * q2)) + yp3 * (2 * (q0 * q1 + q2 * q3)) + zp3 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
[dX + m * (xp4 * (1 - 2 * (q2**2 + q3**2)) + yp4 * (2 * (q1 * q2 - q0 * q3)) + zp4 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp4 * (2 * (q1 * q2 + q0 * q3)) + yp4 * (1 - 2 * (q1**2 + q3**2)) + zp4 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp4 * (2 * (q1 * q3 - q0 * q2)) + yp4 * (2 * (q0 * q1 + q2 * q3)) + zp4 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
[dX + m * (xp5 * (1 - 2 * (q2**2 + q3**2)) + yp5 * (2 * (q1 * q2 - q0 * q3)) + zp5 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp5 * (2 * (q1 * q2 + q0 * q3)) + yp5 * (1 - 2 * (q1**2 + q3**2)) + zp5 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp5 * (2 * (q1 * q3 - q0 * q2)) + yp5 * (2 * (q0 * q1 + q2 * q3)) + zp5 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
|
||||
|
||||
]
|
||||
)
|
||||
|
||||
@@ -100,8 +114,28 @@ A = A_ohne_zahlen.subs(zahlen)
|
||||
#print(J_zahlen.evalf(n=3))
|
||||
|
||||
# Parameterschätzung
|
||||
P = sp.eye(3)
|
||||
schwellenwert = 1e-3
|
||||
alle_kleiner = True
|
||||
|
||||
x0 = sp.Matrix([zahlen[dX], zahlen[dY], zahlen[dZ], zahlen[m], zahlen[q0], zahlen[q1], zahlen[q2], zahlen[q3]])
|
||||
P = sp.eye(15)
|
||||
N = A.T * P * A
|
||||
l = sp.Matrix([p1[0], p1[1], p1[2]])
|
||||
l = sp.Matrix([p1[0], p1[1], p1[2], p2[0], p2[1], p2[2], p3[0], p3[1], p3[2], p4[0], p4[1], p4[2], p5[0], p5[1], p5[2]])
|
||||
# ToDo: Prüfen, ob n mit l oder mit dl!
|
||||
n = A.T * P * l
|
||||
print(n.evalf(n=3))
|
||||
Qxx = N.evalf(n=30).inv()
|
||||
dx = Qxx * n
|
||||
x = x0 + dx
|
||||
|
||||
|
||||
print(x0.evalf(n=3))
|
||||
print(dx.evalf(n=3))
|
||||
print(x.evalf(n=3))
|
||||
|
||||
for i in range (dx.rows):
|
||||
wert = float(dx[i])
|
||||
if abs(wert) >= schwellenwert:
|
||||
print("Warnung")
|
||||
alle_kleiner = False
|
||||
|
||||
if alle_kleiner: print("Ausgleichung fertig!")
|
||||
155
Vorbereitungen_Fabian/Transformation_Helmert_V2.py
Normal file
155
Vorbereitungen_Fabian/Transformation_Helmert_V2.py
Normal file
@@ -0,0 +1,155 @@
|
||||
import sympy as sp
|
||||
from sympy.algebras.quaternion import Quaternion
|
||||
|
||||
#ToDo: Achtung: Die Ergebnisse sind leicht anders, als in den Beispielrechnung von Luhmann (Rundungsfehler bei Luhmann?)
|
||||
#ToDo: Automatische Ermittlung der Anzahl Nachkommastellen für Test auf Orthonormalität integrieren!
|
||||
#Beipsiel aus Luhmann S. 76
|
||||
# Ausgangssystem
|
||||
p1 = sp.Matrix([110, 100, 110])
|
||||
p2 = sp.Matrix([150, 280, 100])
|
||||
p3 = sp.Matrix([300, 300, 120])
|
||||
p4 = sp.Matrix([170, 100, 100])
|
||||
p5 = sp.Matrix([200, 200, 140])
|
||||
|
||||
# Zielsystem
|
||||
P1 = sp.Matrix([153.559, 170.747, 150.768])
|
||||
P2 = sp.Matrix([99.026, 350.313, 354.912])
|
||||
P3 = sp.Matrix([215.054, 544.420, 319.003])
|
||||
P4 = sp.Matrix([179.413, 251.030, 115.601])
|
||||
P5 = sp.Matrix([213.431, 340.349, 253.036])
|
||||
|
||||
#1) Näherungswertberechnung
|
||||
m0 = (P2 - P1).norm() / (p2 - p1).norm()
|
||||
|
||||
U = (P2 - P1) / (P2 - P1).norm()
|
||||
W = (U.cross(P3 - P1)) / (U.cross(P3 - P1)).norm()
|
||||
V = W.cross(U)
|
||||
|
||||
u = (p2 - p1) / (p2 - p1).norm()
|
||||
w = (u.cross(p3 - p1)) / (u.cross(p3 - p1)).norm()
|
||||
v = w.cross(u)
|
||||
|
||||
R = sp.Matrix.hstack(U, V, W) * sp.Matrix.hstack(u, v, w).T
|
||||
|
||||
XS = (P1 + P2 + P3) / 3
|
||||
xS = (p1 + p2 + p3) / 3
|
||||
|
||||
Translation = XS - m0 * R * xS
|
||||
|
||||
|
||||
#print(m0.evalf())
|
||||
#print(R.evalf())
|
||||
#print(Translation.evalf())
|
||||
|
||||
# 2) Test auf orthonormale Drehmatrix bei 3 Nachkommastellen!
|
||||
if R.T.applyfunc(lambda x: round(float(x), 3)) == R.inv().applyfunc(lambda x: round(float(x), 3)) and (R.T * R).applyfunc(lambda x: round(float(x), 3)) == sp.eye(3).applyfunc(lambda x: round(float(x), 3)) and ((round(R.det(), 3) == 1.000 or round(R.det(), 3) == -1.000)):
|
||||
print("R ist Orthonormal!")
|
||||
else:
|
||||
print("R ist nicht Orthonormal!")
|
||||
|
||||
# Testmatrix R aus Luhmann S. 66
|
||||
#R = sp.Matrix([
|
||||
# [0.996911, -0.013541, -0.077361],
|
||||
# [0.030706, 0.973820, 0.225238],
|
||||
# [0.072285, -0.226918, 0.971228]
|
||||
#])
|
||||
|
||||
# 3) Quaternionen berechnen
|
||||
# ToDo: Prüfen, ob Vorzeichen bei q0 richtig ist!
|
||||
#ToDo: q0 stimmt nicht mit Luhmann überein!
|
||||
|
||||
q = Quaternion.from_rotation_matrix(R)
|
||||
q0_wert = q.a
|
||||
q1_wert = q.b
|
||||
q2_wert = q.c
|
||||
q3_wert = q.d
|
||||
|
||||
|
||||
|
||||
# 4) Funktionales Modell
|
||||
liste_Punkte = ["P1", "P2", "P3", "P4", "P5"]
|
||||
liste_unbekannte = ["dX", "dY", "dZ", "dm", "dq0", "dq1", "dq2", "dq3"]
|
||||
liste_beobachtungen =[]
|
||||
for punkt in liste_Punkte:
|
||||
liste_beobachtungen.append(f"X_{punkt}")
|
||||
liste_beobachtungen.append(f"Y_{punkt}")
|
||||
liste_beobachtungen.append(f"Z_{punkt}")
|
||||
|
||||
|
||||
|
||||
|
||||
dX, dY, dZ, m, q0, q1, q2, q3, xp1, yp1, zp1, xp2, yp2, zp2, xp3, yp3, zp3, xp4, yp4, zp4, xp5, yp5, zp5 = sp.symbols('dX dY dZ m q0 q1 q2 q3 xp1 yp1 zp1 xp2 yp2 zp2 xp3 yp3 zp3 xp4 yp4 zp4 xp5 yp5 zp5')
|
||||
|
||||
#print(Translation[0])
|
||||
|
||||
zahlen = {dX: Translation[0], dY: Translation[1], dZ: Translation[2], m: m0, q0: q0_wert, q1: q1_wert, q2: q2_wert, q3: q3_wert, xp1: p1[0], yp1: p1[1], zp1: p1[2], xp2: p2[0], yp2: p2[1], zp2: p2[2], xp3: p3[0], yp3: p3[1], zp3: p3[2], xp4: p4[0], yp4: p4[1], zp4: p4[2], xp5: p5[0], yp5: p5[1], zp5: p5[2]}
|
||||
|
||||
#print(zahlen[zp1])
|
||||
|
||||
f = sp.Matrix(
|
||||
[[dX + m * (xp1 * (1 - 2 * (q2**2 + q3**2)) + yp1 * (2 * (q1 * q2 - q0 * q3)) + zp1 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp1 * (2 * (q1 * q2 + q0 * q3)) + yp1 * (1 - 2 * (q1**2 + q3**2)) + zp1 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp1 * (2 * (q1 * q3 - q0 * q2)) + yp1 * (2 * (q0 * q1 + q2 * q3)) + zp1 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
[dX + m * (xp2 * (1 - 2 * (q2**2 + q3**2)) + yp2 * (2 * (q1 * q2 - q0 * q3)) + zp2 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp2 * (2 * (q1 * q2 + q0 * q3)) + yp2 * (1 - 2 * (q1**2 + q3**2)) + zp2 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp2 * (2 * (q1 * q3 - q0 * q2)) + yp2 * (2 * (q0 * q1 + q2 * q3)) + zp2 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
[dX + m * (xp3 * (1 - 2 * (q2**2 + q3**2)) + yp3 * (2 * (q1 * q2 - q0 * q3)) + zp3 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp3 * (2 * (q1 * q2 + q0 * q3)) + yp3 * (1 - 2 * (q1**2 + q3**2)) + zp3 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp3 * (2 * (q1 * q3 - q0 * q2)) + yp3 * (2 * (q0 * q1 + q2 * q3)) + zp3 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
[dX + m * (xp4 * (1 - 2 * (q2**2 + q3**2)) + yp4 * (2 * (q1 * q2 - q0 * q3)) + zp4 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp4 * (2 * (q1 * q2 + q0 * q3)) + yp4 * (1 - 2 * (q1**2 + q3**2)) + zp4 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp4 * (2 * (q1 * q3 - q0 * q2)) + yp4 * (2 * (q0 * q1 + q2 * q3)) + zp4 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
[dX + m * (xp5 * (1 - 2 * (q2**2 + q3**2)) + yp5 * (2 * (q1 * q2 - q0 * q3)) + zp5 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp5 * (2 * (q1 * q2 + q0 * q3)) + yp5 * (1 - 2 * (q1**2 + q3**2)) + zp5 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp5 * (2 * (q1 * q3 - q0 * q2)) + yp5 * (2 * (q0 * q1 + q2 * q3)) + zp5 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
|
||||
|
||||
]
|
||||
)
|
||||
|
||||
A_ohne_zahlen = f.jacobian([dX, dY, dZ, m, q0, q1, q2, q3])
|
||||
A = A_ohne_zahlen.subs(zahlen)
|
||||
|
||||
#print(J)
|
||||
#print(J_zahlen.evalf(n=3))
|
||||
|
||||
# Parameterschätzung
|
||||
schwellenwert = 1e-3
|
||||
alle_kleiner = True
|
||||
|
||||
x0 = sp.Matrix([zahlen[dX], zahlen[dY], zahlen[dZ], zahlen[m], zahlen[q0], zahlen[q1], zahlen[q2], zahlen[q3]])
|
||||
P = sp.eye(15)
|
||||
N = A.T * P * A
|
||||
#l = sp.Matrix([p1[0], p1[1], p1[2], p2[0], p2[1], p2[2], p3[0], p3[1], p3[2], p4[0], p4[1], p4[2], p5[0], p5[1], p5[2]])
|
||||
|
||||
R_matrix = sp.Matrix([[1 - 2 * (q2_wert**2 + q3_wert**2), 2 * (q1_wert * q2_wert - q0_wert * q3_wert), 2 * (q0_wert * q2_wert + q1_wert * q3_wert)],
|
||||
[2 * (q1_wert * q2_wert + q0_wert * q3_wert), 1 - 2 * (q1_wert**2 + q3_wert**2), 2 * (q2_wert * q3_wert - q0_wert * q1_wert)],
|
||||
[2 * (q1_wert * q3_wert - q0_wert * q2_wert), 2 * (q0_wert * q1_wert + q2_wert * q3_wert), 1 - 2 * (q1_wert**2 + q2_wert**2)]])
|
||||
|
||||
liste_punkte_ausgangssystem = [p1, p2, p3, p4, p5]
|
||||
liste_l_berechnet_0 = [Translation + m0 * R_matrix * p for p in liste_punkte_ausgangssystem]
|
||||
l_berechnet_0 = sp.Matrix.vstack(*liste_l_berechnet_0)
|
||||
|
||||
l = sp.Matrix([P1[0] - p1[0], P1[1] - p1[1], P1[2] - p1[2], P2[0] - p2[0], P2[1] - p2[1], P2[2] - p2[2], P3[0] - p3[0], P3[1] - p3[1], P3[2] - p3[2], P4[0] - p4[0], P4[1] - p4[1], P4[2] - p4[2], P5[0] - p5[0], P5[1] - p5[1], P5[2] - p5[2]])
|
||||
|
||||
# ToDo: Prüfen, ob n mit l oder mit dl!
|
||||
n = A.T * P * l
|
||||
Qxx = N.evalf(n=30).inv()
|
||||
dx = Qxx * n
|
||||
x = x0 + dx
|
||||
|
||||
|
||||
|
||||
print(l.evalf(n=3))
|
||||
print(l_berechnet_0.evalf(n=3))
|
||||
#print(x0.evalf(n=3))
|
||||
#print(dx.evalf(n=3))
|
||||
#print(x.evalf(n=3))
|
||||
|
||||
for i in range (dx.rows):
|
||||
wert = float(dx[i])
|
||||
if abs(wert) >= schwellenwert:
|
||||
#print("Warnung")
|
||||
alle_kleiner = False
|
||||
|
||||
if alle_kleiner: print("Ausgleichung fertig!")
|
||||
246
Vorbereitungen_Fabian/Transformation_Helmert_V3.py
Normal file
246
Vorbereitungen_Fabian/Transformation_Helmert_V3.py
Normal file
@@ -0,0 +1,246 @@
|
||||
import sympy as sp
|
||||
from sympy.algebras.quaternion import Quaternion
|
||||
|
||||
#ToDo: Achtung: Die Ergebnisse sind leicht anders, als in den Beispielrechnung von Luhmann (Rundungsfehler bei Luhmann?)
|
||||
#ToDo: Automatische Ermittlung der Anzahl Nachkommastellen für Test auf Orthonormalität integrieren!
|
||||
#Beipsiel aus Luhmann S. 76
|
||||
# Ausgangssystem
|
||||
p1 = sp.Matrix([110, 100, 110])
|
||||
p2 = sp.Matrix([150, 280, 100])
|
||||
p3 = sp.Matrix([300, 300, 120])
|
||||
p4 = sp.Matrix([170, 100, 100])
|
||||
p5 = sp.Matrix([200, 200, 140])
|
||||
|
||||
# Zielsystem
|
||||
P1 = sp.Matrix([153.559, 170.747, 150.768])
|
||||
P2 = sp.Matrix([99.026, 350.313, 354.912])
|
||||
P3 = sp.Matrix([215.054, 544.420, 319.003])
|
||||
P4 = sp.Matrix([179.413, 251.030, 115.601])
|
||||
P5 = sp.Matrix([213.431, 340.349, 253.036])
|
||||
|
||||
#1) Näherungswertberechnung
|
||||
m0 = (P2 - P1).norm() / (p2 - p1).norm()
|
||||
|
||||
U = (P2 - P1) / (P2 - P1).norm()
|
||||
W = (U.cross(P3 - P1)) / (U.cross(P3 - P1)).norm()
|
||||
V = W.cross(U)
|
||||
|
||||
u = (p2 - p1) / (p2 - p1).norm()
|
||||
w = (u.cross(p3 - p1)) / (u.cross(p3 - p1)).norm()
|
||||
v = w.cross(u)
|
||||
|
||||
R = sp.Matrix.hstack(U, V, W) * sp.Matrix.hstack(u, v, w).T
|
||||
|
||||
XS = (P1 + P2 + P3) / 3
|
||||
xS = (p1 + p2 + p3) / 3
|
||||
|
||||
Translation = XS - m0 * R * xS
|
||||
|
||||
|
||||
#print(m0.evalf())
|
||||
#print(R.evalf())
|
||||
#print(Translation.evalf())
|
||||
|
||||
# 2) Test auf orthonormale Drehmatrix bei 3 Nachkommastellen!
|
||||
if R.T.applyfunc(lambda x: round(float(x), 3)) == R.inv().applyfunc(lambda x: round(float(x), 3)) and (R.T * R).applyfunc(lambda x: round(float(x), 3)) == sp.eye(3).applyfunc(lambda x: round(float(x), 3)) and ((round(R.det(), 3) == 1.000 or round(R.det(), 3) == -1.000)):
|
||||
print("R ist Orthonormal!")
|
||||
else:
|
||||
print("R ist nicht Orthonormal!")
|
||||
|
||||
# Testmatrix R aus Luhmann S. 66
|
||||
#R = sp.Matrix([
|
||||
# [0.996911, -0.013541, -0.077361],
|
||||
# [0.030706, 0.973820, 0.225238],
|
||||
# [0.072285, -0.226918, 0.971228]
|
||||
#])
|
||||
|
||||
# 3) Quaternionen berechnen
|
||||
# ToDo: Prüfen, ob Vorzeichen bei q0 richtig ist!
|
||||
#ToDo: q0 stimmt nicht mit Luhmann überein!
|
||||
|
||||
q = Quaternion.from_rotation_matrix(R)
|
||||
q0_wert = q.a
|
||||
q1_wert = q.b
|
||||
q2_wert = q.c
|
||||
q3_wert = q.d
|
||||
|
||||
|
||||
|
||||
# 4) Funktionales Modell
|
||||
liste_Punkte = ["P1", "P2", "P3", "P4", "P5"]
|
||||
liste_unbekannte = ["dX", "dY", "dZ", "dm", "dq0", "dq1", "dq2", "dq3"]
|
||||
liste_beobachtungen =[]
|
||||
for punkt in liste_Punkte:
|
||||
liste_beobachtungen.append(f"X_{punkt}")
|
||||
liste_beobachtungen.append(f"Y_{punkt}")
|
||||
liste_beobachtungen.append(f"Z_{punkt}")
|
||||
|
||||
|
||||
|
||||
|
||||
dX, dY, dZ, m, q0, q1, q2, q3, xp1, yp1, zp1, xp2, yp2, zp2, xp3, yp3, zp3, xp4, yp4, zp4, xp5, yp5, zp5 = sp.symbols('dX dY dZ m q0 q1 q2 q3 xp1 yp1 zp1 xp2 yp2 zp2 xp3 yp3 zp3 xp4 yp4 zp4 xp5 yp5 zp5')
|
||||
|
||||
#print(Translation[0])
|
||||
|
||||
|
||||
#print(zahlen[zp1])
|
||||
|
||||
f = sp.Matrix(
|
||||
[[dX + m * (xp1 * (1 - 2 * (q2**2 + q3**2)) + yp1 * (2 * (q1 * q2 - q0 * q3)) + zp1 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp1 * (2 * (q1 * q2 + q0 * q3)) + yp1 * (1 - 2 * (q1**2 + q3**2)) + zp1 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp1 * (2 * (q1 * q3 - q0 * q2)) + yp1 * (2 * (q0 * q1 + q2 * q3)) + zp1 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
[dX + m * (xp2 * (1 - 2 * (q2**2 + q3**2)) + yp2 * (2 * (q1 * q2 - q0 * q3)) + zp2 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp2 * (2 * (q1 * q2 + q0 * q3)) + yp2 * (1 - 2 * (q1**2 + q3**2)) + zp2 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp2 * (2 * (q1 * q3 - q0 * q2)) + yp2 * (2 * (q0 * q1 + q2 * q3)) + zp2 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
[dX + m * (xp3 * (1 - 2 * (q2**2 + q3**2)) + yp3 * (2 * (q1 * q2 - q0 * q3)) + zp3 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp3 * (2 * (q1 * q2 + q0 * q3)) + yp3 * (1 - 2 * (q1**2 + q3**2)) + zp3 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp3 * (2 * (q1 * q3 - q0 * q2)) + yp3 * (2 * (q0 * q1 + q2 * q3)) + zp3 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
[dX + m * (xp4 * (1 - 2 * (q2**2 + q3**2)) + yp4 * (2 * (q1 * q2 - q0 * q3)) + zp4 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp4 * (2 * (q1 * q2 + q0 * q3)) + yp4 * (1 - 2 * (q1**2 + q3**2)) + zp4 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp4 * (2 * (q1 * q3 - q0 * q2)) + yp4 * (2 * (q0 * q1 + q2 * q3)) + zp4 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
[dX + m * (xp5 * (1 - 2 * (q2**2 + q3**2)) + yp5 * (2 * (q1 * q2 - q0 * q3)) + zp5 * (2 * (q0 * q2 + q1 * q3)))],
|
||||
[dY + m * (xp5 * (2 * (q1 * q2 + q0 * q3)) + yp5 * (1 - 2 * (q1**2 + q3**2)) + zp5 * (2 * (q2 * q3 - q0 * q1)))],
|
||||
[dZ + m * (xp5 * (2 * (q1 * q3 - q0 * q2)) + yp5 * (2 * (q0 * q1 + q2 * q3)) + zp5 * (1 - 2 * (q1**2 + q2**2)))],
|
||||
|
||||
|
||||
]
|
||||
)
|
||||
|
||||
A_ohne_zahlen = f.jacobian([dX, dY, dZ, m, q0, q1, q2, q3])
|
||||
|
||||
|
||||
#print(J)
|
||||
#print(J_zahlen.evalf(n=3))
|
||||
|
||||
# Parameterschätzung
|
||||
schwellenwert = 1e-3
|
||||
anzahl_iterationen = 0
|
||||
alle_kleiner_vorherige_iteration = False
|
||||
|
||||
|
||||
P = sp.eye(15)
|
||||
|
||||
#l = sp.Matrix([p1[0], p1[1], p1[2], p2[0], p2[1], p2[2], p3[0], p3[1], p3[2], p4[0], p4[1], p4[2], p5[0], p5[1], p5[2]])
|
||||
|
||||
|
||||
liste_punkte_ausgangssystem = [p1, p2, p3, p4, p5]
|
||||
|
||||
|
||||
|
||||
#l = sp.Matrix([P1[0] - p1[0], P1[1] - p1[1], P1[2] - p1[2], P2[0] - p2[0], P2[1] - p2[1], P2[2] - p2[2], P3[0] - p3[0], P3[1] - p3[1], P3[2] - p3[2], P4[0] - p4[0], P4[1] - p4[1], P4[2] - p4[2], P5[0] - p5[0], P5[1] - p5[1], P5[2] - p5[2]])
|
||||
l = sp.Matrix([P1[0], P1[1], P1[2], P2[0], P2[1], P2[2], P3[0], P3[1], P3[2], P4[0], P4[1], P4[2], P5[0], P5[1], P5[2]])
|
||||
# ToDo: Prüfen, ob n mit l oder mit dl!
|
||||
|
||||
while True:
|
||||
if anzahl_iterationen == 0:
|
||||
zahlen_0 = {dX: Translation[0], dY: Translation[1], dZ: Translation[2], m: m0, q0: q0_wert, q1: q1_wert,
|
||||
q2: q2_wert,
|
||||
q3: q3_wert, xp1: p1[0], yp1: p1[1], zp1: p1[2], xp2: p2[0], yp2: p2[1], zp2: p2[2], xp3: p3[0],
|
||||
yp3: p3[1], zp3: p3[2], xp4: p4[0], yp4: p4[1], zp4: p4[2], xp5: p5[0], yp5: p5[1], zp5: p5[2]}
|
||||
x0 = sp.Matrix([zahlen_0[dX], zahlen_0[dY], zahlen_0[dZ], zahlen_0[m], zahlen_0[q0], zahlen_0[q1], zahlen_0[q2], zahlen_0[q3]])
|
||||
R_matrix_0 = sp.Matrix([[1 - 2 * (q2_wert ** 2 + q3_wert ** 2), 2 * (q1_wert * q2_wert - q0_wert * q3_wert),
|
||||
2 * (q0_wert * q2_wert + q1_wert * q3_wert)],
|
||||
[2 * (q1_wert * q2_wert + q0_wert * q3_wert), 1 - 2 * (q1_wert ** 2 + q3_wert ** 2),
|
||||
2 * (q2_wert * q3_wert - q0_wert * q1_wert)],
|
||||
[2 * (q1_wert * q3_wert - q0_wert * q2_wert), 2 * (q0_wert * q1_wert + q2_wert * q3_wert),
|
||||
1 - 2 * (q1_wert ** 2 + q2_wert ** 2)]])
|
||||
liste_l_berechnet_0 = [Translation + m0 * R_matrix_0 * p for p in liste_punkte_ausgangssystem]
|
||||
l_berechnet_0 = sp.Matrix.vstack(*liste_l_berechnet_0)
|
||||
dl_0 = l - l_berechnet_0
|
||||
|
||||
A_0 = A_ohne_zahlen.subs(zahlen_0)
|
||||
N = A_0.T * P * A_0
|
||||
n_0 = A_0.T * P * dl_0
|
||||
Qxx_0 = N.evalf(n=30).inv()
|
||||
dx = Qxx_0 * n_0
|
||||
x = x0 + dx
|
||||
q_norm = sp.sqrt(x[4] ** 2 + x[5] ** 2 + x[6] ** 2 + x[7] ** 2)
|
||||
x[4] /= q_norm
|
||||
x[5] /= q_norm
|
||||
x[6] /= q_norm
|
||||
x[7] /= q_norm
|
||||
anzahl_iterationen += 1
|
||||
print(f"Iteration Nr.{anzahl_iterationen} abgeschlossen")
|
||||
print(dx.evalf(n=3))
|
||||
|
||||
else:
|
||||
#print("Im else-Block")
|
||||
zahlen_i = {dX: float(x[0]), dY: float(x[1]), dZ: float(x[2]), m: float(x[3]), q0: float(x[4]), q1: float(x[5]),
|
||||
q2: float(x[6]),
|
||||
q3: float(x[7]), xp1: p1[0], yp1: p1[1], zp1: p1[2], xp2: p2[0], yp2: p2[1], zp2: p2[2], xp3: p3[0],
|
||||
yp3: p3[1], zp3: p3[2], xp4: p4[0], yp4: p4[1], zp4: p4[2], xp5: p5[0], yp5: p5[1], zp5: p5[2]}
|
||||
#print("zahlen_i")
|
||||
R_matrix_i = sp.Matrix([[1 - 2 * (zahlen_i[q2] ** 2 + zahlen_i[q3] ** 2), 2 * (zahlen_i[q1] * zahlen_i[q2] - zahlen_i[q0] * zahlen_i[q3]),
|
||||
2 * (zahlen_i[q0] * zahlen_i[q2] + zahlen_i[q1] * zahlen_i[q3])],
|
||||
[2 * (zahlen_i[q1] * zahlen_i[q2] + zahlen_i[q0] * zahlen_i[q3]), 1 - 2 * (zahlen_i[q1] ** 2 + zahlen_i[q3] ** 2),
|
||||
2 * (zahlen_i[q2] * zahlen_i[q3] - zahlen_i[q0] * zahlen_i[q1])],
|
||||
[2 * (zahlen_i[q1] * zahlen_i[q3] - zahlen_i[q0] * zahlen_i[q2]),
|
||||
2 * (zahlen_i[q0] * zahlen_i[q1] + zahlen_i[q2] * zahlen_i[q3]),
|
||||
1 - 2 * (zahlen_i[q1] ** 2 + zahlen_i[q2] ** 2)]])
|
||||
#print("R_matrix_i")
|
||||
liste_l_berechnet_i = [sp.Matrix([zahlen_i[dX], zahlen_i[dY], zahlen_i[dZ]]) + zahlen_i[m] * R_matrix_i * p for p in liste_punkte_ausgangssystem]
|
||||
#print("liste_l_berechnet_i")
|
||||
l_berechnet_i = sp.Matrix.vstack(*liste_l_berechnet_i)
|
||||
#print("l_berechnet_i")
|
||||
dl_i = l - l_berechnet_i
|
||||
#print("dl_i")
|
||||
A_i = A_ohne_zahlen.subs(zahlen_i).evalf(n=3)
|
||||
#print("A_i")
|
||||
N_i = A_i.T * P * A_i
|
||||
#print("N_i")
|
||||
n_i = A_i.T * P * dl_i
|
||||
# print("n_i")
|
||||
Qxx_i = N_i.evalf(n=30).inv()
|
||||
#print("Qxx_i")
|
||||
n_i = A_i.T * P * dl_i
|
||||
#print("n_i")
|
||||
dx = Qxx_i * n_i
|
||||
#print("dx")
|
||||
x += dx
|
||||
q_norm = sp.sqrt(x[4] ** 2 + x[5] ** 2 + x[6] ** 2 + x[7] ** 2)
|
||||
x[4] /= q_norm
|
||||
x[5] /= q_norm
|
||||
x[6] /= q_norm
|
||||
x[7] /= q_norm
|
||||
# print("x")
|
||||
anzahl_iterationen += 1
|
||||
print(f"Iteration Nr.{anzahl_iterationen} abgeschlossen")
|
||||
print(dx.evalf(n=3))
|
||||
|
||||
alle_kleiner = True
|
||||
for i in range(dx.rows):
|
||||
wert = float(dx[i])
|
||||
if abs(wert) > schwellenwert:
|
||||
alle_kleiner = False
|
||||
|
||||
|
||||
if alle_kleiner and alle_kleiner_vorherige_iteration:
|
||||
break
|
||||
|
||||
alle_kleiner_vorherige_iteration = alle_kleiner
|
||||
|
||||
print(l.evalf(n=3))
|
||||
print(l_berechnet_0.evalf(n=3))
|
||||
print(f"x = {x.evalf(n=3)}")
|
||||
|
||||
#Neuberechnung Zielsystem
|
||||
zahlen_i = {dX: float(x[0]), dY: float(x[1]), dZ: float(x[2]), m: float(x[3]), q0: float(x[4]), q1: float(x[5]),
|
||||
q2: float(x[6]),
|
||||
q3: float(x[7]), xp1: p1[0], yp1: p1[1], zp1: p1[2], xp2: p2[0], yp2: p2[1], zp2: p2[2], xp3: p3[0],
|
||||
yp3: p3[1], zp3: p3[2], xp4: p4[0], yp4: p4[1], zp4: p4[2], xp5: p5[0], yp5: p5[1], zp5: p5[2]}
|
||||
# print("zahlen_i")
|
||||
R_matrix_i = sp.Matrix(
|
||||
[[1 - 2 * (zahlen_i[q2] ** 2 + zahlen_i[q3] ** 2), 2 * (zahlen_i[q1] * zahlen_i[q2] - zahlen_i[q0] * zahlen_i[q3]),
|
||||
2 * (zahlen_i[q0] * zahlen_i[q2] + zahlen_i[q1] * zahlen_i[q3])],
|
||||
[2 * (zahlen_i[q1] * zahlen_i[q2] + zahlen_i[q0] * zahlen_i[q3]), 1 - 2 * (zahlen_i[q1] ** 2 + zahlen_i[q3] ** 2),
|
||||
2 * (zahlen_i[q2] * zahlen_i[q3] - zahlen_i[q0] * zahlen_i[q1])],
|
||||
[2 * (zahlen_i[q1] * zahlen_i[q3] - zahlen_i[q0] * zahlen_i[q2]),
|
||||
2 * (zahlen_i[q0] * zahlen_i[q1] + zahlen_i[q2] * zahlen_i[q3]),
|
||||
1 - 2 * (zahlen_i[q1] ** 2 + zahlen_i[q2] ** 2)]])
|
||||
# print("R_matrix_i")
|
||||
liste_l_berechnet_i = [sp.Matrix([zahlen_i[dX], zahlen_i[dY], zahlen_i[dZ]]) + zahlen_i[m] * R_matrix_i * p for p in
|
||||
liste_punkte_ausgangssystem]
|
||||
# print("liste_l_berechnet_i")
|
||||
l_berechnet_i = sp.Matrix.vstack(*liste_l_berechnet_i)
|
||||
print("")
|
||||
print(f"l_berechnet_final: {l_berechnet_i.evalf(n=3)}")
|
||||
Reference in New Issue
Block a user