From: Erik Andresen <erik@vontaene.de>
Date: Fri, 27 Jul 2018 07:01:37 +0000 (+0200)
Subject: Added simple kalman
X-Git-Url: https://defiant.homedns.org/gitweb/?a=commitdiff_plain;h=a122ce94f6adfb634d2c3ec74f2b12c782115cc7;p=pyshared.git

Added simple kalman
---

diff --git a/simple_kalman.py b/simple_kalman.py
new file mode 100755
index 0000000..d3d36ef
--- /dev/null
+++ b/simple_kalman.py
@@ -0,0 +1,100 @@
+#!/usr/bin/env python
+# -*- coding: iso-8859-15 -*-
+
+# Implementation of simple Kalman in python based on
+# Xilinx DSP Magazine #1 (2005) - http://www.xilinx.com/publications/
+
+import numpy as np
+
+class simple_kalman:
+	def __init__(self, x_est, P_est, Q, R):
+		# Convert arguments to vectors/matrices if not given
+		if type(x_est) != type(np.array([0])): x_est = np.array(x_est) 
+		if type(P_est) != type(np.matrix([0])): P_est = np.matrix(P_est)
+		if type(Q) != type(np.matrix([0])): Q = np.matrix(Q)
+		if type(R) != type(np.matrix([0])): R = np.matrix(R)
+
+		self.x_est = x_est # Systemzustand
+		self.P_est = P_est # Fehlerkovarianz
+		self.Q = Q # Systemrauschen
+		self.R = R # Varianz des Messfehlers
+		self.I = np.eye(P_est.shape[0])
+
+	def run(self, y):
+		# Korrektur mit der Messung
+		# (1) Berechnung der Kalman Verstärkung
+		K = self.P_est * (self.R + self.P_est).I
+		# (2) Korrektur der Schätzung mit der Messung y
+		x = self.x_est + K*(y - self.x_est)
+		# (3) Korrektur der Fehlerkovarianzmatrix
+		P = (self.I-K)*self.P_est
+		#
+		# Prädiktion
+		# (1) Prädiziere den Systemzustand
+		self.x_est = x
+		# (2) Präzidiere die Fehlerkovarianzmatrix
+		self.P_est = P + self.Q
+
+		if x.shape == (1,1): return x.item(0,0)
+		return x
+
+
+if __name__ == '__main__':
+	import random
+	from matplotlib.pyplot import *
+
+	#
+	# 1d Test
+	#
+	orig = [0.5 for i in range(100)]
+	y = [orig[i] + (random.random()-0.5)/5 for i in range(0, len(orig))]
+	x = []
+
+	x_est = 0.0 # Systemzustand
+	P_est = 0.1 # Fehlerkovarianz
+	Q = 10e-6 # Systemrauschen
+	R = 0.0035 # Varianz des Messfehlers
+	p = simple_kalman(x_est, P_est, Q, R)
+
+	for i in range(0, 100):
+		# Messwert
+		x.append(p.run(y[i]))
+
+	plot(orig, label="Orig")
+	plot(range(100), y, 'o', label="Noise")
+	plot(x, label="Filtered")
+	ylim(0, 2)
+	legend()
+	show()
+
+	#
+	# 2d Test
+	#
+	orig1 = [0.5 for i in range(100)]
+	orig2 = [1.5 for i in range(100)]
+	y1 = [orig1[i] + (random.random()-0.5)/5 for i in range(0, len(orig1))]
+	y2 = [orig2[i] + (random.random()-0.5)/5 for i in range(0, len(orig2))]
+	x1 = []
+	x2 = []
+
+	x_est = np.array([[0.0, 0.0]]).T # Systemzustand
+	P_est = 0.1 * np.eye(2) # Fehlerkovarianz
+	Q = 10e-6 * np.eye(2) # Systemrauschen
+	R = np.matrix([[0.0035, 0.0], [0.0, 0.0035]]) # Varianz des Messfehlers
+	p = simple_kalman(x_est, P_est, Q, R)
+
+	for i in range(0, 100):
+		# Messwert
+		a = p.run(np.array([[y1[i], y2[i]]]).T)
+		x1.append(a.item((0, 0)))
+		x2.append(a.item((1, 0)))
+
+	plot(orig1, label="Orig")
+	plot(orig2, label="Orig")
+	plot(range(100), y1, 'o', label="Noise")
+	plot(range(100), y2, 'o', label="Noise")
+	plot(x1, label="Filtered")
+	plot(x2, label="Filtered")
+	ylim(0, 2)
+	legend()
+	show()