asr_vosk: Allow to handle keyword and command in one sentence
[ros_wild_thumper.git] / scripts / dwm1000.py
index bf58657..b16ff9e 100755 (executable)
@@ -1,7 +1,7 @@
 #!/usr/bin/env python
 # -*- coding: iso-8859-15 -*-
 
-VISULAIZE = False
+VISUALIZE = False
 
 import threading
 import struct
@@ -15,32 +15,10 @@ from time import sleep
 from std_msgs.msg import Float32
 from nav_msgs.msg import Odometry
 from wild_thumper.srv import DWM1000Center, DWM1000CenterResponse
-if VISULAIZE:
+from pyshared.simple_kalman import simple_kalman
+if VISUALIZE:
        import matplotlib.pyplot as plt
 
-class simple_kalman:
-       def __init__(self, x_est, P_est, Q, R):
-               self.x_est = x_est # Systemzustand
-               self.P_est = P_est # Fehlerkovarianz
-               self.Q = Q # Systemrauschen
-               self.R = R # Varianz des Messfehlers
-
-       def run(self, y):
-               # Korrektur mit der Messung
-               # (1) Berechnung der Kalman Verstärkung
-               K = self.P_est/(self.R + self.P_est)
-               # (2) Korrektur der Schätzung mit der Messung y
-               x = self.x_est + K*(y - self.x_est)
-               # (3) Korrektur der Fehlerkovarianzmatrix
-               P = (1-K)*self.P_est
-               #
-               # Prädiktion
-               # (1) Prädiziere den Systemzustand
-               self.x_est = x
-               # (2) Präzidiere die Fehlerkovarianzmatrix
-               self.P_est = P + self.Q
-
-               return x
 
 class DW1000(threading.Thread):
        def __init__(self, name, addr, offset):
@@ -73,10 +51,10 @@ class DW1000(threading.Thread):
 
        def run(self):
                last_val = 10
-               while True:
+               while not rospy.is_shutdown():
                        val = self.get_value()
-                       if abs(val - last_val)  > 10:
-                               print "Ignoring values too far apart %s: %.2f - %.2f" % (self.name, val, last_val)
+                       if abs(val - last_val)  > 50:
+                               rospy.logwarn("Ignoring values too far apart %s: %.2f - %.2f", self.name, val, last_val)
                        elif not isnan(val):
                                self.dist = val + self.offset
                                self.last_update = datetime.now()
@@ -86,16 +64,10 @@ class DW1000(threading.Thread):
 
 class Position:
        def __init__(self):
-               # Varianz des Messfehlers
-               Rx = 0.2
-               Ry = 0.05
-               # Fehlerkovarianz
-               P_est_x = 0.02
-               P_est_y = 0.01
-               # Systemrauschen
-               Q = 0.002
-               self.filter_x = simple_kalman(1.0, P_est_x, Q, Rx)
-               self.filter_y = simple_kalman(0.0, P_est_y, Q, Ry)
+               R = np.matrix([[0.5, 0.0], [0.0, 0.2]])         # Varianz des Messfehlers
+               P_est = 1e50 * np.eye(2)                        # Fehlerkovarianz
+               Q = np.matrix([[0.001, 0.0], [0.0, 0.0001]])    # Systemrauschen
+               self.filter_xy = simple_kalman(np.array([[1.0, 0.0]]).T, P_est, Q, R)
                self.speed_x = 0
                self.speed_y = 0
                self.speed_z = 0
@@ -107,16 +79,12 @@ class Position:
                self.speed_y = msg.twist.twist.linear.y
                self.speed_z = msg.twist.twist.angular.z
 
-       """
-       TODO:
-       - variance of kalman should be dependant on distance
-       """
        def filter(self, x, y):
                # Correct estimation with speed
                current_time = rospy.Time.now()
                dt = (current_time - self.last_time).to_sec()
                # Subtract vehicle speed
-               pos = np.array([self.filter_x.x_est, self.filter_y.x_est])
+               pos = self.filter_xy.x_est.T
                # translation
                pos -= np.array([self.speed_x*dt, self.speed_y*dt])
                # rotation
@@ -124,16 +92,25 @@ class Position:
                                [np.sin(self.speed_z*dt),  np.cos(self.speed_z*dt)]])
                pos = np.dot(pos, rot)
                # copy back
-               self.filter_x.x_est = pos[0]
-               self.filter_y.x_est = pos[1]
+               self.filter_xy.x_est = pos.T
 
                # run kalman if new measurements are valid
+               pos = None
                if x != None and y != None:
-                       x = self.filter_x.run(x)
-                       y = self.filter_y.run(y)
+                       pos = self.filter_xy.run(np.array([[x, y]]).T)
+
+                       # Update variance
+                       dist = np.linalg.norm([x, y])
+                       cov_00 = np.polyval([0.018, 0.0, 0.0], dist)
+                       cov_11 = np.polyval([0.006, 0.0, 0.0], dist)
+                       self.filter_xy.set_measure_cov(np.matrix([[cov_00, 0.0], [0.0, cov_11]]))
                else:
-                       x = self.filter_x.x_est
-                       y = self.filter_y.x_est
+                       pos = self.filter_xy.x_est
+                       self.filter_xy.P_est = np.matrix(1e50 * np.eye(2)) # reset estimate when lost
+
+
+               x = pos.item((0, 0))
+               y = pos.item((1, 0))
 
                self.last_time = current_time
                return x,y
@@ -143,14 +120,14 @@ def handle_center_call(req):
        diff = dwleft.distance_valid() - dwright.distance_valid()
        dwleft.offset -= diff/2
        dwright.offset += diff/2
-       print "Centering to %.2f %.2f" % (dwleft.offset, dwright.offset)
+       rospy.loginfo("Centering to %.2f %.2f", dwleft.offset, dwright.offset)
        return DWM1000CenterResponse()
 
 if __name__ == "__main__":
-       rospy.init_node('DWM1000')
-       dwleft  = DW1000("uwb_dist_left",  0xc2, +0.02)
-       dwright = DW1000("uwb_dist_right", 0xc0, -0.02)
-       dist_l_r = 0.285
+       rospy.init_node('DWM1000', log_level=rospy.DEBUG)
+       dwleft  = DW1000("uwb_dist_left",  0xc2, -0.04)
+       dwright = DW1000("uwb_dist_right", 0xc0, +0.04)
+       dist_l_r = 0.285 # Distance between both DWM1000
        rate = rospy.Rate(10)
        pos = Position()
        tf_broadcaster = tf.broadcaster.TransformBroadcaster()
@@ -159,16 +136,18 @@ if __name__ == "__main__":
        while not rospy.is_shutdown() and dwleft.is_alive() and dwright.is_alive():
                dist_left = dwleft.distance_valid()
                dist_right = dwright.distance_valid()
+               x = None
+               y = None
                if dist_left == None or dist_right == None:
-                       print "no valid sensor update"
+                       rospy.logerr_throttle(10, "no valid sensor update %r %r" % (dist_left, dist_right))
                        # run kalman prediction only
-                       pos.filter(None, None)
+                       x,y = pos.filter(None, None)
                else:
                        dir = "left" if (dist_left < dist_right) else "right"
 
                        diff = abs(dist_left - dist_right)
                        if diff >= dist_l_r:
-                               # difference to high, correct to maximum
+                               # difference too high, correct to maximum
                                off = diff - dist_l_r + 0.01
                                if dist_left > dist_right:
                                        dist_left -= off/2
@@ -176,21 +155,20 @@ if __name__ == "__main__":
                                else:
                                        dist_left += off/2
                                        dist_right -= off/2
-                       print "%.2f %.2f %.2f %.2f %s" % (dwleft.distance(), dwright.distance(), dist_left, dist_right, dir)
+                       rospy.logdebug("%.2f %.2f %.2f %.2f %s", dwleft.distance(), dwright.distance(), dist_left, dist_right, dir)
 
                        a_r = (-dist_right**2 + dist_left**2 - dist_l_r**2) / (-2*dist_l_r)
                        x = dist_l_r/2 - a_r
                        t = dist_right**2 - a_r**2
                        if t >= 0:
                                y = sqrt(t)
-                               print x,y
+                               rospy.logdebug("x=%.2f, y=%.2f", x, y)
                                # Rotate 90 deg
                                x, y = (y, -x)
 
                                x, y = pos.filter(x, y)
-                               tf_broadcaster.sendTransform((x, y, 0.0), (0, 0, 0, 1), rospy.Time.now(), "uwb_beacon", "base_footprint")
 
-                               if VISULAIZE:
+                               if VISUALIZE:
                                        circle_left = plt.Circle((-dist_l_r/2, 0), dwleft.distance, color='red', fill=False)
                                        circle_right = plt.Circle((dist_l_r/2, 0), dwright.distance, color='green', fill=False)
                                        plt.gca().add_patch(circle_left)
@@ -200,6 +178,7 @@ if __name__ == "__main__":
                                        plt.show()
                        else:
                                # No current position, still need up update kalman prediction
-                               pos.filter(None, None)
+                               x, y = pos.filter(None, None)
+               tf_broadcaster.sendTransform((x, y, 0.0), (0, 0, 0, 1), rospy.Time.now(), "uwb_beacon", "base_footprint")
 
                rate.sleep()