def update(self, value):
if self.maximum is None:
self.maximum = value
else:
self.maximum = numpy.maximum(value, self.maximum)
if self.minimum is None:
self.minimum = value
else:
self.minimum = numpy.minimum(value, self.minimum)
self.mean_accum.update(value)
self.mean = self.mean_accum.get_value()
value_norm = value - self.mean
P = outer(value_norm, value_norm)
self.covariance_accum.update(P)
self.covariance = self.covariance_accum.get_value()
try:
self.information = pinv(self.covariance, rcond=1e-2)
except LinAlgError:
filename = "pinv-failure"
with open(filename + ".pickle", "w") as f:
self.last_value = value
pickle.dump(self, f)
raise JobFailed("Did not converge; saved on %s" % filename)
def process_data(self, data):
y = data.sensels
y_dot = data.sensels_dot
u = data.commands
# n = y.shape[0]
# k = u.shape[0]
self.y_dot_stats.update(y_dot)
self.y_stats.update(y)
self.u_stats.update(u)
y_norm = dot(self.y_stats.information, y - self.y_stats.mean)
u_norm = dot(self.u_stats.information, u - self.u_stats.mean)
# Assuming model of the form:
#
# y_dot = mu + M y + N u + T(y,u)
#
# assuming that mu = 0 and M = 0
#
# Then:
# N = E{ y_dot u~ }
Ns = outer(y_dot, u_norm)
self.N.update(Ns)
N = self.N.get_value()
Nu = dot(N, u)
Ts = outer(u_norm, outer(y_norm, y_dot - Nu))
self.T.update(Ts)
# this assumes the normal bilinear model
bT = outer(u_norm, outer(y - self.y_stats.mean, y_dot))
self.bT.update(bT)
# this assumes the normal bilinear model
bTn = outer(u_norm, outer(y_norm, y_dot))
self.bTn.update(bTn)
