px4-firmware/airdata/python/wind_est_derivation.py

189 lines
6.0 KiB
Python

# -*- coding: utf-8 -*-
"""
Created on Tue Nov 1 19:14:39 2016
@author: roman
"""
from sympy import *
# q: quaternion describing rotation from frame 1 to frame 2
# returns a rotation matrix derived form q which describes the same
# rotation
def quat2Rot(q):
q0 = q[0]
q1 = q[1]
q2 = q[2]
q3 = q[3]
Rot = Matrix([[q0**2 + q1**2 - q2**2 - q3**2, 2*(q1*q2 - q0*q3), 2*(q1*q3 + q0*q2)],
[2*(q1*q2 + q0*q3), q0**2 - q1**2 + q2**2 - q3**2, 2*(q2*q3 - q0*q1)],
[2*(q1*q3-q0*q2), 2*(q2*q3 + q0*q1), q0**2 - q1**2 - q2**2 + q3**2]])
return Rot
# take an expression calculated by the cse() method and write the expression
# into a text file in C format
def write_simplified(P_touple, filename, out_name):
subs = P_touple[0]
P = Matrix(P_touple[1])
fd = open(filename, 'a')
is_vector = P.shape[0] == 1 or P.shape[1] == 1
# write sub expressions
for index, item in enumerate(subs):
fd.write('float ' + str(item[0]) + ' = ' + str(item[1]) + ';\n')
# write actual matrix values
fd.write('\n')
if not is_vector:
iterator = range(0,sqrt(len(P)), 1)
for row in iterator:
for column in iterator:
fd.write(out_name + '(' + str(row) + ',' + str(column) + ') = ' + str(P[row, column]) + ';\n')
else:
iterator = range(0, len(P), 1)
for item in iterator:
fd.write(out_name + '(' + str(item) + ') = ' + str(P[item]) + ';\n')
fd.write('\n\n')
fd.close()
########## Symbolic variable definition #######################################
# model state
w_n = Symbol("w_n", real=True) # wind in north direction
w_e = Symbol("w_e", real=True) # wind in east direction
k_tas = Symbol("k_tas", real=True) # true airspeed scale factor
state = Matrix([w_n, w_e, k_tas])
# process noise
q_w = Symbol("q_w", real=True) # process noise for wind states
q_k_tas = Symbol("q_k_tas", real=True) # process noise for airspeed scale state
# airspeed measurement noise
r_tas = Symbol("r_tas", real=True)
# sideslip measurement noise
r_beta = Symbol("r_beta", real=True)
# true airspeed measurement
tas_meas = Symbol("tas_meas", real=True)
# ground velocity variance
v_n_var = Symbol("v_n_var", real=True)
v_e_var = Symbol("v_e_var", real=True)
#################### time varying parameters ##################################
# vehicle velocity
v_n = Symbol("v_n", real=True) # north velocity in earth fixed frame
v_e = Symbol("v_e", real=True) # east velocity in earth fixed frame
v_d = Symbol("v_d", real=True) # down velocity in earth fixed frame
# unit quaternion describing vehicle attitude, qw is real part
qw = Symbol("q_att[0]", real=True)
qx = Symbol("q_att[1]", real=True)
qy = Symbol("q_att[2]", real=True)
qz = Symbol("q_att[3]", real=True)
q_att = Matrix([qw, qx, qy, qz])
# sampling time in seconds
dt = Symbol("dt", real=True)
######################## State and covariance prediction ######################
# state transition matrix is zero because we are using a stationary
# process model. We only need to provide formula for covariance prediction
# create process noise matrix for covariance prediction
state_new = state + Matrix([q_w, q_w, q_k_tas]) * dt
Q = diag(q_w, q_k_tas)
L = state_new.jacobian([q_w, q_k_tas])
Q = L * Q * Transpose(L)
# define symbolic covariance matrix
p00 = Symbol('_P(0,0)', real=True)
p01 = Symbol('_P(0,1)', real=True)
p02 = Symbol('_P(0,2)', real=True)
p12 = Symbol('_P(1,2)', real=True)
p11 = Symbol('_P(1,1)', real=True)
p22 = Symbol('_P(2,2)', real=True)
P = Matrix([[p00, p01, p02], [p01, p11, p12], [p02, p12, p22]])
# covariance prediction equation
P_next = P + Q
# simplify the result and write it to a text file in C format
PP_simple = cse(P_next, symbols('SPP0:30'))
P_pred = Matrix(PP_simple[1])
write_simplified(PP_simple, "cov_pred.txt", 'P_next')
############################ Measurement update ###############################
# airspeed fusion
tas_pred = Matrix([((v_n - w_n)**2 + (v_e - w_e)**2 + v_d**2)**0.5]) * k_tas
# compute true airspeed observation matrix
H_tas = tas_pred.jacobian(state)
# simplify the result and write it to a text file in C format
H_tas_simple = cse(H_tas, symbols('HH0:30'))
write_simplified(H_tas_simple, "airspeed_fusion.txt", 'H_tas')
K = P * Transpose(H_tas)
denom = H_tas * P * Transpose(H_tas) + Matrix([r_tas])
denom = 1/denom.values()[0]
K = K * denom
K_simple = cse(K, symbols('KTAS0:30'))
write_simplified(K_simple, "airspeed_fusion.txt", "K")
P_m = P - K*H_tas*P
P_m_simple = cse(P_m, symbols('PM0:50'))
write_simplified(P_m_simple, "airspeed_fusion.txt", "P_next")
# sideslip fusion
# compute relative wind vector in vehicle body frame
relative_wind_earth = Matrix([v_n - w_n, v_e - w_e, v_d])
R_body_to_earth = quat2Rot(q_att)
relative_wind_body = Transpose(R_body_to_earth) * relative_wind_earth
# small angle approximation of side slip model
beta_pred = relative_wind_body[1] / relative_wind_body[0]
# compute side slip observation matrix
H_beta = Matrix([beta_pred]).jacobian(state)
# simplify the result and write it to a text file in C format
H_beta_simple = cse(H_beta, symbols('HB0:30'))
write_simplified(H_beta_simple, "beta_fusion.txt", 'H_beta')
K = P * Transpose(H_beta)
denom = H_beta * P * Transpose(H_beta) + Matrix([r_beta])
denom = 1/denom.values()[0]
K = K*denom
K_simple = cse(K, symbols('KB0:30'))
write_simplified(K_simple, "beta_fusion.txt", 'K')
P_m = P - K*H_beta*P
P_m_simple = cse(P_m, symbols('PM0:50'))
write_simplified(P_m_simple, "beta_fusion.txt", "P_next")
# wind covariance initialisation via velocity
# estimate heading from ground velocity
heading_est = atan2(v_n, v_e)
# calculate wind speed estimate from vehicle ground velocity, heading and
# airspeed measurement
w_n_est = v_n - tas_meas * cos(heading_est)
w_e_est = v_e - tas_meas * sin(heading_est)
wind_est = Matrix([w_n_est, w_e_est])
# calculate estimate of state covariance matrix
P_wind = diag(v_n_var, v_e_var, r_tas)
wind_jac = wind_est.jacobian([v_n, v_e, tas_meas])
wind_jac_simple = cse(wind_jac, symbols('L0:30'))
write_simplified(wind_jac_simple, "cov_init.txt", "L")