I am trying to find user distances from beacon positions using trilateration. I am having a mean squared error function that I am trying to minimize using non-linear least squares but I am not getting correct results. Any help is appreciated. The code is below.
def mse(self, user_pos, positions, distances):
mse = 0.0
for pos, dist in zip(positions, distances):
distance = great_circle((user_pos[0], user_pos[1]), (pos[0], pos[1])).meters
mse += math.pow(distance - dist, 2.0)
return mse/len(positions)
def least_squares_func(self, positions, distances):
# Returns users coordinates
return least_squares(self.mse, [0,0], args=(positions, distances)).x
Starting position in least_squares is [0,0] but after changing it I am not getting much different results.
Input example: positions = [(5.0, -6.0), (13.0, -15.0), (21.0, -3.0)] distances = [8.06, 13.97, 23.32]
