Skip to content

GPS Noise Computation #1041

@rafaeljer

Description

@rafaeljer

I have some doubts regarding how GPS noise is being computed.

if (gps_noise_) {
noise_gps_pos_.X() = gps_xy_noise_density_ * sqrt(dt) * randn_(rand_);
noise_gps_pos_.Y() = gps_xy_noise_density_ * sqrt(dt) * randn_(rand_);
noise_gps_pos_.Z() = gps_z_noise_density_ * sqrt(dt) * randn_(rand_);
noise_gps_vel_.X() = gps_vxy_noise_density_ * sqrt(dt) * randn_(rand_);
noise_gps_vel_.Y() = gps_vxy_noise_density_ * sqrt(dt) * randn_(rand_);
noise_gps_vel_.Z() = gps_vz_noise_density_ * sqrt(dt) * randn_(rand_);
random_walk_gps_.X() = gps_xy_random_walk_ * sqrt(dt) * randn_(rand_);
random_walk_gps_.Y() = gps_xy_random_walk_ * sqrt(dt) * randn_(rand_);
random_walk_gps_.Z() = gps_z_random_walk_ * sqrt(dt) * randn_(rand_);
}

Noise density for XY has, as unit, (m) / sqrt(hz) and the noise has as unit (m).

static constexpr double kDefaultGpsXYNoiseDensity = 2.0e-4; // (m) / sqrt(hz)

So, by the current formula, noise_gps_pos_.X() = (m) / sqrt(hz) * (1/sqrt(Hz)) = (m) / (Hz).

So, it would make me sense that the formula divides noise density by sqrt(df) [sqrt(s)], so that the result is in (m) instead of multiplying:
noise_gps_pos_.X() = ((m) / sqrt(hz)) / (1/sqrt(Hz)) = (m)

The formula I referred is also used in other places of the code, like here

double sigma_g_d = 1 / sqrt(dt) * imu_parameters_.gyroscope_noise_density;

If you could clarify, I would appreciate.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions