Skip to content

Latest commit

 

History

History
48 lines (33 loc) · 2.16 KB

switching_state_estimators.md

File metadata and controls

48 lines (33 loc) · 2.16 KB

Switching State Estimators

This page shows you which state estimators are available and how you can switch between them.

:::tip EKF2 is the default and should be used unless you have a reason not to (in particular on vehicles with a GNSS/GPS). The Q-Estimator can be used if you don't have GPS, and is commonly used in multicopter racers. :::

Available Estimators

The available estimators are:

  • EKF2 attitude, position and wind states estimator (recommended) - An extended Kalman filter estimating attitude, 3D position / velocity and wind states.

  • LPE position estimator (deprecated) - An extended Kalman filter for 3D position and velocity states.

    :::warning LPE is deprecated. It works (at time of writing, in PX4 v1.14) but is no longer supported or maintained.

:::

  • Q attitude estimator - A very simple, quaternion based complementary filter for attitude. It does not require a GPS, magnetometer, or barometer.

How to Enable Different Estimators

To enable a particular estimator enable its parameter and disable the others:

  • EKF2_EN - EKF2 (default/recommended)
  • ATT_EN - Q Estimator (quaternion based attitude estimator)
  • LPE_EN - LPE (not supported for Fixed-wing)

:::warning It is important to enable one, and only one, estimator. If more than one is enabled, the first to publish the UOrb topics vehicle_attitude or vehicle_local_position is used. If none are enabled then the topics are not published. :::

:::info For FMU-v2 (only) you will also need to build PX4 to specifically include required estimator (e.g. EKF2: make px4_fmu-v2, LPE: make px4_fmu-v2_lpe). This is required because FMU-v2 is too resource constrained to include both estimators. Other Pixhawk FMU versions include both. :::