Bayesian Estimation and Tracking: A Practical Guide

20.3 The Sensor Fusion Observational Model

The task is to define the functional dependence for the transformation of the state vector x_n into the components of the sensor fusion observation vector z_n. If the observation noise is included, this transformation can be written as

(20.2)

where . The observations consist of two separate components, the IMU components and the photogrammetric components, which must be fused together to provide an asynchronous stream of observation data.

20.3.1 The Inertial Measurement Unit Component of the Observation Model

The IMU package outputs data from three orthogonal accelerometers and three orthogonal rate gyroscopes. The IMU observation vector at time t_n can be written as

(20.3)

where z_x,n, z_y,n, and z_z,n are the accelerations (in G's) measured by the x-forward, y-right, and z-down accelerometers, respectively, and z_ϕ,n, z_θ,n and z_ϕ,n are rotation rates (in degrees) measured by the roll (counterclockwise about the x-axis), pitch (counterclockwise about the y-axis), and yaw (counterclockwise about the z-axis) rate gyroscopes, respectively. Note that since the IMU package is attached to the rigid body, it rotates as the body rotates so that the measurements ...