Hybrid Positioning and Sensor Integration - Global Navigation Satellite Systems: Signal,Theory and Applications - page 343

3. Multi sensor integration
Navigation is the continuous positioning which is the process of monitoring and controlling
the movement of a vehicle from one place to another. Inertial navigation is the self-
determination of the instantaneous position and other parameters of motion of a vehicle by
measuring specific force, angular velocity, and time in a previously selected coordinate
system. The basic concept is to determine the vehicle velocity and position by real time
integration of governing differential equations.

Figure. 1 shows an overview of data processing of sensor integration from navigation as a
hybrid positioning to mapping by direct geo-referencing with a laser scanner. In this paper,
the following data are acquired and those data are integrated: base station GPS data, remote
station GPS data, IMU data, digital images, and laser range data. Although the data are
acquired in different frequencies, they are synchronized with each other by GPS time.

First, differential or kinematic GPS post processing is conducted. Second, the processed GPS
data and the IMU data are integrated by a Kalman filter to estimate the sensor trajectory.
The bundle block adjustment (BBA) of the digital images is then made to acquire geo-
referenced images and exterior orientations with the support of GPS and IMU data, which
are the sensor position and attitude, as an external aid. Also, VMS and other sensors can be
considered as external aids if GPS accuracy is not enough to support IMU in urban areas.
Finally, GPS data, IMU data, and the external aides are combined to regenerate high-
precision and time-series sensor position and attitude. Finally, these hybrid positioning data
are used for the geo-referencing of the laser range data and for the construction of a digital
surface model as 3D point cloud data.

Laser range
data

Digital images
(external aids)

IMU data

GPS data

Post-processing GPS data Pure navigation process

Integration of GPS/IMU by Kalman filtering

Bundle block adjustment to estimate sensor orientations

Trajectory estimation by integration of GPS/IMU data and External Source

Geo referencing of laser range data

Construction of DSM

Fig. 1. Overview of data processing for multi sensor integration

4. Hybrid positioning

In general, IMUs exhibit position errors that tend to increase with time in an unbounded
manner. This degradation occurs due to errors of initialization of IMUs and inertial sensor