Global Positioning Systems, Inertial Navigation, and Integration

Chapter 9.5: INERTIAL SYSTEMS TECHNOLOGIES: SYSTEM-LEVEL ERROR MODELS

9.5 SYSTEM-LEVEL ERROR MODELS

The system-level implementation models discussed in the previous section are for the internal implementation of the inertial navigation system, itself. These are models for the peculiarities of the sensors and software that contribute to navigation errors. They are used in INS design analysis for predicting performance as a function of component characteristics. They are also used within the implementation for compensating the corrupting influence of sensor and software error tolerances on the measured and inferred vehicle dynamic variables.

The system-level error models in this section are for implementing GNSS/INS integration. These models represent how the resulting navigation errors will propagate over time, as functions of the error parameters of the inertial sensors. They are used in two ways:

  1. In so-called "loosely coupled" approaches for keeping track of the uncertainty in the INS navigation solution to use in a Kalman gain for combining the GNSS and INS navigation solutions to maintain an integrated navigation solution. When GNSS signals are not available, the model can still be used to propagate the estimated INS errors and subtract them from the uncompensated INS navigation solution. The resulting compensated INS navigation solution can then be used to speed up detection and reacquisition of GNSS signals if and when they become available again.
  2. In more "tightly coupled" approaches using GNSS measurements to estimate and compensate for random variations in the calibration parameters of individual sensors in the INS. These approaches continually re-calibrate the INS when GNSS signals are available. They are functionally similar to the loosely coupled approaches, in that they still carry forward the calibrationcompensated INS navigation solution when GNSS signals are unavailable. Unlike the loosely coupled approaches, however, they are feedback-based and not as sensitive to modeling errors.

Model Diversity There is no universal INS error model for GNSS/INS integration, because there is no standard design for an INS. There may be differences between different GNSS systems, and between generations of GPS satellites, but GNSS error models are all quite similar. Differences between error models for INSs, on the other hand, can be anything but minor. There are some broad INS design types (e.g., gimbaled vs. strapdown), but there are literally thousands of different inertial sensor designs that can be used for each INS type.

Methodology We present here a variety of inertial system error models, which will be sufficient for many of the sensors in common use, but not for every conceivable inertial sensor. For applications with sensor characteristics different from those used here, the use of these error models in GNSS/INS integration will serve to illustrate the general integration methodology, so that users can apply the same methodology to GNSS/INS integration with other sensor error models, as well.

UNLIMITED FREE
ACCESS
TO THE WORLD'S BEST IDEAS

SUBMIT
Already a GlobalSpec user? Log in.

This is embarrasing...

An error occurred while processing the form. Please try again in a few minutes.

Customize Your GlobalSpec Experience

Category: Accelerometers
Finish!
Privacy Policy

This is embarrasing...

An error occurred while processing the form. Please try again in a few minutes.