Lever Arm Estimation

Introduction

Qinertia features a powerful module used to estimate and refine automatically mechanical installation parameters such as the GNSS primary and secondary antenna lever arms. You can use this module if you have no way to precisely measure lever arms or alignment but also to check entered settings consistency.

The automatic estimation algorithm needs enough dynamics and good GNSS conditions to observe correctly the lever arms. Please read the dedicated Lever arms & alignment calibration article to get more details.

General Concepts

SBG Systems Inertial Navigation System (INS) as well as Qinertia can estimate several mechanical installation parameters such as the GNSS primary and secondary lever arms. This estimation is done using additional observation states in the INS Extended Kalman Filter (EKF).

To estimate lever arms and alignments, a specific mode has to be enabled in the navigation filter. This is required to offer the best robustness and accuracy when lever arms are perfectly measured and known.

When this specific mode is enabled, the navigation filter will try to find the best lever arms and alignment parameters that minimize INS / GNSS position discrepancies. In this mode, the overall INS solution is less accurate and robust so you should only use this mode to estimate mechanical parameters and not to get an INS solution.

Always try to input an initial guess for all lever arms. The more accurate the initial guess is the better and faster the estimation will be.

Acquisition Pattern

The navigation filter need some motion and dynamics to correctly estimate lever arms and alignment parameters. For example, if there is no roll/pitch motion it will be almost impossible to correctly estimate the lever arm vertical component.

Moreover, the INS need an accurate RTK/PPK position to refine correctly the lever arms. The INS tries to estimate lever arms with an accuracy of 1 cm so the INS obviously need an input position with the same level of accuracy.

SBG Systems strongly recommends to have a dedicated acquisition and processing to estimate the lever arms using figure of 8-patterns in a good open sky environment.

Real Time vs Post Processed

All SBG Systems Inertial Navigation Systems have a built in mechanical installation parameters estimation module so you can do it in real time. However, using Qinertia with a post processed solution is much more efficient. Post processing can process the same acquisition in forward and backward directions and process it three times. As a result, the lever arm can converge to the right values much faster than using a real time solution.

More importantly, if you can only apply low dynamics and motions, such as for a big vessel, Qinertia can estimate the lever arms accurately while the real time solution could just fail. Post processing can also be mandatory if you can't have in real time a RTK position quality.

Finally, Qinertia gives more detailed quality control indicators to assess the estimation accuracy and reliability.

Using Qinertia GUI

This is the easiest and most common way to use Qinertia to estimate mechanical installation parameters such as lever arm and alignment.

Project Settings

Qinertia can estimate several mechanical installation parameters such as GNSS antennas lever arm, odometer lever arm, DVL alignment and lever arm. You first have to configure the project correctly to indicate which parameters have to be estimated and which have been entered accurately and don't need to be estimated.

This is done from the INS Project Settings in each aiding panel. The screenshot below illustrates this for GNSS aiding and you can indicate if the primary and secondary GNSS lever arms have to be estimated or not.

Rough lever arm

Even if Qinertia auto estimation module is very powerful, try to enter initial lever arm guess as close as possible to real values to improve estimation robustness and quality.

Process the mission

When you start an INS processing, Qinertia will enable the automatic estimation for each parameter that has been indicated as being not accurate. During this specific processing, Qinertia will try to estimate the mechanical installation parameters and report the refined lever arms with associated quality metrics. The screenshot below shows the results for a GNSS primary and secondary lever arms estimation.

The default display shows easy to understand quality metrics computed using the estimated parameter standard deviation and stability. If you click on the Quality label, you can access these advanced metrics as shown on the screenshot below:

Apply the estimated parameters

During the auto estimation processing, Qinertia focus on lever arm estimation and not on the INS solution accuracy and reliability. You shouldn't use the INS solution that has been computed for the automatic estimation but rather launch a new processing with the lever arms refined by Qinertia.

If you are happy with the estimated lever arms and alignments, simply check the use checkbox to update the project configuration with the new values. You can then launch right away the final INS computation or close the dialog box and do it later by yourself

Using Command Line Interface

Qinertia also provides a full featured Command Line Interface (CLI) that can be used to automate mechanical installation parameters estimation. For each mechanical parameter to estimate, the preciseSetup flag has to be set to false in the JSON process file.

You can then use Qinertia CLI to process the project and an additional block in the processing Session report JSON file is added with all details regarding each estimated parameter.