Relative Preintegration for Inertial Navigation Drift Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inertial navigation systems face errors due to accumulated drift over time, requiring frequent initial conditions for integration, which increases operational complexity and reduces accuracy, especially when used for long distances or with vibrations, and preintegration methods still require coordinate transformations for error correction.
Innovation Solution
An inertia-based navigation apparatus and method using relative preintegration, where a first sensor detects motion and a second sensor measures inertia data, allowing for error correction without coordinate transformation by calculating preintegration values and using external correction values to estimate and correct pose information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If inertial navigation system uses integration of acceleration and velocity components, then position and velocity can be calculated continuously, but navigation errors accumulate over time
Solution Approach 1:
The patent applies preintegration by calculating preintegration values in advance for a preset time period. This preliminary action allows the system to prepare navigation data ahead of time, reducing the need for continuous real-time integration and thereby minimizing error accumulation while maintaining continuous navigation capability.
Solution Approach 2:
The patent uses pose information from a first sensor (such as GPS or visual odometry) as feedback to correct the pose information calculated from inertial measurement unit (IMU) data. This feedback mechanism continuously compensates for drift and accumulation errors, maintaining high navigation accuracy over extended periods.
2Measurement precision
If new initial condition is used every time for integration, then gravity information can be removed, but operation speed decreases remarkably
Solution Approach 1:
The patent precalculates and stores preintegration values for gravity removal in advance. Instead of performing gravity removal calculations every time a new initial condition is needed, the system uses the precomputed values, significantly reducing computational load and increasing operation speed while maintaining accurate gravity removal.
3Measurement precision
If coordinate transformation is performed to compare pose information from different sensors, then pose information can be compared, but processing time increases
Solution Approach 1:
The patent introduces a coordinate transformation matrix as an intermediary that pre-establishes the relationship between different coordinate systems. By using this precomputed transformation matrix, the system can quickly compare pose information from different sensors without performing complex real-time coordinate transformations, thus reducing processing time while maintaining comparison accuracy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces error accumulation and operational complexity by directly comparing pose information from different sensors, enhancing navigation accuracy and speed without the need for frequent initial conditions or coordinate transformations.
Implementation Method 1
a first sensor detecting movement of a moving body and outputting motion information about the moving body
Implementation Method 2
a second sensor measuring inertia data which is information about a rotational acceleration and a translational acceleration of the moving body
Implementation Method 3
calculating a preintegration value by preintegrating the inertia data of the second sensor
Implementation Method 4
comparing second pose information, which is pose information of the moving body obtained by the second sensor, with first pose information, which is pose information of the moving body obtained by the first sensor without coordinate transformation
Data Source
AI summary
An inertia-based navigation apparatus and an inertia-based navigation method based on relative preintegration are provided. The inertia-based navigation apparatus includes: a first sensor detecting and outputting motion information about a moving body which is moving, based on a first coordinate system; a second sensor detecting and outputting inertia data about a translational acceleration and a rotational angular velocity related to the movement of the moving body, based on a second coordinate system; and a controller determining, at every first time, pose information about a position, a velocity and an attitude of the moving body in a reference coordinate system, based on the motion information and the inertia data.