Heading Error Partitioning in Low-Performance IMU Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inertial navigation systems with low-performance inertial measurement units (IMUs) and relative aiding sensors are prone to filter inconsistency, particularly in heading estimates, due to the inability of relative aiding sensors to independently register in inertial space, leading to rapid growth of heading errors without absolute aiding sources.
Innovation Solution
The repartitioned model partitions the heading error into two separate states: an initial-heading error and an accumulated heading-change error, decoupling them from the relative aiding measurements and ensuring they correspond to navigation quantities explicitly computed in the INS, thereby reducing divergence and improving consistency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a low-performance IMU with only relative aiding sensors is used, then device complexity and cost are reduced, but heading accuracy deteriorates rapidly over time due to filter inconsistency
Solution Approach 1:
The patent segments the heading error into two distinct components: initial-heading error and accumulated heading-change error. This segmentation allows each error type to be modeled and corrected independently, preventing the filter inconsistency that would otherwise cause rapid degradation of heading accuracy in low-performance IMU systems
Solution Approach 2:
The patent changes the parameter representation by introducing separate error states for initial heading and accumulated heading change. This parameter transformation enables the filter to properly account for different error sources, maintaining heading accuracy despite using low-performance sensors
2Device complexity
If relative aiding sensors are used instead of absolute aiding sensors, then device complexity is reduced, but filter consistency deteriorates leading to rapid heading error growth
Solution Approach 1:
The patent introduces an intermediary computational framework that processes relative aiding measurements through a repartitioned error model. This intermediary structure transforms relative sensor data into consistent heading estimates by separating initial and accumulated error components, maintaining filter consistency without requiring absolute aiding sensors
3Loss of information
If infrequent absolute aiding is used, then loss of information from continuous absolute measurements is avoided, but heading accuracy deteriorates due to error accumulation
Solution Approach 1:
The patent applies preliminary correction by separating and initially correcting the initial-heading error component before processing accumulated heading changes. This preliminary action prevents error propagation and maintains accuracy even when absolute aiding measurements are infrequent
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of mitigating filter inconsistency of a heading estimate in an inertial navigation system is provided. The method includes inputting inertial measurements from at least one low-performance inertial measurement unit (IMU) in the inertial navigation system; inputting an initial-heading from at least one heading-information source; inputting an initial-heading uncertainty level associated with an error in the inputted initial-heading; initializing an initial-heading estimate with the inputted initial-heading; initializing an initial-heading uncertainty estimate with the inputted initial-heading uncertainty level; initializing an accumulated heading-change estimate to zero at a startup of the at least one low-performance IMU; and initializing an accumulated heading-change uncertainty estimate with an initial accumulated heading-change uncertainty level that is a value less than the inputted initial-heading uncertainty level; and periodically updating the accumulated heading-change estimate with the inertial measurements input from the at least one low-performance IMU.