Measurement Rejection in Aided Navigation Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inertial navigation systems (INS) that combine information from inertial sensors and global navigation satellite systems (GNSS) face significant accuracy degradation due to erroneous GNSS signals in environments like urban canyons, where multipath reflections and obstructions lead to inconsistent measurements, affecting sensor fusion algorithms.
Innovation Solution
A system with a processor and inertial sensors that employs a combined measurement validation methodology using a normalized measurement residual monitor and a Chi-squared monitor to select between aiding measurement validation processes, rejecting inconsistent GNSS measurements and updating navigation error states accordingly, thereby improving navigation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If GNSS measurements are used in sensor fusion algorithms, then navigation solution is available, but navigation accuracy significantly degrades due to erroneous GNSS signals in urban canyon environments
Solution Approach 1:
The system performs preliminary validation of GNSS measurements before they are used in the sensor fusion algorithm. By computing measurement residuals and comparing them against thresholds (using normalized residual monitor or Chi-squared monitor), the system proactively identifies and rejects erroneous measurements before they can degrade the navigation solution accuracy.
Solution Approach 2:
The patent introduces an intermediary validation layer between the GNSS receiver and the sensor fusion algorithm. This intermediary component computes measurement residuals and determines whether GNSS measurements are consistent with the sensor fusion model, acting as a mediator that filters out erroneous measurements while allowing valid ones to pass through to the navigation solution.
2Measurement precision
If measurement validation processes are implemented to reject erroneous GNSS measurements, then navigation accuracy is improved, but device complexity increases
Solution Approach 1:
The validation system monitors changes in measurement residual parameters and compares them against predefined thresholds. By changing the monitoring approach from direct measurement acceptance to residual-based parameter comparison, the system achieves accurate error detection without requiring complex validation logic. The use of standardized statistical monitors (normalized residual and Chi-squared) provides a systematic parameter-based approach to measurement validation.
Solution Approach 2:
The patent extracts the measurement validation function as a separate, modular component that operates independently from the main sensor fusion algorithm. By taking out the validation logic into a distinct measurement monitoring module, the system improves navigation accuracy through rigorous checking while keeping the overall device complexity manageable through functional separation and modularity.
Data Source
AI summary
A system comprises a processor onboard a vehicle, an onboard aiding source, and onboard inertial sensors. The processor includes a prediction module that propagates navigation error states, and a residual computation module that computes a measurement residual and measurement residual variance based on aiding measurements and navigation error states. A measurement monitoring module comprises a measurement monitor selection switch that selects between a normalized measurement residual monitor, and a Chi squared measurement monitor. When selected, the normalized measurement residual monitor performs a first aiding measurement validation process that determines whether a normalized measurement residual is less than a first user selected threshold. If yes, then the aiding measurement is valid. When selected, the Chi squared measurement monitor performs a second aiding measurement validation process that determines whether a measurement residual error squared is less than a second user selected threshold. If yes, then the aiding measurement is valid.