Adaptive Stochastic Filtering for Vehicle Navigation Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle navigation systems, such as GPS, have limited accuracy, which is inadequate for applications like autonomous vehicle operation and can lead to unforeseen consequences due to potential flaws in navigation directions.
Innovation Solution
A method and system that use adaptive stochastic filtering to assess the accuracy of navigation systems by calculating the probability of an estimated vehicle position being within an error bound of a triangulated position, allowing for the detection of navigation system faults and triggering tasks when accuracy thresholds are not met.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional GPS navigation systems are used, then the system is simple and widely available, but the accuracy is limited to 2-3 meters which is inadequate for autonomous vehicle operation
Solution Approach 1:
The patent combines GPS triangulated position data with vehicle motion sensor data (accelerometers, gyroscopes, odometers) to create a hybrid navigation system. The GPS provides periodic position fixes while motion sensors provide continuous position estimation through dead reckoning, merging both data sources to achieve higher accuracy than GPS alone while maintaining system simplicity
Solution Approach 2:
The patent introduces an intermediary computational layer (filtering algorithm) that processes and fuses GPS data with motion sensor data. This intermediary processing layer reconciles the coarse GPS measurements with the continuous but drift-prone motion sensor data, producing a more accurate estimated position without requiring more complex hardware
2Reliability
If GPS accuracy is relied upon for autonomous navigation, then the navigation system operates autonomously, but the potential for unforeseen consequences increases due to navigation system flaws and errors
Solution Approach 1:
The patent implements a feedback mechanism where the system continuously monitors its own navigation accuracy by comparing the GPS-derived position with the motion-sensor-derived position. When the discrepancy exceeds a threshold, the system generates a fault flag and can trigger corrective actions, creating a self-monitoring loop that improves reliability without requiring external ground truth
Solution Approach 2:
The navigation system performs self-diagnosis and self-monitoring by using its own motion sensor data to validate GPS accuracy. The system serves its own validation needs by computing whether the estimated position based on motion data plausibly matches the GPS position within expected error bounds, eliminating the need for external reference systems
Data Source
AI summary
A method for monitoring performance of a navigation system onboard a vehicle is provided. The method obtains, at a vehicle electronic control unit (ECU), a triangulated vehicle location and vehicle motion data, the triangulated vehicle location obtained from the navigation system; computes an estimated vehicle location, using the triangulated vehicle location and the vehicle motion data; calculates a probability that the estimated vehicle location exists within an error bound of the triangulated vehicle location; and when the probability indicates that the estimated vehicle location does not exist within the error bound, performs a task onboard the vehicle.