Pose Estimation Uncertainty Scaling for Transportation Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating the pose of a transportation vehicle do not effectively differentiate between accurate and inaccurate or implausible pose estimations, leading to poor fusion results in sensor data processing.
Innovation Solution
A method that scales the uncertainty estimation of pose estimations based on a comparison with a priori information, such as a comparison trajectory, to weight and filter out implausible pose estimations, ensuring that accurate pose estimations contribute more significantly to the overall result.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing methods for estimating pose are used, then the pose estimation process is simple, but the accuracy of pose estimation is poor due to inability to differentiate between accurate and inaccurate estimations
Solution Approach 1:
The patent applies preliminary action by comparing pose estimations with a priori information (such as comparison trajectories from maps or previous measurements) before final fusion. This preliminary comparison allows the system to identify and weight implausible estimations in advance, improving accuracy without adding complex real-time processing during the fusion step itself.
Solution Approach 2:
The patent changes the parameter of uncertainty estimation by scaling it based on the comparison between measured pose and a priori information. When the measured pose deviates significantly from expected values, the uncertainty parameter is increased, effectively reducing the weight of implausible estimations in the fusion process.
2Reliability
If all pose estimations are treated equally in fusion, then the fusion process is straightforward, but the reliability of the result deteriorates due to inclusion of outlier measurements
Solution Approach 1:
The patent applies local quality by assigning different uncertainty levels to different pose estimations based on their local consistency with a priori information. Instead of treating all estimations uniformly, each estimation is evaluated individually against expected values, and its uncertainty is scaled accordingly, allowing the fusion process to weigh each estimation appropriately.
Solution Approach 2:
The patent uses feedback by continuously comparing pose estimations with a priori information and using this comparison to adjust uncertainty levels. This feedback loop ensures that estimations consistent with expected behavior are weighted more heavily, while outliers are automatically downweighted, improving fusion reliability.
3Measurement precision
If uncertainty estimation is not scaled, then the processing is computationally efficient, but the representation of true certainty level is inaccurate
Solution Approach 1:
The patent applies partial action by selectively scaling uncertainty only for estimations that show significant deviation from a priori information. Instead of uniformly processing all estimations with complex scaling, the system focuses computational effort on identifying and adjusting uncertain cases, maintaining efficiency while improving accuracy.
Data Source
AI summary
A method, a device, and a computer-readable storage medium with instructions for determining the location of a datum detected by a transportation vehicle wherein at least one pose estimation is ascertained. An uncertainty of the at least one pose estimation is determined wherein the uncertainty of the pose estimation includes a process of scaling an uncertainty estimation of the pose estimation, wherein the scaling process is based on a comparison of the pose estimation with a priori information. The at least one pose estimation is fused solely with at least one additional pose estimation with a weighting according to the uncertainties.