Lane Course Determination Using Compensating B-Spline Curves
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current driver assistance systems face challenges in accurately determining lane geometry, especially at intersections, due to the noisy measurements provided by simple GPS receivers, which are insufficient for precise lane course inference from recorded journeys.
Innovation Solution
A method using a compensating B-spline curve is employed to determine the lane course from multiple noisy recorded journeys, where the curve is iteratively refined based on position information from various routes, independent of their order, to produce a more accurate and representative lane course, suitable for safety-critical systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If simple GPS receivers are used to record journeys, then device complexity is reduced and cost is lowered, but measurement precision deteriorates due to noisy position measurements
Solution Approach 1:
The patent combines position measurements from multiple different journeys into a single dataset. By merging data from many GPS recordings that all pass through the same lane, the system accumulates sufficient position information to accurately determine the lane course despite each individual measurement being noisy.
Solution Approach 2:
The patent creates a mathematical model (compensating curve) that copies the essential geometric characteristics of the lane from numerous noisy measurements. The B-spline curve serves as an idealized copy of the actual lane path, filtering out measurement noise while preserving the true lane geometry.
2Ease of manufacture
If lane course is determined from individual noisy journeys, then data processing is simplified, but manufacturing precision deteriorates as exact lane geometry cannot be inferred
Solution Approach 1:
The system merges position information from multiple journeys that traverse the same lane. By combining these datasets and identifying common spatial patterns, the method achieves accurate lane geometry determination without requiring complex processing of each individual journey separately.
Solution Approach 2:
The patent transforms the problem from determining lane geometry from single-journey data to determining it from multi-journey aggregated data. This parameter change in the input data characteristics enables accurate lane course inference while maintaining computational feasibility through the B-spline modeling approach.
3Measurement precision
If precise position determinations are required for accurate lane modeling, then lane course accuracy is improved, but device complexity increases as simple GPS receivers become insufficient
Solution Approach 1:
Instead of relying on complex hardware to provide precise position data, the patent creates a mathematical copy (B-spline curve) of the lane geometry from aggregated noisy measurements. This computational copying approach achieves high accuracy without requiring sophisticated positioning hardware.
Solution Approach 2:
The patent replaces the need for complex precision positioning hardware with a computational data processing system. By substituting mechanical/physical precision requirements with mathematical modeling and statistical aggregation, the system achieves accurate lane determination using simple GPS receivers.
4Measurement precision
If multiple journeys are processed together to reduce noise, then measurement precision is improved, but loss of time increases due to additional data processing
Solution Approach 1:
The patent changes the processing approach from analyzing each journey separately to simultaneously processing multiple journeys as a unified dataset. This parameter change in the processing methodology enables noise reduction through aggregation while the B-spline fitting algorithm maintains computational efficiency.
Solution Approach 2:
The system creates a single mathematical representation (compensating curve) that captures the essential lane geometry from all journeys. This copying approach consolidates multiple datasets into one model, reducing the overall processing time compared to analyzing each journey individually and then combining results.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention discloses a method for determining the course of a lane with the aid of a multiplicity of recorded routes using a compensation curve, wherein each route of the multiplicity of routes comprises individual indications of the position of the vehicle in the order of passing during the particular journey; wherein the method comprises: selecting a route from the multiplicity of routes; determining the number of checkpoints for the compensation curve on the basis of the selected route; determining an initial compensation curve on the basis of the number of checkpoints and the selected route from the multiplicity of routes; repeating the following step until an abort criterion has been satisfied: determining a new compensation curve using the previously determined compensation curve and individual position indications of the routes of the multiplicity of routes, wherein the process of determining the new compensation curve is independent of the order of position indications in one of the respective routes and is independent of the fact that a position determination used is assigned to a route from the multiplicity of routes; outputting the compensation curve determined last as the lane course.