Vehicle Lane Selection via Trajectory Pairing and Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lane selection systems for motor vehicles face challenges due to measurement noise and reliance on current speed and yaw rate assumptions, leading to poor route selection, especially with road markings and boundaries providing ambiguous trajectory indications.
Innovation Solution
A control system that uses environmental sensors to identify and form pairs of road boundaries and markings, determining sampling points and calculating section lengths to improve lane selection, while also estimating trajectories of other vehicles based on detected data to reduce noise and enhance route choice accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If trajectory estimation is based on current speed and yaw rate assumptions, then the route selection can be performed quickly, but the measurement noise increases and route selection accuracy deteriorates
Solution Approach 1:
The system performs preliminary actions by capturing road boundaries and markings in advance to establish multiple possible trajectories before route selection is needed. These pre-captured trajectories are stored and can be quickly retrieved during route selection, eliminating the need for real-time trajectory estimation based on noisy sensor data while maintaining fast response times.
Solution Approach 2:
The system introduces an intermediary element - pre-captured road boundary and marking data - that mediates between the need for fast route selection and accurate trajectory estimation. Instead of directly estimating trajectories from noisy speed and yaw rate sensors, the system uses these pre-captured visual references as an intermediate reference frame for more accurate and faster route determination.
2Loss of information
If road markings and boundaries are used to indicate trajectory, then trajectory information is available, but the ambiguous indications lead to poor lane selection
Solution Approach 1:
The system segments the trajectory determination process by separating the capture of road boundaries/markings from the interpretation of trajectory. Multiple discrete road features are captured and stored as separate trajectory options, then evaluated against multiple criteria (consistency checks, pairing with other trajectories, geometric validity) to select the most accurate lane, rather than relying on ambiguous continuous markings.
Solution Approach 2:
The system implements feedback through consistency checks that evaluate whether captured road boundaries and markings form coherent, geometrically valid trajectories. Multiple candidate trajectories are generated and then validated against each other and against expected road geometry patterns, with feedback loops that refine the selection until the most consistent and accurate lane is identified.
Data Source
AI summary
Control system, which is adapted and determined for use in a motor vehicle to identify road boundaries and/or road markings based on environmental data obtained by at least one environmental sensor associated with the motor vehicle. The at least one environmental sensor is adapted to provide the environmental data, which reproduces the area in front of the motor vehicle, to an electronic control of the control system. The control system is at least adapted and determined to capture road boundaries and/or road markings with the at least one environmental sensor. A respective trajectory is determined from the captured road boundaries and/or road markings. Pairs are formed from the determined trajectories, wherein a pair comprises two trajectories respectively. Sampling points are determined at predetermined distances for each pair of first and second trajectories. Distances perpendicular to a course of the first trajectory are determined from the determined sampling points to the second trajectory and a length of a section is calculated along the course of the first trajectory, for which length the determined distances are within a predetermined value range, and/or distances perpendicular to a course of the second trajectory are determined from the determined sampling points to the first trajectory, and a length of a section is calculated along the course of the second trajectory, for which length the determined distances are within the predetermined value range. A set of pairs of trajectories is selected based on a criterion for pairing to determine at least one lane and/or at least one trajectory for the motor vehicle based on this set.