Lane Course Estimation via Vehicle Trajectory Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing lane detection methods in vehicles are limited by the visibility of lane markings and rely heavily on camera or LIDAR systems, which struggle in poor weather conditions or without markings, and can be erroneous due to movements of vehicles ahead or incorrect trajectory tracking.

Innovation Solution

A method and system that estimate the lane course by detecting and analyzing the trajectories of other vehicles moving in the same direction, assigning these trajectories to possible lanes, and selecting the most accurate lane based on trajectory data, while filtering out outliers and lane changes to provide a reliable and accurate lane estimation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If lane detection is based on camera or LIDAR systems detecting lane markings, then lane course can be recognized under optimal conditions, but detection range is limited and reliability deteriorates in poor visibility conditions or when markings are absent

Engineering Contradiction:
Improvelane detection precisionVSAvoidlane detection reliability
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent uses other vehicles as intermediary objects to infer lane course information. Instead of directly detecting lane markings, the system tracks trajectories of surrounding vehicles and uses their collective movement patterns as a mediator to determine the actual lane course, especially when markings are invisible or unreliable

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent inverts the traditional approach by not detecting lane markings directly, but rather detecting vehicle trajectories and inferring lane markings from them. This inversion allows the system to determine lane course indirectly through vehicle behavior patterns rather than direct visual detection

Inventive Principle:
Principle #13The other way round (Inversion)

2Adaptability or versatility

If lane course is determined solely by trajectory of a single vehicle driving ahead, then lane course can be estimated when a vehicle is present, but accuracy deteriorates when the vehicle changes lanes or turns

Engineering Contradiction:
Improvelane estimation capabilityVSAvoidlane course accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent merges trajectory data from multiple vehicles into a collective evaluation. By combining trajectory information from several vehicles and analyzing their collective movement patterns, the system achieves more robust lane course determination that is not easily affected by individual vehicle maneuvers such as lane changes or turns

Inventive Principle:
Principle #5Merging (Combining)

3Reliability

If trajectories of multiple vehicles are recorded and evaluated collectively, then lane course estimation becomes more robust, but complexity of trajectory assignment and lane change detection increases

Engineering Contradiction:
Improvelane estimation reliabilityVSAvoidtrajectory evaluation complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the evaluation process into distinct stages: trajectory recording, trajectory assignment to reference vehicles, collective evaluation, and lane change detection. This segmentation allows complex multi-vehicle trajectory analysis to be broken down into manageable processing steps, reducing overall system complexity while maintaining reliability

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3094530B1Method and system for estimating a course of a traffic lane
Publication Date: 2023.05.31 BAYERISCHE MOTOREN WERKE AG
  • EP3094530B1 patent drawingFigure 1~5
  • EP3094530B1 patent drawingFigure 2
  • EP3094530B1 patent drawingFigure 3A~3C

AI summary

In a method for estimating a course of a traffic lane for an ego-vehicle which travels on a road with multiple lanes, the following steps are repeatedly executed: identification of one or more other vehicles moving in the same direction on the road as the ego-vehicle; determination of vehicle positions for each of the other vehicles identified; formation of a trajectory for each of the other vehicles identified based on the respectively determined vehicle positions; assignment of the trajectories to possible traffic lanes based on a course of the trajectories formed in the environment of the ego-vehicle; selection of one of the possible traffic lanes to which at least one trajectory is assigned; and estimation of the course of the traffic lane based on the at least one trajectory assigned to the selected traffic lane.