Lane Line Estimation Using Map-Matched Telemetry and Point Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles struggle to accurately estimate the location of lane lines within roadways, which is crucial for maintaining lane positioning and navigation.
Innovation Solution
A method and system utilizing telemetry data, GPS data, and road-network topology data to estimate lane lines through Gaussian Mixture Model (GMM) and Lane-Point Detection (LPD) models, combined with synthetic gaussian oversampling undersampling technique (SGOUT) and lane-point clustering (LPC), to determine the shape, location, and extent of lane lines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional lane line estimation methods are used, then the system complexity is low, but the measurement precision of lane line location is insufficient
Solution Approach 1:
The patent segments the lane line estimation process into multiple specialized modules: GMM-based point detection for initial lane point identification, LPD model for refined lane line detection, and LPC model for clustering and smoothing. Each module handles a specific aspect of the estimation task, improving overall precision while keeping individual module complexity manageable.
Solution Approach 2:
The patent introduces map-matched telemetry data as an intermediary element that bridges raw GPS data and final lane line estimation. This intermediary layer provides contextual information from pre-built maps to guide and constrain the estimation process, improving accuracy without requiring direct complex processing of raw sensor data alone.
2Measurement precision
If multiple processing models (GMM, LPD, LPC) are applied to improve lane line estimation accuracy, then the measurement precision improves, but the computational time increases
Solution Approach 1:
The patent performs preliminary actions by first generating map-matched telemetry data and identifying candidate lane points using GMM before applying the more computationally intensive LPD and LPC models. This preliminary filtering reduces the data volume and complexity for subsequent processing stages, optimizing the overall computational efficiency.
Solution Approach 2:
The patent applies multiple processing models (GMM, LPD, LPC) in sequence, where each model processes only the necessary portion of data required for its specific function. This partial action approach avoids redundant processing while maintaining high accuracy through the cumulative effect of specialized models.
3Reliability
If map-matched telemetry data is used to improve lane line estimation, then the reliability of lane detection improves, but the device complexity increases
Solution Approach 1:
The patent makes the map-matched telemetry data serve multiple functions: it provides geographic context for lane estimation, constrains the search space for lane points, validates detected lane lines against known road geometry, and enables smoothing through clustering. This multi-functionality improves reliability without proportionally increasing system complexity.
Data Source
AI summary
A method for estimating lane lines receiving, by a controller, telemetry data and map data. The method further includes matching the telemetry data with the map data to generate map-matched telemetry data and determining, using a gaussian mixture model (GMM), a first set of lane line points relating to the lane lines of the roadway within the predetermined region. Also, the method includes determining, using a lane-point detection (LPD) model, a second set of lane line points relating to the lane lanes of the roadway within the predetermined region and determining, using a lane-point clustering (LPC) model, a plurality of lane lines of the roadway based on the first set of lane line points and the second set of lane line points.