Lane Map Altitude Smoothing via Factor Graph Optimization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Autonomous vehicles face challenges in accurately modeling roads due to altitude mapping errors in lane maps, which can lead to navigation issues and safety hazards.

Innovation Solution

A factor graph-based approach is used to encode constraints associated with realistic road characteristics, translating the problem into a nonlinear optimization to minimize and smooth altitude discontinuities in lane maps, utilizing binary flatness, translational, and GPS constraints to generate corrected maps.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional sensor data is used for road mapping, then mapping speed is maintained, but altitude accuracy deteriorates due to discontinuities and errors in lane maps

Engineering Contradiction:
Improvealtitude accuracyVSAvoidmapping system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The road map is divided into discrete segments with associated altitude values and uncertainties. Each segment can be independently processed and optimized, allowing the system to handle complex altitude corrections without overwhelming computational complexity. The segmentation enables targeted correction of discontinuities while maintaining overall mapping integrity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system transforms the mapping problem from fixed altitude values to optimized parameters that include altitude, uncertainty, and continuity constraints. By changing from simple altitude mapping to a parameter optimization framework, the system achieves higher accuracy while managing complexity through mathematical formulation rather than computational brute force.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If altitude smoothing is applied to correct discontinuities, then mapping accuracy improves, but processing time increases

Engineering Contradiction:
Improveroad modeling accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary identification and classification of altitude discontinuities before applying smoothing corrections. By pre-analyzing the map to locate problematic segments, the system can apply targeted corrections only where needed rather than processing the entire map uniformly, reducing overall processing time while maintaining accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The optimization process incorporates feedback mechanisms that adjust correction strength based on local map characteristics and uncertainty levels. The system continuously evaluates the impact of smoothing corrections and adapts the optimization parameters to achieve minimal processing time while maintaining sufficient accuracy for autonomous vehicle operations.

Inventive Principle:
Principle #23Feedback

3Manufacturing precision

If strict altitude continuity is enforced, then mapping precision improves, but adaptability to real road variations deteriorates

Engineering Contradiction:
Improvelane map precisionVSAvoidroad variation accommodation
Core Design Contradiction:
Manufacturing precisionVSAdaptability or versatility

Solution Approach 1:

The system applies different continuity constraints and correction strengths to different segments of the road map based on local characteristics. Areas with clear discontinuities receive strong smoothing, while areas with legitimate altitude variations (such as bridges or hills) maintain their local accuracy. This localized approach preserves precision where needed while accommodating real road variations elsewhere.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The altitude correction process is dynamic rather than static, allowing the system to adjust correction parameters based on real-time analysis of road characteristics. The optimization can adapt to varying road conditions, transitioning between strict continuity enforcement and flexible variation accommodation as needed, thereby maintaining both precision and adaptability.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4365549B1Altitude smoothing
Publication Date: 2025.09.03 EMBARK TRUCKS INC
  • EP4365549B1 patent drawingFigure 1
  • EP4365549B1 patent drawingFigure 2A
  • EP4365549B1 patent drawingFigure 2B

AI summary

A system and method including receiving a digital representation of a road that includes one or more lane representations; defining a factor graph including variable nodes and constraint nodes that encode lane boundary constraints for the lane map; translating the factor graph into a nonlinear optimization problem for altitudes of the lane map; generating a solution to the optimization problem; and generating a corrected lane map with an optimized altitude for the road based on the generated solution.