Route Matching for Autonomous Driving Using HD Map Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The accuracy of automatic driving routes based on high definition maps is low due to unclear correspondence between standard definition maps and high definition maps, leading to random selection of road sections which may not match user needs.

Innovation Solution

A route processing method that determines the matched road section in the high definition map only for sections with unique mapping relationships, filling in uncertain sections based on unique correspondences to improve route accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If random selection is used for road sections with unclear correspondence between standard definition map and high definition map, then the route processing can be completed, but the accuracy of the matched route is low

Engineering Contradiction:
Improveroute matching accuracyVSAvoidroute processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the route matching process into two distinct parts: (1) road sections with clear one-to-one correspondence that are matched directly, and (2) road sections with unclear correspondence that require alternative processing. This segmentation allows the system to apply different strategies to different portions of the route, improving overall accuracy while managing complexity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Instead of directly matching unclear road sections through random selection, the patent inverts the approach by: (a) identifying and matching all clear one-to-one sections first, (b) using the matched sections as reference points, and (c) inferring the uncertain sections based on the established correspondence pattern. This inversion transforms an uncertain matching problem into a constrained inference problem.

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

2Measurement precision

If direct mapping is used between standard definition map and high definition map, then the processing speed is fast, but the accuracy is low due to errors between the two maps

Engineering Contradiction:
Improveroad section correspondence accuracyVSAvoidroute processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies partial action by not attempting to resolve all mapping uncertainties. Instead, it focuses computational resources only on road sections with clear one-to-one correspondence, accepting that some sections will remain uncertain. This partial resolution approach maintains processing speed while improving accuracy for the critical matched sections.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The patent introduces an intermediary verification mechanism that checks whether a road section has clear one-to-one correspondence before attempting direct mapping. This intermediary step acts as a filter, allowing fast direct mapping for verified sections while routing uncertain sections through alternative processing, thus balancing speed and accuracy.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If all road sections are attempted to be matched between standard definition map and high definition map, then completeness is improved, but the reliability of the matched route decreases due to incorrect matches

Engineering Contradiction:
Improveroute matching reliabilityVSAvoidhandling of unclear correspondence
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent applies local quality by assigning different matching strategies to different road sections based on their local characteristics. Road sections with clear geometric and topological correspondence receive direct one-to-one matching, while sections with ambiguous characteristics receive alternative processing. This localized approach ensures high reliability for matched sections without forcing unreliable matches everywhere.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent performs preliminary classification of road sections into 'clear correspondence' and 'unclear correspondence' categories before executing the matching process. This preliminary action allows the system to prepare appropriate matching strategies in advance, ensuring that reliable direct matching is applied only where appropriate, thereby maintaining high overall reliability.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3967980B1Route processing method and apparatus
Publication Date: 2024.01.31 BEIJING BAIDU NETCOM SCI & TECH CO LTD
  • EP3967980B1 patent drawingFigure 1~2
  • EP3967980B1 patent drawingFigure 3~4
  • EP3967980B1 patent drawingFigure 5

AI summary

The present disclosure provides a route processing method and apparatus, which relates to the field of automatic driving in computer technology. The specific implementation includes: acquiring a vehicle driving route according to a standard definition map, where the vehicle driving route comprises multiple standard definition map road sections of the standard definition map; determining at least one first road section from the multiple standard definition map road sections according to a road section mapping relationship between the standard definition map and a high definition map, where each first road section has a unique corresponding target road section in the high definition map; determining the target road section corresponding to each first road section in the high definition map according to the road section mapping relationship; and determining a target driving route corresponding to the vehicle driving route in the high definition map according to the target road section corresponding to each first road section and the high definition map. By determining the target road section for the first road section with a clear correspondence, and then determining the target driving route according to each target road section and the high definition map, the accuracy of the determined matched route can be effectively improved.