Reference Line Generation for Autonomous Driving

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing path planning systems face inefficiencies in generating reference lines due to the need for universal algorithms that fail to account for varying road scenarios and vehicle dynamics, leading to resource wastage and increased calculation time, especially in complex environments like parking lots with different driving difficulties.

Innovation Solution

Classifying path planning points based on driving difficulty, assigning geometric reference line generation for low-difficulty segments and algorithmic processing for high-difficulty segments, using the Hybrid A* algorithm to simulate reference lines, and combining these to form complete reference lines efficiently.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a single universal algorithm is used for all road segments with different driving difficulties, then the system can handle all scenarios uniformly, but system resources are wasted and calculation time increases

Engineering Contradiction:
Improveuniversal algorithm applicabilityVSAvoidreference line generation efficiency
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent segments the global path into multiple local path portions based on driving difficulty characteristics. Each segment is independently classified and processed with appropriate algorithms, avoiding the inefficiency of applying a universal algorithm to all segments while maintaining adaptability to different road scenarios.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different reference line generation methods to different local path portions based on their specific driving difficulty characteristics. Low-difficulty segments use geometric connection methods while high-difficulty segments use algorithmic processing, optimizing resource allocation and calculation efficiency for each local region.

Inventive Principle:
Principle #3Local quality

2Reliability

If algorithmic processing is used for all path planning points, then vehicle dynamics constraints are satisfied, but system resource consumption increases and calculation time is excessive

Engineering Contradiction:
Improvevehicle dynamics constraint satisfactionVSAvoidreference line calculation time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent applies algorithmic processing only to the extent necessary - specifically to high-difficulty local path portions where vehicle dynamics constraints are critical. Low-difficulty segments use simpler geometric methods, avoiding unnecessary computational overhead while ensuring reliability where needed.

Inventive Principle:
Principle #16Partial or excessive action

3Productivity

If geometric connection is used for all path nodes, then reference line generation is fast, but vehicle dynamics constraints and complex road conditions are not satisfied

Engineering Contradiction:
Improvereference line generation speedVSAvoidvehicle dynamics constraint satisfaction
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent applies geometric connection methods only to low-difficulty local path portions where simple geometric relationships suffice. For high-difficulty segments, algorithmic processing is used to ensure vehicle dynamics constraints are satisfied, optimizing the balance between speed and reliability based on local road conditions.

Inventive Principle:
Principle #3Local quality

4Reliability

If reference lines are updated every time a local map is received in complex parking lot scenarios, then the path planning remains current, but system resources are continuously consumed and calculation time accumulates

Engineering Contradiction:
Improvepath planning accuracyVSAvoidsystem resource efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent segments the path planning updates by processing only high-difficulty local path portions that require algorithmic processing, rather than recalculating all reference lines. This selective update approach maintains path planning accuracy in dynamic environments like parking lots while reducing unnecessary computational resource consumption.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3951321B9Method, system, terminal, and storage medium for rapid generation of reference lines
Publication Date: 2024.11.06 ZONGMU TECH SHANGHAI CO LTD
  • EP3951321B9 patent drawingFigure 1
  • EP3951321B9 patent drawingFigure 2
  • EP3951321B9 patent drawingFigure 3

AI summary

Provided are a method and system for rapid generation of a reference driving route, a terminal and a storage medium. According to the method, segmenting route planning points for road sections of different difficulties according to driving difficulty, and performing geometric processing on the parts of low driving difficulty so as to obtain reference driving lines; and processing parts of high difficulty by Hybrid A star algorithm in combination with vehicle dynamics constraint so as to obtain reference driving lines, and splicing the reference driving routes of all the parts to form a complete reference driving route. The present disclosure is low in system resource occupancy rate and short in time consumed by the algorithm.