Road Structure Perception Fusion for Autonomous Driving Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automatic driving systems face challenges in accurately perceiving road structure information, such as lane lines, road edges, and stop lines, due to low accuracy in existing perception methods.
Innovation Solution
A method and apparatus that combine map data with a road structure perceiving model trained using machine learning, where images from different angles are used to generate a final perceiving result by determining a first perceiving result from map data and a second perceiving result from the model, and then fuse these results to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If only map data is used for road structure perception, then the applicable range is comprehensive, but the measurement precision is low
Solution Approach 1:
The patent combines map data-based perception with machine learning model-based perception into a unified fusion system. The first perception result from map data and the second perception result from the trained road structure perception model are merged through a fusion mechanism to produce a final perception result, thereby improving accuracy while maintaining comprehensive applicability
Solution Approach 2:
The perception system is segmented into two independent modules: one processing map data to generate the first perception result, and another using the trained machine learning model to generate the second perception result. This segmentation allows each module to specialize in its strength while the fusion mechanism integrates them to resolve the contradiction between accuracy and complexity
2Measurement precision
If only machine learning model is used for road structure perception, then the measurement precision is improved, but the applicable range is limited
Solution Approach 1:
The system merges the advantages of both map data-based perception (comprehensive applicability) and machine learning model-based perception (high accuracy). The fusion mechanism integrates results from both sources, ensuring the system maintains broad adaptability while achieving improved measurement precision through the combined approach
3Measurement precision
If multiple images from different angles are processed, then the measurement precision is improved, but the loss of time increases
Solution Approach 1:
The road structure perception model is trained in advance using multiple images from different angles during the training phase. This preliminary action pre-processes and learns from diverse angular data, enabling the model to make accurate predictions from single or fewer images during actual operation, thereby reducing real-time processing time while maintaining high accuracy
Data Source
AI summary
The present disclosure provides a method, an apparatus, an electronic device, a storage medium and a program product for perceiving a road structure, and relates to the technical field of artificial intelligence and, in particular, to the technical field of deep learning and automatic driving. A specific implementation includes: determining, based on map data, a first perceiving result characterizing a road structure around a current position; determining, by a road structure perceiving model trained in advance, a second perceiving result characterizing the road structure around the current position; and generating a final perceiving result characterizing the road structure around the current position based on the first perceiving result and the second perceiving result.