Vehicle Object Perception for Occluded Preceding Vehicle Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle perception systems struggle to quickly and accurately identify the nearest object in front of the host vehicle, particularly when the nearest object is obscured, leading to potential misidentification of surrounding objects.
Innovation Solution
The system employs a LiDAR sensor to detect candidate objects, determines ego-lane objects using yaw rate and speed, and identifies the first and second closest preceding objects through lateral position tracking and size conditions, ensuring accurate and rapid perception even when the nearest object is cut out.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the system relies on object size for determining the nearest object, then the identification process becomes simple, but the accuracy deteriorates when the nearest object is obscured or cut out
Solution Approach 1:
The system performs preliminary actions by determining both the first closest preceding object and the second closest preceding object in advance, before any occlusion occurs. This allows the system to quickly switch to the second closest object as backup when the nearest object is cut out, without requiring complex real-time reanalysis
Solution Approach 2:
The patent prepares compensatory measures in advance by identifying a second closest preceding object that can serve as backup. When the nearest object is obscured, this pre-identified backup object immediately compensates for the loss, preventing misidentification without requiring complex emergency processing
2Measurement precision
If the system uses multiple parameters (lateral position, size conditions, ego-lane determination) for object identification, then the accuracy improves, but the processing time increases
Solution Approach 1:
The system performs preliminary filtering and classification of candidate objects using lateral position and size conditions before final identification. By pre-processing candidate objects and organizing them into ego-lane and preceding object categories in advance, the system reduces the computational burden during critical identification moments
Solution Approach 2:
The patent segments the object identification process into distinct stages: candidate object detection, ego-lane object determination, preceding object identification, and closest object selection. Each stage applies specific filters and conditions, breaking down the complex task into manageable segments that can be processed efficiently
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach provides quick and accurate perception of the nearest preceding object, enhancing driving stability by promptly identifying the first and second closest objects, even in situations where the nearest object is obscured.
Implementation Method 1
detecting candidate objects using at least one sensor
Data Source
AI summary
The present disclosure relates to an object perception method and an object perception apparatus. An object perception method may include detecting candidate objects by using at least one sensor; determining objects from a vehicle among the candidate objects; determining preceding objects among the vehicle objects from the vehicle; and determining the first closest preceding object and the second closest preceding object among the preceding objects.