Wheel-to-Lane Distance Estimation for Accurate Vehicle Pose
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous driving systems face challenges in accurately determining the distance between a vehicle and nearby lane markers due to perspective distortions in 2D image processing, which can lead to inaccurate vehicle pose estimation and lane-changing predictions.
Innovation Solution
The system generates wheel and lane segmentation maps from camera images, using deep learning techniques to detect wheels and lanes, and calculates the distance between the wheels and nearby lanes, providing a more accurate representation of the vehicle's position relative to the lane.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2D image processing is used to determine vehicle to lane distance, then the system is simple and computationally efficient, but the distance measurement accuracy deteriorates due to perspective distortions
Solution Approach 1:
The patent transitions from 2D image processing to 3D point cloud processing by using LiDAR data. The system generates a 3D point cloud representation of the environment and uses 3D bounding box parameters (length, width, height, roll, pitch, yaw) to accurately determine vehicle pose and distance to lane markers, eliminating perspective distortion issues inherent in 2D images.
2Measurement precision
If wheel segmentation is used to determine vehicle position, then the distance measurement accuracy improves, but the processing time and computational complexity increase
Solution Approach 1:
The system performs wheel segmentation and lane detection in advance to establish reference frameworks before calculating vehicle pose. By pre-identifying wheel positions in the point cloud and pre-detecting lane marker locations, the system creates ready-to-use reference data that speeds up the subsequent distance calculation process.
Solution Approach 2:
The patent introduces wheel segmentation maps and lane segmentation maps as intermediary representations. These maps serve as intermediate structures that bridge the raw sensor data and the final distance calculation, organizing spatial information in a computationally efficient format that facilitates accurate pose estimation.
Data Source
AI summary
A system and method for determining car to lane distance is provided. In one aspect, the system includes a camera configured to generate an image, a processor, and a computer-readable memory. The processor is configured to receive the image from the camera, generate a wheel segmentation map representative of one or more wheels detected in the image, and generate a lane segmentation map representative of one or more lanes detected in the image. For at least one of the wheels in the wheel segmentation map, the processor is also configured to determine a distance between the wheel and at least one nearby lane in the lane segmentation map. The processor is further configured to determine a distance between a vehicle in the image and the lane based on the distance between the wheel and the lane.