Sensor-Fusion Path Planning for Map-Free Lane Change Decisions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driver assistance systems for autonomous vehicles rely on high-precision vector maps, which are costly to create and maintain, and can lead to path planning errors if the map information does not match the actual environment, posing a safety risk.
Innovation Solution
A path planning system that uses a camera and LIDAR scanner to generate images and analyze them with a deep learning model and probabilistic graph model to identify lane lines and drivable areas, reducing computational complexity and eliminating the need for high-precision vector maps, allowing for real-time lane change planning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-precision vector maps are used to determine lane positions, then path planning accuracy is improved, but system cost and complexity increase significantly
Solution Approach 1:
The patent extracts the essential function of high-precision vector maps (providing lane position information) and implements it through alternative means: using camera and LIDAR sensors to directly detect and identify lane lines in the real environment, thereby eliminating the need for complex vector map systems while maintaining lane position accuracy
Solution Approach 2:
The patent replaces the mechanical/vector-based map system with a sensor-based perception system. Instead of relying on pre-built vector maps that require complex construction and maintenance, the system uses camera and LIDAR sensors to optically and geometrically detect lane lines in real-time, substituting a simpler sensor-based approach for the complex map-based approach
2Reliability
If high-precision vector maps are constructed and maintained, then accurate lane information is obtained, but construction and maintenance costs increase
Solution Approach 1:
The patent enables the system to self-acquire lane information through sensor-based detection rather than relying on externally constructed and maintained vector maps. The camera and LIDAR sensors automatically capture and process lane line information in real-time, making the system self-sufficient and eliminating the need for expensive map construction and maintenance operations
3Productivity
If vector map information is used for path planning, then computational complexity is reduced, but adaptability to real-time environmental changes decreases
Solution Approach 1:
The patent transitions from a static vector map-based approach to a dynamic sensor-based approach. The camera and LIDAR sensors continuously capture real-time environmental data, allowing the path planning system to dynamically adapt to changing lane configurations and environmental conditions while maintaining efficient processing through real-time image and point cloud analysis
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 solution enables real-time path planning without high-precision vector maps, reducing costs and improving safety by relying solely on sensor inputs from cameras and LIDAR scanners, while being portable and expandable, and capable of embedded platform processing.
Implementation Method 1
The LIDAR scanner is configured to capture a plurality of distance data points of the target road
Implementation Method 2
light detection and ranging (LIDAR) scanner
Data Source
AI summary
The path planning system projects a plurality of path distance point clouds to a path image to generate a path and distance point composition image, and then input the path and distance point composition image to a deep learning model and a probabilistic graphical model to obtain a path segmentation image. The path planning system obtains an adjacent lane point cloud and a main lane point cloud. The path planning system clusters the adjacent lane point cloud and the main lane point cloud, calculates an adjacent lane cluster center and a main lane cluster center, and transforms a LIDAR coordinates of the adjacent lane cluster center and the main lane cluster center to a car coordinate. The path planning system obtains a changing path and a main path by smoothing the car coordinate, and selects the changing path or the main path as a driving path according to obstacle information.