Probability-Graph Obstacle Distribution for Realistic Road Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing obstacle simulation methods in high-precision maps fail to accurately represent the distribution and number of obstacles in various scenarios due to limited rule-based arrangements, resulting in a low coverage of real-world conditions.
Innovation Solution
A method and device utilizing a probability graph to simulate obstacle distribution by acquiring point clouds, constructing a labeling data set, dividing it into grids, calculating occurrence probabilities, and selecting obstacles based on these probabilities to achieve a more diverse and realistic simulation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If obstacle arrangement based on rules is used in high-precision map simulation, then the simulation process is simple and fast, but the coverage of real-world scenarios is low and the distribution accuracy is poor
Solution Approach 1:
The patent copies real obstacle distribution patterns from actual point cloud data to generate simulated obstacle arrangements. Instead of using abstract rules, the system directly replicates the spatial distribution characteristics observed in real-world scenarios, preserving the statistical properties and spatial relationships of actual obstacles while enabling scalable simulation.
Solution Approach 2:
The patent transforms the simulation approach by changing from rule-based parameters to data-driven parameters. By extracting statistical parameters (density, distribution patterns, spatial relationships) from real point cloud data, the system dynamically adjusts simulation parameters to match real-world conditions, improving distribution accuracy while maintaining simulation efficiency.
2Device complexity
If high-precision map with main roads only is used, then the map structure is simple and clear, but the coverage of side roads and branch roads is missing leading to low scenario diversity
Solution Approach 1:
The patent makes the simulation system universal by using point cloud data from diverse road scenarios (main roads, side roads, branch roads, intersections) to train the obstacle distribution model. The extracted statistical patterns from this diverse data enable the system to generalize and adapt to various road types and scenarios, making the simulation applicable to multiple contexts beyond simple main roads.
3Ease of manufacture
If rule-based obstacle arrangement is used, then the arrangement method is easy to implement, but it cannot present all possible cases in real scenarios resulting in low coverage
Solution Approach 1:
The patent enables the system to self-adapt to different scenarios by automatically extracting obstacle distribution patterns from real point cloud data without requiring manual rule creation for each scenario. The system self-learns the statistical characteristics and spatial relationships from the data, automatically adjusting to present diverse real-world cases including rare scenarios that would be difficult to anticipate with predefined rules.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
An obstacle distribution simulation method, device and terminal based on a probability graph are provided. The method includes: acquiring a plurality of point clouds of a plurality of frames; acquiring real labeling data of an acquisition vehicle at vehicle labeled positions, and acquiring data of a simulation position of the acquisition vehicle; determining the number of obstacles to be simulated at a position to be simulated; extracting real labeling data of the obstacles, and constructing a labeling data set; dividing the labeling data set into a plurality of grids and calculating occurrence probabilities of the plurality of obstacles; selecting the determined number of obstacles to be simulated according to the occurrence probabilities; and acquiring a position distribution of the selected obstacles to be simulated for the position to be simulated based on the real labeling data of the selected obstacles to be simulated.