Adaptive Obstacle Map Grid for Vehicle Collision Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Classic obstacle maps used in motor vehicle control require fine rasterization, leading to high computational complexity and making real-time prediction of future obstacle maps a challenging problem, especially when dealing with large grid sizes.
Innovation Solution
The method involves creating an obstacle map by subdividing the area around a motor vehicle into cells with dimensions and orientations adapted to the road course, using non-square shapes like rectangles or circular segments, and varying cell sizes based on distance from the vehicle, reducing redundant information and processing load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the obstacle map is finely rasterized to improve control accuracy, then measurement precision is improved, but device complexity increases significantly
Solution Approach 1:
The area surrounding the motor vehicle is segmented into cells with dimensions adapted to the road course, creating a hierarchical structure where the road center line divides the area into left and right sides, and cells are arranged along these divisions. This segmentation reduces the total number of grid points while maintaining necessary spatial resolution for collision avoidance.
Solution Approach 2:
The cell dimensions vary locally based on position relative to the road course. Cells closer to the motor vehicle have smaller dimensions for higher resolution where collision risk is greatest, while cells farther away have larger dimensions. This local adaptation of quality reduces overall computational complexity while preserving accuracy where most needed.
2Device complexity
If the grid size is increased to reduce processing load, then device complexity is reduced, but measurement precision deteriorates
Solution Approach 1:
The obstacle map implementation uses varying cell dimensions based on distance from the motor vehicle. Near-field cells have smaller sizes for high-resolution obstacle detection, while far-field cells have larger sizes. This local quality variation maintains detection accuracy in critical zones while reducing processing load in less critical distant areas.
3Ease of manufacture
If regular square cells are used for simplicity, then ease of manufacture is improved, but measurement precision deteriorates due to projection errors
Solution Approach 1:
The patent abandons symmetric square cells in favor of asymmetric rectangular cells whose dimensions and orientations are adapted to the road course. The cell main axes are oriented parallel to the road direction, creating an asymmetric grid that eliminates projection errors when mapping sensor data to the obstacle map, thereby improving distance estimation accuracy.
Solution Approach 2:
The cell orientations are dynamically adapted to the road course rather than being fixed in a regular grid pattern. The main axes of cells are oriented as a function of the road direction at each location, allowing the grid structure to dynamically conform to the road geometry and improve measurement precision.
Data Source
AI summary
A method for creating an obstacle map of an area surrounding a motor vehicle on a road includes preparing a rasterized obstacle map, scanning an area surrounding the motor vehicle, dividing the surrounding into cells, recording those of the cells that are occupied by an obstacle, and identifying grid points of the obstacle map that correspond to occupied cells in the surrounding area. The dimensions of the cells are determined as a function of the position thereof relative to the course of the road.