Obstacle Map Cell Smoothing for Uneven Surface Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for creating obstacle environment maps in vehicles are computationally and memory-intensive, making real-time obstacle detection challenging on uneven driving surfaces, particularly in parking and maneuvering areas where the assumption of a flat drivable surface is often violated, leading to incorrect obstacle recognition and faulty driver assistance functions.
Innovation Solution
A method that provides a map of surroundings divided into cells with occupancy probabilities and corrected obstacle heights, using smoothed obstacle heights from an initial environment map to determine ground levels and correct obstacle heights, enabling real-time processing with limited computing and storage capacities, and accounting for varying driving surfaces by treating neighboring cell heights and assigning error values where necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If point cloud measurements are used to calculate segmented floor surface, then measurement precision is improved, but device complexity and computing requirements increase significantly
Solution Approach 1:
The environment map is divided into multiple cells, with each cell independently evaluated for ground level determination. This segmentation allows parallel processing of individual cells while maintaining overall accuracy, reducing the computational burden compared to processing the entire point cloud simultaneously.
Solution Approach 2:
The invention extracts only the necessary ground level information from sensor measurements by using a simplified cell-based evaluation approach. Instead of processing all point cloud data, it extracts elevation values at specific cell locations and uses threshold-based filtering to determine ground levels, significantly reducing data processing requirements.
2Device complexity
If simple threshold calculation is used to distinguish road surface from obstacle, then device complexity is reduced, but measurement precision deteriorates on uneven surfaces
Solution Approach 1:
The invention applies different processing approaches to different cells based on local characteristics. Each cell's ground level is determined independently using elevation thresholding, and cells are classified as ground or obstacle based on their specific local elevation values rather than applying a global threshold to all measurements.
Solution Approach 2:
The invention changes the parameter used for obstacle detection from fixed height thresholds to relative elevation differences within each cell. By comparing elevation values against cell-specific ground levels and applying height difference thresholds, the system adapts to local terrain variations while maintaining simple processing logic.
3Measurement precision
If extensive measurement data is collected for each cell to determine ground level, then measurement precision is improved, but loss of time increases due to data processing requirements
Solution Approach 1:
The invention uses a sufficient but not excessive amount of measurement data for ground level determination. Instead of requiring extensive point cloud data for each cell, it uses a limited set of elevation measurements at specific locations within each cell, combined with threshold-based filtering, to achieve adequate ground level accuracy in real-time.
Solution Approach 2:
The invention uses simple, computationally inexpensive calculations for ground level determination that can be quickly discarded and recalculated as new sensor data arrives. The cell-based threshold evaluation method requires minimal processing power, allowing rapid updates as the vehicle moves and new measurements become available.
Data Source
Figure 1
Figure 2~3b
Figure 4a~4b
AI summary
A method is disclosed for making available a map of obstacles in the surroundings for a vehicle, specifically a map of the surroundings which is divided into cells, each of which is assigned an occupation probability and a corrected obstacle altitude, comprising: making available an initial map of the surroundings, specifically a map of the surroundings which is divided into cells, each of which is assigned an occupation probability and an obstacle altitude; determining a ground altitude for each cell of the initial map of the surroundings on the basis of the altitudes of the obstacles in the initial map of the surroundings, wherein the ground altitude is determined by smoothing the altitudes of obstacles in the initial map of the surroundings; determining the corrected altitude of the obstacles for each cell of the map of obstacles in the surroundings on the basis of the ground altitude determined for the corresponding cell of the initial map of the surroundings and the altitude of the obstacles assigned to the corresponding cell of the initial map of the surroundings; determining the respective occupation probability on the basis of the occupancy probability assigned to the corresponding cell of the initial map of the surroundings; making available the map of obstacles in the surroundings.