Vehicle Surround Monitoring Using Occupancy Grid Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle radar systems face challenges in accurately monitoring the surrounding environment, particularly in distinguishing stationary objects and determining the type of parking space, which affects the performance of rear cross traffic alert systems.
Innovation Solution
An apparatus and method utilizing a sensor unit and controller to detect stationary objects, map them to a grid map, calculate occupancy probability parameters, apply clustering algorithms to identify continuous structures, and adjust the rear cross traffic alert system based on the type of parking space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If general radar detection is used to monitor surrounding environment, then detection coverage is provided, but accuracy in distinguishing stationary objects and determining parking types is insufficient
Solution Approach 1:
The patent divides the detection space into a grid map with multiple grids, where each grid can independently represent the presence or absence of stationary objects. This segmentation allows the system to precisely locate and classify objects in specific regions without requiring complex global analysis, thereby improving detection accuracy while maintaining manageable system complexity.
Solution Approach 2:
The patent introduces a probabilistic dimension by calculating occupancy probability parameters for each grid based on multiple radar detections. Instead of binary detection, the system uses probability values to represent the likelihood of object presence, adding a dimensional layer of information that significantly improves stationary object distinction and parking type determination accuracy.
2Measurement precision
If occupancy probability calculation is applied to each grid, then stationary object detection accuracy is improved, but computational load increases
Solution Approach 1:
The patent applies occupancy probability calculation only to specific grids where stationary objects are detected, rather than uniformly processing the entire grid map. This localized approach concentrates computational resources on relevant regions, improving object location accuracy while reducing overall computational load by avoiding unnecessary processing of empty spaces.
Solution Approach 2:
The system performs clustering algorithms and detailed analysis only on grids with high occupancy probability values, rather than processing all grids equally. This partial action approach focuses computational power on the most likely object locations, achieving high accuracy in critical areas while minimizing waste of computational resources on low-probability regions.
3Measurement precision
If clustering algorithm is used to identify continuous structures, then parking type determination accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary grouping of grids based on occupancy probability thresholds before applying clustering algorithms. By pre-identifying and grouping high-probability grids, the system reduces the input data size for clustering operations, thereby improving parking type identification accuracy while significantly reducing the processing time required for clustering analysis.
Solution Approach 2:
The system extracts and focuses on edge grids from clustered groups to determine parking types, rather than analyzing all grids in the cluster. This extraction of critical information (edge grids that define the boundaries and characteristics of parked vehicles) allows accurate parking type identification while minimizing processing time by avoiding redundant analysis of internal cluster grids.
Data Source
AI summary
Disclosed herein are an apparatus and method for monitoring a surrounding environment of a vehicle, the apparatus including a sensor unit including a plurality of detection sensors for detecting an object outside a vehicle according to a frame at a predefined period, and a controller configured to extract a stationary object from among the outside objects detected by the sensor unit, to map the extracted stationary object to a grid map, to calculate an occupancy probability parameter, indicative of a probability that the stationary object will be located on a grid of the grid map, from the result of mapping, and to monitor the surrounding environment of the vehicle by specifying a grid on which the stationary object is located in the grid map, based on the occupancy probability parameter.