Dynamic Spatial Scenario Analysis Using Angular Sector Time Curves
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complex task of reliably recognizing dynamic spatial scenarios, such as traffic situations, in sensor data for modern vehicles is challenging due to the need for precise data preparation to prevent information loss during processing by artificial neural networks.
Innovation Solution
A method and system for preparing sensor data by generating a display of the time curve of an angular sector covered by another object from the perspective of an ego object, which is then processed by an artificial neural network to improve data processing efficiency and reduce information loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If sensor data is prepared using traditional grid maps to represent driving situations, then the data can be processed by artificial neural networks, but information loss occurs during the preparation process
Solution Approach 1:
The patent transforms the traditional 2D grid map representation into a 3D spatiotemporal representation by adding the time dimension. This is achieved by stacking multiple grid maps representing different time points to form a volumetric data structure, allowing the neural network to capture temporal dynamics while preserving spatial information, thereby reducing information loss during data preparation
Solution Approach 2:
The patent applies preprocessing operations to the sensor data before feeding it to the neural network. This includes generating bird's-eye view representations from camera images, calculating angular sectors covered by objects, and constructing the spatiotemporal grid map structure in advance. These preliminary actions organize the data in a format that preserves critical information while making it suitable for neural network processing
2Reliability
If more sensor data is collected to improve scenario recognition accuracy, then identification reliability increases, but data processing requirements and complexity increase
Solution Approach 1:
The patent extracts and focuses on the most critical features from the sensor data, such as the angular sector coverage by objects, their positions in bird's-eye view, and temporal changes in these parameters. By selecting only the essential information needed for scenario recognition, the system maintains high reliability while reducing the overall data volume that requires processing
Solution Approach 2:
The patent segments the complex sensor data into distinct components: spatial information from bird's-eye views, angular sector measurements, temporal sequences, and object characteristics. This segmentation allows each component to be processed independently and efficiently, reducing the computational burden while maintaining the integrity needed for reliable scenario recognition
Data Source
AI summary
The invention relates to a method and a system for preparing data on dynamic spatial scenarios, to a computer-supported method, to a system for training artificial neural networks, to a computer-supported method, and to a system for analyzing sensor data. A display of a time curve of an angular sector covered by another object from the perspective of an ego object is generated. The time curve is ascertained from sensor data, and the sensor data characterizes a dynamic spatial scenario with respect to the ego object and at least one other object.