Radar Object Detection Using Concatenated Point and Map Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for object detection in automotive perception systems using radar data are inefficient in predicting the location and motion of objects, particularly due to the high dimensionality of radar point clouds and the need for improved data representation for effective processing.
Innovation Solution
A computer-implemented method that combines radar detection data and radar map data by converting them into suitable representations, concatenating these representations, and preprocessing the data to remove vehicle ego-location effects, allowing for efficient object detection using neural networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If radar detection data and map data are combined into concatenated data for object detection, then detection accuracy is improved, but data processing complexity increases
Solution Approach 1:
The patent segments the processing of radar detection data and map data by converting them into separate standardized representations (point representation and map representation) before concatenation. This segmentation allows each data type to be processed independently into a unified format, reducing overall processing complexity while maintaining detection accuracy.
Solution Approach 2:
The patent applies parameter changes by transforming radar detection data into point representations with specific dimensions and map data into map representations with matching dimensions. This standardization of data parameters enables efficient concatenation and processing, improving detection accuracy without proportionally increasing processing complexity.
2Stability of the object's composition
If radar data is preprocessed to remove vehicle ego-location effects, then detection consistency is improved, but processing time increases
Solution Approach 1:
The patent applies preliminary action by performing ego-location compensation during the data preprocessing stage, before object detection is executed. This advance processing removes vehicle motion effects from the data, ensuring detection consistency across multiple frames without adding time pressure to the critical detection path.
Solution Approach 2:
The patent introduces an intermediary preprocessing step that transforms raw radar data into ego-location compensated data. This intermediary representation serves as a stable foundation for subsequent detection operations, improving consistency while allowing time-consuming operations to be performed separately from the main detection loop.
3Loss of information
If high-dimensional radar point cloud data is used for object detection, then detection completeness is improved, but processing efficiency decreases
Solution Approach 1:
The patent applies dimensionality change by converting high-dimensional radar point cloud data into a standardized point representation with predetermined dimensions. This transformation projects the high-dimensional data into a manageable format that retains essential detection information while enabling efficient processing through consistent dimensional structure.
Solution Approach 2:
The patent segments the high-dimensional radar data processing by dividing it into distinct conversion steps: raw radar data → point representation → concatenated representation. This segmentation allows the system to handle high-dimensional data in manageable portions, maintaining detection completeness while improving processing efficiency through structured intermediate representations.
Data Source
AI summary
A computer implemented method for object detection includes: determining a grid, the grid comprising a plurality of grid cells; determining, for a plurality of time steps, for each grid cell, a plurality of respective radar detection data, each radar detection data indicating a plurality of radar properties; determining, for each time step, a respective radar map indicating a pre-determined radar map property in each grid cell; converting the respective radar detection data of the plurality of grid cells for the plurality of time steps to a point representation of pre-determined first dimensions; converting the radar maps for the plurality of time steps to a map representation of pre-determined second dimensions, wherein the pre-determined first dimensions and the pre-determined second dimensions are at least partially identical; concatenating the point representation and the map representation to obtain concatenated data; and carrying out object detection based on the concatenated data.