Radar Object Boundary Recognition via Range-Doppler Map Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Radar sensor systems face high computational burden in determining object boundaries due to the complexity of calculating observation angles and the difficulty in finding sufficient detections, especially when stationary objects are passed by the host vehicle, leading to inefficient recognition of stationary objects in traffic spaces.
Innovation Solution
The method involves dividing the Range-Doppler map into evaluation regions separated by parallel lines, allowing for the selection of detections with extremal values along the range and Doppler axes, reducing computational effort and ensuring detections are distributed across the axes, thereby enhancing boundary recognition reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If calculations are carried out for all detections to determine object boundaries, then measurement precision is improved, but computational effort increases significantly
Solution Approach 1:
The Range-Doppler map is divided into multiple evaluation regions separated by separation lines extending parallel to the range axis and Doppler axis. This segmentation allows the system to process detections in distributed regions rather than handling all detections globally, reducing computational complexity while maintaining boundary recognition accuracy through localized extremal value calculations in each region
Solution Approach 2:
The patent extracts only the essential information needed for boundary determination by identifying selected detections with extremal values (maximum or minimum range or Doppler shift) in each evaluation region. This extraction approach discards redundant detection data while preserving the critical boundary-defining detections, thereby reducing computational effort without sacrificing measurement precision
2Productivity
If a threshold is used to select only the strongest detections, then computational burden is reduced, but sufficient detections cannot be found in all range regions
Solution Approach 1:
By dividing the Range-Doppler map into multiple evaluation regions with separation lines, the patent ensures that detections are systematically distributed across different range regions. This segmentation prevents the threshold-based selection from missing detections in any particular region, as each region independently identifies its own selected detections with extremal values, guaranteeing coverage reliability while maintaining computational efficiency
Solution Approach 2:
The patent applies local quality by determining selected detections based on extremal values within each local evaluation region rather than applying a global threshold. This allows each region to adaptively identify its own relevant detections based on local characteristics, ensuring that even weaker detections in specific regions are captured if they represent boundary information, thus improving detection coverage reliability
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach significantly reduces computational effort and enhances the reliability of object boundary recognition by focusing calculations on a limited number of selected detections, ensuring reliable identification of object boundaries with reduced ghost detections and improved detection efficiency across the range and Doppler axes.
Implementation Method 1
Radar sensor systems often use the known Doppler effect to gather information relating to objects moving relative to the host vehicle. The Doppler effect or Doppler shift is a change in frequency observed when a wave source moves relative to the receiver.
Data Source
AI summary
A method includes identifying, from a reflected radar signal, a plurality of single detections corresponding to object surface spots detected by the radar sensor system, wherein the positions of the single detections in a Range-Doppler-map are deter-mined, wherein at least a region of the Range-Doppler map is divided into a plurality of adjacent evaluation regions separated by separation lines, wherein the separation lines extend parallel to one of the range axis and the Doppler axis. For each evaluation region, at least one selected detection is determined which has, among the detections present in the respective evaluation region, an extremal value with respect to the other axis of the range axis and the Doppler axis, and a boundary of the at least one object is determined based on the selected detections.