CA-CFAR Log Detector Threshold Computation for Automotive Radar
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing CFAR algorithms in radar systems face challenges in balancing performance, complexity, and computational requirements, particularly in automotive radar applications, due to high dynamic range requirements and computational complexity, leading to suboptimal target detection.
Innovation Solution
The implementation of a cell-average-constant false alarm rate (CA-CFAR) log detector that optimizes log-law detector results with a log CA-CFAR correction factor, reducing dynamic range requirements and improving detection probability by mapping the log detector distribution to the squared-law detector distribution, and using OS-CFAR processes to compute thresholds with reduced computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a square-law detector is used with CA-CFAR algorithm, then the detection performance is improved, but the dynamic range requirement increases by a factor of 2
Solution Approach 1:
The patent transforms the detection parameter from linear scale to logarithmic scale by applying a log detector. This parameter change compresses the dynamic range while preserving detection performance through the development of a corrected CA-CFAR algorithm that accounts for the logarithmic transformation of the noise distribution.
2Measurement precision
If OS-CFAR algorithm is used for multi-target detection, then the detection capability is improved, but the computational complexity increases to O(n²)
Solution Approach 1:
The patent applies segmentation by dividing the Range-Doppler Matrix into separate dimensions (range bins and Doppler bins) and applying 1D OS-CFAR independently to each dimension. This reduces the computational complexity from O((MN)²) for 2D processing to O(MN(M+N)) for separated 1D processing, making real-time multi-target detection feasible.
3Productivity
If windowing with OS-CFAR is used to reduce computational load, then the processing speed is improved, but the detection performance deteriorates
Solution Approach 1:
The patent changes the statistical parameter from using the raw kth largest sample to using the kth largest sample of logarithmically transformed data. This parameter transformation, combined with the corrected threshold computation, maintains detection performance while enabling the use of log detectors that reduce dynamic range requirements.
Data Source
AI summary
A vehicle radar system, apparatus and method use a radar control processing unit generate compressed radar data signals, to apply the compressed radar data signals to a log detector to generate log detector sample values, and to generate a first log cell-average constant false alarm rate (CA-CFAR) threshold from the log detector sample values by computing and adding an average sample value SAVG from the log detector sample values, a probability of false alarm factor α, and a log CA-CFAR correction factor β, where the first log CA-CFAR threshold may be used with a second log CA-CFAR threshold to generate an ordered statistics CA-CFAR threshold for the compressed radar data signals by sorting the first and second log CA-CFAR thresholds by magnitude to form a sorted list of log CA-CFAR thresholds, and then selecting a kth threshold from the sorted list of log CA-CFAR thresholds as the OS-CA-CFAR threshold.