Radar Device Noise Variance Reduction via CIR Tap Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Ultra-wideband (UWB) radar systems face challenges in detecting moving radar targets due to unidentified noise sources that introduce variance changes in the received signals, making these changes indistinguishable from statistical changes caused by moving targets.
Innovation Solution
A method for operating a radar device that involves performing pulse radar measurements, determining specified training and target sets of measured channel impulse response (CIR) taps, calculating auto-covariance and cross-covariance matrices, and using linear prediction to subtract predicted values from target CIR taps, thereby reducing noise variance and enhancing target detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pulse radar measurements are performed to detect radar targets, then target detection capability is improved, but noise variance increases making targets undetectable
Solution Approach 1:
The patent applies preliminary action by performing training measurements before actual target detection. During the training phase, the system collects CIR taps and computes auto-covariance and cross-covariance matrices to establish statistical models of the channel response. This preliminary statistical characterization enables subsequent noise compensation without interfering with actual target detection operations.
Solution Approach 2:
The patent introduces an intermediary approach by using predicted CIR tap values as a mediator between the raw noisy measurements and the final target detection. The linear prediction model generates estimated CIR tap values that represent the expected channel response, which are then subtracted from the actual measurements to produce compensated signals for target detection.
2Reliability
If noise compensation is performed using training sets, then noise variance is reduced, but measurement complexity increases
Solution Approach 1:
The patent applies parameter changes by transforming the statistical parameters of the CIR taps through computation of auto-covariance and cross-covariance matrices. By changing from raw signal values to covariance-based statistical parameters, the system captures the essential noise characteristics while reducing the dimensionality and complexity of the data representation.
3Object-affected harmful factors
If linear prediction is applied to subtract predicted values from target CIR taps, then noise is suppressed, but computational requirements increase
Solution Approach 1:
The patent replaces direct signal processing with statistical modeling approaches. Instead of using complex adaptive filtering or iterative optimization methods, the system substitutes a linear prediction model based on covariance statistics. This substitution simplifies the computational burden while maintaining effective noise suppression through the mathematical properties of covariance matrices.
Data Source
AI summary
A radar device is configured to perform a method that includes performing pulse radar measurements and determining measured CIR-taps. The method also includes a specified training set of measured CIR-taps; determining a specified target set of measured CIR-taps in a region of interest with expected radar targets; determining an auto-covariance matrix of CIR-taps of the specified training set; determining a cross-covariance vector between CIR-taps of the specified training set and a specified target CIR tap of the specified target set; and predicting a value for the specified target CIR-tap. The predicted value is subtracted from the corresponding value of the specified target CIR tap.