Radar Target Separation Using Interleaved Model Orders
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional radar target separation methods, particularly those based on beamformers and maximum likelihood approaches, face inefficiencies due to high computational requirements when optimizing multiple model orders, leading to increased processing effort.
Innovation Solution
A grid-based method is employed to calculate only the highest model order, with lower orders derived as interim results, reducing computational load by interleaving model orders and utilizing pre-calculated values stored in a look-up table, and employing a simplified cost function for efficient separation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If separate optimization is performed for each model order to achieve accurate target separation, then measurement precision is improved, but computational complexity increases significantly
Solution Approach 1:
The patent pre-calculates and stores the steering matrix and correlation matrix in a look-up table before actual target separation is needed. These pre-computed values are reused across multiple model orders, eliminating the need to recompute them for each optimization iteration and significantly reducing the computational burden while maintaining separation accuracy.
Solution Approach 2:
The patent computes the cost function and its gradient in a unified manner that works across multiple model orders simultaneously. By formulating the optimization problem to handle model order N and all lower model orders in a single computational framework, the solution achieves universal applicability without requiring separate optimization routines for each model order.
2Reliability
If multiple model orders are calculated separately to determine the correct number of radar targets, then reliability is improved, but processing time increases
Solution Approach 1:
The steering matrix and correlation matrix are pre-calculated and stored in a look-up table before model order selection is performed. This preliminary computation allows the cost function evaluation to proceed efficiently for multiple model orders without repeating expensive matrix operations, thereby reducing processing time while maintaining reliable model order detection.
Solution Approach 2:
The patent combines the computation of cost functions for multiple model orders into a single unified optimization process. By merging the evaluations and using the same pre-computed matrices across different model orders, the method reduces redundant calculations and achieves faster processing while reliably identifying the correct model order through information criteria comparison.
3Measurement precision
If high-resolution model-based methods are used to separate radar targets, then measurement precision is improved, but computational requirements increase
Solution Approach 1:
The patent pre-computes the steering matrix A(θ) and correlation matrix R and stores them in a look-up table before actual target separation is performed. These pre-calculated matrices are reused across multiple model orders and optimization iterations, eliminating redundant computations and significantly reducing the energy consumption associated with high-resolution model-based separation while maintaining separation precision.
Solution Approach 2:
The patent changes the computational approach by reformulating the cost function to utilize pre-computed matrices and by computing gradients in a manner that leverages these pre-calculated values. This parameter change in the computational methodology reduces the energy requirements for evaluating multiple model orders while maintaining the high-resolution separation capability through accurate maximum likelihood estimation.
Data Source
AI summary
A method for the model-based, high-resolution separation of radar targets for a radar sensor, in which the radar sensor initially generates radar data by capturing radar targets by sampling a field of view of the radar sensor. Various model orders are calculated for the number of radar targets with the aid of a grid-based method and are interleaved in one another. A highest model order is specified, and only the highest model order is calculated, so that the lower model calculations are implicitly produced from the calculation of the highest model order.