Martin Willame,Laurent Storrer,Hasan Can Yildirim,François Horlin,Jérôme Louveaux
Abstract:Recent advancements in Wi-Fi sensing have sparked interest in exploiting OFDM modulated communication signals for target detection and tracking. In this study, we address the angle-based localization of multiple targets using a multistatic OFDM radar. While the maximum likelihood approach optimally merges data from each radar pair comprised by the system, it entails a complex multi-dimensional search process. Leveraging pre-estimation of the targets' parameters obtained via the MUSIC algorithm, our method decouples this multi-dimensional search into a single two-dimensional estimator per target. The proposed alternating summation method allows the computation of a combined likelihood map aggregating contributions from each radar pair, enabling target detection via peak selection. Besides reducing computational complexity, the method effectively captures target interactions and accommodates varying radar pair localization abilities. Also, it requires transmitting only the estimated channel covariance matrices of each radar pair to the central processor. Numerical simulations demonstrate superior performance over existing approaches.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the angle - based multi - target localization problem in multi - static OFDM radar systems. Specifically, the researchers hope to develop an effective signal - level fusion method to jointly estimate the angle of departure (AoD) and the angle of arrival (AoA) of multiple targets, thereby improving the accuracy of target detection and localization.
### Problem Background
In recent years, the development of Wi - Fi sensing technology has made it possible to use communication signals modulated by orthogonal frequency - division multiplexing (OFDM) for target detection and tracking. In multi - static radar systems, multiple radar pairs (RP) work together, and the data of each RP is fused by the central processing unit to enhance the sensing ability of the system. However, the traditional maximum - likelihood (ML) method faces a complex high - dimensional search problem when dealing with multi - target localization, resulting in high computational complexity.
### Research Motivation
To overcome this challenge, the author proposes a maximum - likelihood fusion method based on alternating summation. This method uses the MUSIC algorithm to pre - estimate the parameters of the target, decomposes the originally complex 2K - dimensional ML estimation problem into K two - dimensional search problems, thereby significantly reducing the computational complexity. In addition, this method can effectively capture the interaction between different targets and consider the differences in the localization capabilities of each radar pair under different geometric conditions.
### Main Contributions
1. **Signal - level Fusion Method**: A signal - level fusion method based on the ML framework is proposed for the joint AoD/AoA localization of K targets in multi - static OFDM radar systems. By using the MUSIC algorithm to pre - estimate the target parameters, the alternating summation method effectively decomposes the 2K - dimensional ML estimation problem into K two - dimensional search problems for each target.
2. **Target Interaction and Localization Capability**: Compared with previous works, the new method can not only capture the interaction between targets, but also consider the differences in the localization capabilities of each radar pair under different geometric conditions without increasing the search dimension.
3. **Performance Verification**: Numerical simulation results show that this method has higher localization accuracy and hit rate compared with existing methods when the signal - to - noise ratio (SNR) changes.
### Method Overview
- **System Model**: Describes the configuration of the multi - static OFDM radar system, including multiple transmit - receive pairs (RP), and defines the channel model.
- **Maximum - likelihood Fusion**: Derives the ML combination rule to determine the target position by summing the local log - likelihood functions. By using the MUSIC algorithm to pre - estimate the target parameters, the high - dimensional search problem is decomposed into multiple low - dimensional problems.
- **Alternating Summation Method**: Fuses the local log - likelihood functions of multiple radar pairs into a comprehensive log - likelihood map by the alternating summation method, and determines the target position by peak selection.
### Conclusion
The alternating summation method proposed in the paper achieves efficient multi - target localization in multi - static OFDM radar systems, significantly reduces the computational complexity, and improves the localization accuracy. Future research can further combine accurate data association techniques to improve the pre - estimated results obtained from different radar pairs.
### Formula Summary
- **Channel Vector**:
\[
h_{p,q}(\Psi_p)=A_p(\Psi_p)\alpha_{p,q}
\]
where \(\Psi_p = (\Phi_p,\Theta_p)\) is the set of target angles of the \(p\) - th radar pair, \(A_p(\Psi_p)\) is the joint AoD/AoA steering matrix, and \(\alpha_{p,q}\) is the channel coefficient vector.
- **Local Log - likelihood Function**:
\[
L_p(\gamma)=-\frac{1}{2\sigma_p^2}\sum_{q = 0}^{Q_p-1}\|eh_{p,q}-A_p(\Psi_p)\alpha_{p,q}\|^2
\]
- **ML - estimated Channel Coefficient**:
\[
\hat{\alpha}_{p,q}(\Psi_p)=A_p^+(\Psi_p)eh_{p,q}
\]
- **Local per - target Log - likelihood Function**:
\[
L_p^{(k)}(x_k,y_k)=\frac{Q_p}{2\sigma_