A Deep Learning Approach to Multipath Component Detection in Power Delay Profiles

TL;DR

This paper explores a deep learning framework combining autoencoders and clustering to improve multipath component detection in Power Delay Profiles, demonstrating superior performance over traditional methods.

Contribution

It introduces a novel deep learning framework using transformer autoencoders and DBSCAN clustering for robust multipath detection in PDPs, outperforming existing approaches.

Findings

Transformer autoencoder achieves higher reconstruction accuracy.

The framework improves detection robustness in noisy environments.

Proposed method outperforms traditional techniques in F1 score.

Abstract

Power Delay Profile (PDP) plays a crucial role in wireless communications, providing information on multipath propagation and signal strength variations over time. Accurate detection of peaks within PDP is essential to identify dominant signal paths, which are critical for tasks such as channel estimation, localization, and interference management. Traditional approaches to PDP analysis often struggle with noise, low resolution, and the inherent complexity of wireless environments. In this paper, we evaluate the application of traditional and modern deep learning neural networks to reconstruction-based anomaly detection to detect multipath components within the PDP. To further refine detection and robustness, a framework is proposed that combines autoencoders and Density-Based Spatial Clustering of Applications with Noise (DBSCAN) clustering. To compare the performance of individual models, a relaxed F1 score strategy is defined. The experimental results show that the proposed framework with transformer-based autoencoder shows superior performance both in terms of reconstruction and anomaly detection.