# Deep Learning Based Pilot Design for Multi-user Distributed Massive MIMO Systems

## TL;DR

This paper introduces a deep learning-based pilot design method for multi-user distributed massive MIMO systems that optimizes pilot power allocation to minimize channel estimation error, outperforming existing methods.

## Contribution

It proposes a novel deep neural network approach to optimize pilot power allocation in distributed massive MIMO, reducing mean square error in channel estimation.

## Key findings

- Achieves lower sum MSE than traditional methods
- Utilizes unsupervised learning for training the DNN
- Demonstrates low complexity and improved performance

## Abstract

This letter proposes a deep learning based pilot design scheme to minimize the sum mean square error (MSE) of channel estimation for multi-user distributed massive multiple-input multiple-output (MIMO) systems. The pilot signal of each user is expressed as a weighted superposition of orthonormal pilot sequence basis, where the power assigned to each pilot sequence is the corresponding weight. A multi-layer fully connected deep neural network (DNN) is designed to optimize the power allocated to each pilot sequence to minimize the sum MSE, which takes the channel large-scale fading coefficients as input and outputs the pilot power allocation vector. The loss function of the DNN is defined as the sum MSE, and we leverage the unsupervised learning strategy to train the DNN. Simulation results show that the proposed scheme achieves better sum MSE performance than other methods with low complexity.