Pilot Allocation and Assignment Optimization in User-Centric Distributed Massive MIMO Networks

Distributed massive multiple-input multiple-output (MIMO) networks, also known as cell-free, are promising solution to increase efficiency in beyond 5G systems. Pilot-based uplink (UL) channel estimation directly influences transmission efficiency, as it is used to mitigate interference and noise from user equipment (UEs). In this context, this work uses genetic algorithm (GA) as a tool to optimize pilot allocation and assignment and maximize spectral efficiency (SE). First, we define the optimal amount of samples allocated to channel estimation that balances accuracy and overhead. Generally, this lead to fewer pilots than UEs. Therefore, the pilot assignment is also optimized to decrease interference between UEs reusing the same pilot. The results show that the optimal number of pilots presents a similar behavior when the number of UEs increases. The average SE is improved when GA is used to optimize pilot assignment compared with the baseline solution