Implementation of platform for long‐term evolution cell perspective resource utilization analysis

Abstract

As wireless communication continues to develop in limited frequency resource environments, it is becoming important to identify the state of spectrum utilization and predict the amount needed in future. It is essential to collect reliable information for data analysis. This paper introduces a platform that enables the gathering of the scheduling information of a long-term evolution (LTE) cellular system without connecting to the network. A typical LTE terminal can confirm its assigned resource information using the configuration parameters delivered from a network. However, our platform receives and captures only the LTE signal over the air and then enables the estimation of the data relevant to scheduling for all terminals within an LTE cell.After extracting the control channel signal without loss from all LTE subframes, it detects valid downlink control information using the proposed algorithm, which is based on the error vector magnitude of depatterned symbols. We verify the reliability of the developed platform by comparing it with real data from mobile phones and service operators. The average difference in resource block utilization is only 0.28%.

Bi‐LSTM model with time distribution for bandwidth prediction in mobile networks

“…As the proposed Bi-LSTM model with TD model employs RSRQ, RSRP, and RSSI as inputs, the number of resource blocks is implicit in the proposed model and used for bandwidth prediction. However, the number of resource blocks allotted to a UE may vary depending on the number of active users in a cell, which affects the prediction of the UE available bandwidth [27]. Nevertheless, the proposed bandwidth prediction model uses RSRQ without considering the number of users owing to the lack of this information in the dataset.…”