Self-optimization of uplink power control in LTE
- Autores: Ana Belén Vallejo Mora
- Directores de la Tesis: Matías Toril Genovés (dir. tes.), Salvador Luna Ramírez (dir. tes.)
- Lectura: En la Universidad de Málaga ( España ) en 2024
- Idioma: inglés
- Tribunal Calificador de la Tesis: Fernando Ruiz Vega (presid.), María Luisa Marí Altozano (secret.), Liam Fallon (voc.)
- Programa de doctorado: Programa de Doctorado en Ingeniería de Telecomunicación por la Universidad de Málaga
- Materias:
- Enlaces
- Tesis en acceso abierto en: RIUMA
- Resumen
In today's heterogeneous communication networks, where different technologies such as Global System for Mobile communication (GSM)/Universal Mobile Telecommunications System (UMTS)/Long Term Evolution (LTE)/ New Radio (NR) coexist, the ability to self-manage and automate the maximum number of processes is essential. Only in this way, a better network performance can be achieved while reducing operational costs (OPEX, Operating Expenses). This was the reason why the concept of Self-Organizing Networks (SON) emerged, which includes the functionalities of self-configuration, self-optimization and self-healing.
Self-optimization consists of the automatic adjustment of network parameters, whose settings are dynamically updated whenever there is a change in user traffic and mobility trends, or new equipment or functionality is introduced into the network. Examples of optimization use cases are the updating of neighbor cell list, or capacity and coverage optimization through procedures such as Remote Electrical Tilt (RET) or UpLink Power Control (ULPC).
The first tests carried out in live networks showed that in LTE the UpLink (UL) is much more restrictive than the DownLink (DL). This is due to the limited power of user terminals and the interference that exists in this link, coming from a large number of User Equipments (UE) in neighbor cells. To alleviate this effect, a possible strategy would be to adjust the UL nominal power (P0), i.e., the transmit power of each UE.
In this thesis, a preliminary analysis shows the impact of P0 changes in ULPC on network performance indicators. The sensitivity analysis of said parameter is carried out in a commercial LTE network, and the acquired knowledge is the starting point for designing the optimization methods presented in this work.
Once the impact of P0 parameter is understood, a model is devised to accurately determine the neighbor cells that cause the highest UL interference on the cell under study. This model is especially important in situations with degraded performance due to interference problems. The inter-cell interference coupling is based on the correlation between fluctuations of received UL interference in the cell under study and fluctuations of load in neighbor cells. These coupling factors are calculated by combining two methods, Multiple Linear Regression (MLR) and Principal Component Analysis (PCA), and using data from cell traces of a real LTE network. After calculating the coupling factors between cells, a new performance indicator is designed and implemented, which represents the contribution of each neighbor cell to the received UL interference in the cell under study. This contribution is measured as a percentage of its total UL interference.
The initial field trial carried out during this thesis showed that the UL performance is severely degraded with group mobility in subway cells. To relief this effect, a self-optimization algorithm is proposed to adjust P0 parameter. This algorithm is also validated on a real LTE network.
Finally, it is well known that operators demand the need to have an adequate tool to estimate the impact of P0 changes on network performance, which is not available today in commercial products. Therefore, one of the objectives of this thesis was to have an analytical model capable of predicting the effect of P0 changes on certain relevant performance indicator (UL interference and UL SINR), both in the modified cell and in its neighbor cells. The study is focused on two low-level indicators, UL interference and UL SINR, since their values affect many high-level indicators, which are relevant to operators.
All the methods developed in this thesis are conceived to be integrated in a centralized commercial optimization tool in the network management system. Both for the analysis of the problem and for the validation of the designed optimization solutions, performance statistics and connection traces taken from real networks are considered. In their development, the constraints imposed by the operator and the vendor are taken into account, paying special attention to the computational efficiency of solution methods. As a proof of concept, the solutions are initially tested and validated in a system-level simulator. The scenario considered in the simulator is the same LTE scenario (number of cells, cell locations and orientations and configured P0 per cell) as in the real network, where a field trial is carried out later.
