Raquel Barco

Automated Diagnosis for UMTS Networks Using Bayesian Network Approach

This paper presents an automated diagnosis in troubleshooting (TS) for Universal Mobile Telecommunications System (UMTS) networks using a Bayesian network (BN) approach. An automated diagnosis model is first described using the Naive Bayesian Classifier. To increase the performance of the diagnosis model, the entropy minimization discretization (EMD) method is incorporated into the model to select optimal segments for the discretization of the input symptoms. In the first phase, the diagnosis model is constructed using a dynamic simulator. The simulator TS platform allows generation of a large amount of data required to study the relations between faults and symptoms. In the second phase, the diagnosis model is adapted to a real UMTS network using counters and key performance indicators (KPIs) recovered from an Operations and Maintenance Center (OMC). Results for the automated diagnosis using both network simulator and real UMTS network measurements illustrate the efficiency of the proposed TS approach and its importance to mobile network operators.

A unified framework for self-healing in wireless networks

Future wireless networks will be Heterogeneous Networks (HetNet), as they will comprise multiple Radio Access Technologies (RAT) and network architectures. The complexity of managing and operating such networks is forcing operators to find new strategies if they want to remain competitive. In this context, Self-Organizing Networks (SON) have arisen as one of the key technologies for enhancing operation and minimizing costs in future networks. SON functionalities can be classified as self-configuration, self-optimization, and self-healing. Although the first two categories have received considerable attention in the last years, the studies devoted to self-healing have been scarce. In addition, existing references use different terminology and concepts and they focus only on partial aspects of self-healing for a particular RAT. This article unifies previous research and proposes a reference model for self-healing in which terms and functions are unambiguously defined. In addition, this article presents a survey on the state of the art, in which the main references on self-healing are outlined. Finally, the main research challenges in self-healing that will have to be faced in the near future are summarized.

System for automated diagnosis in cellular networks based on performance indicators

This paper presents a system for automated diagnosis of problems in a cellular network, which comprises a method and a model. The reasoning method, based on a naive Bayesian classifier, can be applied to the identification of the fault cause in GSM/GPRS, 3G or multi-systems networks. A diagnosis model for GSM/GPRS radio access networks is also described, whose elements are available in the network management systems (NMSs) of most networks. It is shown that the statistical relations among the elements, that is the quantitative part of the model, under certain assumptions, can be completely specified by means of the parameters of beta density functions. In order to support the theoretical concepts, a model has been built based on data from a real network and the automated diagnosis system has been used to classify problems in a cellular network, showing that the solution is easily implemented and that the diagnosis accuracy is very high, therefore leading to a reduction in the operational costs of running the network. Copyright

On the Potential of Handover Parameter Optimization for Self-Organizing Networks

Self-organizing networks (SONs) aim to raise the level of automated operation in next-generation networks. One of the use cases defined in this field is the optimization of the handover (HO) process, which involves a tradeoff between the amount of signaling load due to HOs and the quality of the active connections in the network. In this paper, first, a sensitivity analysis of the two main HO parameters, i.e., the HO margin (HOM) and the time-to-trigger (TTT), is carried out for different system load levels and user speeds in a Long-Term Evolution (LTE) network. Second, a fuzzy logic controller (FLC) that adaptively modifies HOMs is designed for HO optimization. In this case, different parameter optimization levels (network-wide, cell-wide, and cell-pair-wide) and the impact of measurement errors have been considered. Results of the sensitivity analysis show that tuning HOMs is an effective solution for HO optimization in LTE networks. In addition, the FLC is shown as an effective technique to adapt HOM to different network conditions so that the signaling load in the network is decreased while an admissible level of call dropping is achieved.

Optimization of a Fuzzy Logic Controller for Handover-Based Load Balancing

In Self-Organizing Networks (SON), load balancing has been recognized as an effective means to increase network performance. In cellular networks, cell load balancing can be achieved by tuning handover parameters, for which a Fuzzy Logic Controller (FLC) usually provides good performance and usability. Operator experience can be used to define the behavior of the FLCs. However, such a knowledge is not always available and hence optimization techniques must be applied in the controller design. In this work, a fuzzy

Fuzzy Rule-Based Reinforcement Learning for Load Balancing Techniques in Enterprise LTE Femtocells

Q

Mobility-based strategies for traffic steering in heterogeneous networks

-Learning algorithm is proposed to find the optimal set of fuzzy rules in an FLC for traffic balancing in GSM-EDGE Radio Access Network (GERAN). Load balancing is performed by modifying handover margins. Simulation results show that the optimized FLC provides a significant reduction in call blocking.

Learning of model parameters for fault diagnosis in wireless networks

Mobile-broadband traffic has experienced a large increase over the past few years. Femtocells are envisioned to cope with such a demand of capacity in indoor environments. Since those small cells are low-cost nodes, a thorough deployment is not typically performed, particularly in enterprise scenarios. As a result, the matching between traffic demand and network resources is rarely optimal. In this paper, several load balancing techniques based on self-tuning of femtocell parameters are designed to solve localized congestion problems. In particular, these techniques are implemented by fuzzy logic controllers (FLC) and fuzzy rule-based reinforcement learning systems (FRLSs). Performance assessment is carried out in a dynamic system-level simulator. Results show that the combination of FLC and FRLS produces an increase in performance that is significantly higher than if techniques are implemented alone. Both the response time and the final value of performance indicators are improved.

Management architecture for location-aware self-organizing LTE/LTE-a small cell networks

The large increase in size and complexity experienced by cellular networks in recent years has led to a new paradigm known as heterogeneous networks, or HetNets. In this context, networks with different cell sizes, radio access technologies, or carrier frequencies can be deployed in the same environment. As the coverage area of each of these networks is typically overlapped, operators have some degree of freedom to modify user distributions across the networks (i.e., traffic steering) in order to improve network performance. This article introduces different mechanisms of traffic steering in HetNets, clarifying the specific goals that operators can set and focusing on those techniques that adjust mobility parameters, which are typically more attractive to achieve these goals. In addition, some challenging issues arising from particular HetNet deployments are discussed and illustrated by example use cases, which are applicable to an early stage of LTE deployment. Finally, a fuzzy-logic-based algorithm that optimizes network parameters for traffic steering is proposed.

Automated troubleshooting of a mobile communication network using Bayesian networks

Self-management is essential for Beyond 3G (B3G) systems, where the existence of multiple access technologies (GSM, GPRS, UMTS, WLAN, etc.) will complicate network operation. Diagnosis, that is, fault identification, is the most difficult task in automatic fault management. This paper presents a probabilistic system for auto-diagnosis in the radio access part of wireless networks, which comprises a model and a method. The parameters of the model are thresholds for the discretization of Key Performance Indicators (KPIs) and probabilities. In this paper, some techniques are proposed for the automatic learning of those model parameters. In order to support the theoretical concepts, experimental results are examined, based on data from a live network. It has been proven that calculating parameters from network statistics, instead of being defined by diagnosis experts, highly increases the performance of the diagnosis system. In addition, the proposed techniques enhance the results obtained with continuous diagnosis models previously exposed in the literature.