AbstractA heterogeneous wireless network needs to maintain seamless mobility and service continuity; for this reason, we have proposed an approach based on the combination of particle swarm optimization (PSO) and an adaptive neuro-fuzzy inference system (ANFIS) to forecast a handover during a movement of a mobile terminal from a serving base station to target base station. Additionally, the handover decision is made by considering several parameters, such as peak data rate, latency, packet loss, and power consumption, to select the best network for handover from an LTE to an LTE-A network. The performance efficiency of the new hybrid approach is determined by computing different statistical parameters, such as root mean square error (RMSE), coefficient of determination (R2), mean square error (MSE), and error standard deviation (StD). The execution of the proposed approach has been performed using MATLAB software. The simulation results show that the hybrid PSO-ANFIS model has better performance than other approaches in terms of prediction accuracy and reduction of handover latency and the power consumption in the network.
One of the most challenging issues in the future of heterogeneous wireless networks is to maintain the quality of service (QoS) during a vertical handover (Kassar, Kervella & Pujolle, 2008). In this paper, vertical handover is done between two heterogeneous networks that are LTE and LTE-A. From the existing work, we can conclude that the most important issue is to ensure the Quality of Service (QoS) when the mobile station moves away from one base station towards another (Miyim, Ismail & Nordin, 2014; Aboelezz, Nafea & Zaki, 2020). Thus, the major problems are a large number of handovers, delay in a correct handover decision, deficiency of maintaining a seamless connectivity and service continuity, and high energy consumption in the network. For this reason, a hybrid computing-technique-based handover decision system design is applied to make more intelligent, comprehensive and quick decisions to select the optimal network between LTE and LTE-A (Hashim & Abido, 2019). Therefore, the primary focus was to improve the QoS while the mobile terminal moves across the heterogeneous wireless networks, to reduce the number of handovers and the energy consumption in the network, and to maintain the required connectivity (Davaasambuu, 2018; Chinnappan & Balasubramanian, 2016).
An adaptive neuro-fuzzy inference-system-based handover decision system design is used to make more intelligent, comprehensive, and quick decisions to select the best network (Bin, Xiaofeng & Zianzhong, 2013). Indeed, the primary focus was to improve the quality of service (QoS) while the mobile terminal moves across the heterogeneous wireless networks. Then, we have proposed an approach based on the combination of adaptive neuro-fuzzy inference system (ANFIS) and particle swarm optimization (PSO) with four different parameters (peak data rate, latency, packet loss, and power consumption) to enhance the accuracy of our handover decision algorithm (Benaatou, Latif & Pla, 2017).
The main contribution of this study is to propose a new hybrid approach combining a particle swarm optimization algorithm and an adaptive network-based fuzzy inference system for decision making of handover in wireless networks by using the statistical parameters (Benaatou, Latif & Pla, 2019) the coefficient of determination (R2), the root mean square error (RMSE), the mean square error (MSE), and error standard deviation (StD) to make a comparison between the real and predicted values for the ANFIS and PSO-ANFIS models (Jang, 1993).
The rest of this paper is structured as follows. In the next section, we describe the proposed approach and the structure of the ANFIS model that corresponds to it. Following that, we present the analytical results, the data evaluation for ANFIS-PSO modelling, and the implementation of the PSO-ANFIS algorithm. Next, we carry out a comparison between vertical handover algorithms. Finally, we present the conclusions.
Please refer to PDF download for the full paper.