Call Admission Control In Lte Femtocell/Macrocell Systems Based On Fuzzy Neural Network

As most of bussiness occurs in the room according to the experience of theoperator, it is an important part to enhance the signal coverage of indoor area. Theconcept of “Femtocell” is introduced in order to meet the requirements of highquality bussiness in indoor environment. But interference between them is animportant problem because its deployment is always random. Therefore Femtocelluplink interference is firstly analysed and simulated. The simulation results are usedin the problem of call admission control for mobile users.Many factors are considered for mobile users to access to networks. The inputparameters for multiple attribute decision are always precise values. But in real life,many concepts can not be described by the mathematical parameters accurately, sofuzzy mathematics is introduced in this paper. Membership function is an importantpart in fuzzy mathematics and the determination of its weights a big problem. Hereneural network is used to solve this problem because its self learning ability. Aboveall fuzzy neural network is used in the call admission control in this paper.In this paper user speed pre-decision module is proposed in order to avoid thePing-pong effect in handover procedure. The users in high speed are accessed to theMacrocell because of the small range of the Femtocell. Besides network loadbalancing plays a crucial role in improving network source utilization, so thenetwork load is introduced as a key parameter in access control procedure.This paper gets an intelligent call admission control algorithm throughestablishing fuzzy neural network model of four levels and which is trained byexpert data sample. By the establishment of the user movement model, the accesscontrol procedure could be truly reflected in this heterogeneous environment. Withthis simulation scenario, the parameters of this algorithm and the traditionalalgorithm only based on RSS during the procedure of access and handover could besimulated, compared and analyzed.Simulation results show that the algotrithm proposed in this paper improves theperformance of Ping-pong rate comparing to the traditional algorithm which is onlybased on received signal strength. This algorithm also has a lower blocking rate andhandoff dropping rate because it has a better utilization of network resource. Thecall admission control algorithm based on fuzzy neural network can enable thenetwork to accommodate more users.