A continuous-space analytical approach for relay node placement in hybrid cellular and ad hoc networks

Abstract

Abstract-A hybrid network is composed of a cellular component and an ad hoc component connected by a relay node, for the purpose of coverage extension and/or capacity improvement. In this paper, we analyze the capacity of such a hybrid network by employing a continuous-space analytical methodology based on circular geometry for uniformly distributed nodes. To achieve maximal overall capacity, the relay node needs to be placed in an optimum location between the base station and the mobile station located at the boundary of the hybrid network. Numerical results show that for obtaining the optimum overall capacity for the hybrid network, the placement of the relay node should be in a range which is neither too close nor too far away from the base station. For a given node density and path-loss coefficient, a precise location for relay node placement to achieve maximum overall capacity can be found using the presented method. I. INTRODUCTIONTremendous amount of interest from the research community has been raised in recent years to enhance the coverage of the mobile cellular networks while maintaining better link quality and low cost. The improvement in the capacity of such networks has also been an active research topic. While there have been many attempts for increasing the cellular range (similarly for the capacity), there are practical and/or natural fundamental limits [1] on these quantities. For example, transmission power of a Base Station (BS) can not be increased beyond the limit defined by regulations. Similarly, the transmission power of a Mobile Station (MS) has to be kept within an optimum level due to, for example, interference and battery life considerations. When the coverage area of a BS becomes larger, interference increases due to higher number of interfering nodes, leading to reduced capacity and lower supported user data rate. By reducing the coverage area of the BS, the benefit of spatial reuse of spectrum is achieved. However, this benefit has to be traded off with the installation cost of a BS. As a consequence, the limits on the number of BSs in an area may create uncovered geographical spots. Moreover, the quality of signal in a circular region near a BS is better than in the far region (assuming omnidirectional antennas). Thus, the question of how to improve and extend the coverage (as well as capacity) of the network in the far region from the BS remains valid.On the other hand, a good amount of research work has been carried out by the scientific community in the field of mobile ad hoc networks [2]. In an ad hoc network, two MSs can communicate with each other, either directly or via multihop relay, but the range of communication between two neighboring nodes is relatively short compared with the range between an MS and the BS in cellular networks. Some other advantages of ad hoc networks are that they can be easily deployed in an infrastructure-less manner and are of economically low cost.A natural question is what happens if we integrate both the pure cellular and pure ad hoc network t...