An indoor resource optimisation for TDMA incorporating a convex space building configuration and accurate ray-tracing

Abstract

The contents of this work are designed for planning wireless telecommunication systems in indoor environments with application intended for time division multiple access (TDMA) systems. The term resource optimisation mentioned in the title refers to the optimisation of the resources that make up a communication system, such as the power, server and mobile unit costs. Defined here is a resource optimisation algorithm that uses at its core a ray-tracing engine. The ray-tracing provides propagation signal strength (coverage) information. The base stations are positioned within predefined cells contained in the building and can be moved subject to a number of constraints at each iteration of the optimisation. Each constraint is related to the capacity requirement of a user, in other words the signal to interference requirement. For an optimisation procedure to be reliable, it is expected that the ray-tracing solution be calculated to a very high degree of accuracy so as to provide the algorithm with adequate information. Using a full wave solution such as an integral equation technique instead to give better results is virtually impossible due to the computational complexity of the problem. In the past authors have relied on empirical coverage calculations that are known to work excellently in outdoor models but are not at all reliable in indoor propagation due to the diffraction and reflection paths being neglected. The problem using a ray-tracing engine at the heart of an optimisation algorithm is that it may require the continual creation of a visibility list every time a server is moved, if that visibility list is dependent on the location of the transmitter. Many well-known ray-tracing algorithms depend on a visibility algorithm which assumes the transmitter is at a fixed location. It will be shown that this is not the case for the new ray-tracing algorithm described here. This means that the ray-tracing allows the optimisation procedures to operate with reasonable computational times, while providing far more accurate propagation coverage in the building than empirical methods. The description of an algorithm that converts the building into convex spaces is one of the main achievements of this work, showing how the building is broken up, and how a visibility algorithm is formed. The optimisation algorithms presented by previous authors use path-loss based models instead of signal to interference ratio based models. This is adequate for GSM but not for TDMA systems where the servers are not communicating directly with one another by cabling, in other words each acts as a separate entity. The chapters of this thesis have a very deliberate progression, going through a process of building the inner workings of an optimisation tool before describing the optimisation itself. The majority of the work consists of using well-known ray-tracing methods such as method of images for reflections, uniform theory of diffraction (UTD) for diffraction paths, and transmissions through and inside different types of media. The shortcomings of ray-tracing are highlighted and novel techniques to correct reflection, transmission and diffraction coefficients are presented in great detail. A brief look at linear programming methods is followed by a downlink and uplink optimisation procedure married together with the ray-tracing results. The algorithm uses a method previously described in the literature, with considerable alteration making it novel in its own right. An attempt is then made to obtain good simulations of a TDMA system finding the best positions of the servers whilst giving the highest possible bit rates to the users of the system.