Modelling of Radiowave Propagation through Finite Vegetation Structures based on RET Telmo R. Fernandes *† , Rafael F. S. Caldeirinha *† , Miqdad O. Al-Nuaimi † , Jürgen Richter † *Instituto Politécnico de Leiria, ESTG, Leiria, Portugal *Instituto de Telecomunicações, Pólo de Coimbra, Coimbra, Portugal † Radiowave Propagation and Systems Design Research Unit, University of Glamorgan, CF 37 1DL, United Kingdom Abstract 1 In this paper a numerical discrete approach to the RET (Radiowave Energy Transfer) applied to vegetation is presented. The proposed method has the advantage of allowing us to estimate the radio signal attenuation of well defined finite and non-homogeneous vegetation structures, by dividing the forest in several elementary volumes. A new discrete scatter or “phase” function is also presented, leading to increased validity to a wider range of RET input parameters. Simulation results are compared with appropriate measurements made at two different frequencies, i.e. 1.9 and 12 GHz. I. INTRODUCTION Costumer demand for robust broadband mobile and fixed radio communication systems has been growing considerably and is expected to keep rising in the foreseeable future. Such modern radio communication systems will vastly rely on suitable planning tools and on the availability of accurate radiowave propagation models. Radio propagation is influenced by any scatter or obstacle present in the radio path, which may include buildings, terrain profile, vegetation, etc. In urban and rural macrocells, vegetation of any sort is a special type of obstacle that is likely to interfere with the radio path [1]. Models associated with radiowave propagation through vegetation are therefore essential in planning of reliable communication systems [1]. Most of the existing radiowave propagation models, although simple to implement, contain relatively few parameters related to the physical characteristics of the vegetation medium, e.g. leaf size, trunks and twigs dimensions, foliation state, etc. Essentially, three types of vegetation propagation models are available from literature [1]: • Empirical models; • Semi-empirical models; • Theoretical models. Empirical models mostly rely on measurement data and curve fitting, as no vegetation attributes are taken into account, their accuracy is confined to the scenarios in which the measurement data was acquired. Semi-empirical models, even though they are, to a certain extent, dependant of The authors would like to thank the UK vegetation consortium formed by Qinetiq, Rutherford Appleton Lab and the Universities of Portsmouth and Glamorgan, for providing the measurement data presented in this paper. vegetation parameters, their overall performance may not be satisfactory in some cases [1]. Theoretical models, like RET (Radiowave Energy Transfer), are normally more complex as they are hooked on electromagnetic properties and physical characteristics of the medium. The RET, in particular, has been extensively used in vegetation modelling [1]. Providing that certan parameters are established for a particular vegetation type, the RET is able to estimate the attenuation as a function of vegetation depth [1] [2]. Although, the RET is able give a good estimation of the scattered radio signal inside a vegetation medium, its applicability may be limited [1]. For instances, the RET is mostly applicable to idealised homogeneous vegetation media, with an infinite dimension air to vegetation interface, when illuminated by a uniform plane wave. In practice, the air to vegetation interface is finite, as it is limited by the ground and the forest dimensions. Forests are normally non homogeneous as they are formed by several different species with different dimensions and foliation states. Also trees tend to grow more on the edges of the forest as they are more exposed to sun light. In addition, the wave illuminating the interface may not always be uniform, e.g. when narrow beamwidth antennas are used. In order to overcome these limitations, a d iscrete approach to the RET (DRET) is presented in this paper. The RET and its extension to the DRET are presented and compared in section II. Section III presents the measurement procedure used to validate both the RET and DRET. The validation of the models is given in section IV where simulation results of DRET and the actual measurements are compared and differences are discussed. From this, conclusions and further work to improve DRET performance is presented in section V. II. MODEL DESCRIPTION A. Description of RET The RET attempts to model the vegetation as a statistically homogenous medium of scatterers ds , which are characterised by three intrinsic parameters, i.e. an absorption cross section a k , a scatter cross section s k and a scatter directional profile ( ) , ' pss , also known as “phase” function. The incident and emanating directions over the scatterers ds is illustrated in Fig. 1. The RET equations describe the radio signal propagating through the medium in terms of its