Using efficient ray‐tracing techniques to predict propagation losses in indoor environments

Abstract

the complete procedure of training and testing, a neural network denoted with 2 N [N] 7 5 was selected. Average and worst case errors for this model are presented in Table 1 as well. It can be observed from Table 1 that, by using the modified neural structure, the accuracy of modeling is improved, compared with the basic simple neural structure. Once trained, the ANN noise model provides an instantaneous response for different input vectors covering the whole operating range. As an illustration, Figure 3 shows the plots of the magnitude of optimum reflection coefficient versus frequency, obtained by using the model 2 N [N] 7 5, at two different bias points: (1) V ce ϭ 2 V, I c ϭ 5 mA; (2) V ce ϭ 2 V, f c ϭ 20 mA where the second one does not belong to the training set. The corresponding measured data are marked in the same figure. An excellent agreement of simulated noise characteristics with measured values can be observed. Finally, it should be pointed out that the developed ANN HBT noise models are implemented within the standard circuit simulator ADS as user-defined library elements, enabling very fast simulation for microwave circuit design. CONCLUSIONSA new, ANN-based approach can be used successfully for modeling the noise parameters of HBTs. For developing a neural noise model, only a number of measured data is needed. That gives an advantage to ANN approach in comparison with other modeling approaches, especially when the noise generating mechanisms of the device are too complex or not well known. Developed neural models are characterized by high accuracy together with the efficiency and simplicity and therefore are convenient for CAD purposes. REFERENCES 1. S. Lee, Direct extraction of base-collector model parameters for AlGaAs/ ABSTRACT: This paper addresses an efficient implementation of the 2.5D ray-tracing propagation model to predict propagation losses in indoor environments. The analysis is based on the Shooting and Bouncing Rays technique and Uniform Theory of Diffraction. Besides the lineof-sight propagation, we consider that the radio waves may experience reflection, refraction, and diffraction (NLOS). The simulated results are validated by the measurement carried out in a teaching building at 1.8 GHz. Validation is also provided by comparisons with published data.

Radiowave Propagation Over an Irregular Surface – The GO and UTD Approach

“…The simulation results, based on a combination of GO and UTD, are in very good agreement with measurements and with other calculation methods (e.g. Physical Optics combined with Physical Theory of Diffraction), which are very often used in an outdoor [4]- [7], as well as an indoor environment [8]- [10]. Propagation over an Manuscript received May 6, 2013; accepted November 22, 2013.…”