304 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 48, NO. 2, MAY2006 Radio Frequency Electromagnetic Fields in Large Conducting Enclosures: Effects of Apertures and Human Bodies on Propagation and Field-Statistics Martin P. Robinson, Janet Clegg, and Andy C. Marvin, Member, IEEE Abstract—Radio frequency propagation in an electrically large resonant chamber (a screened room) was simulated by two models: a statistical combination of multiple resonant modes and a compu- tational electromagnetic simulation [the transmission line matrix (TLM) method]. The purpose of this work was to investigate the effects of passengers and windows on electromagnetic fields (EMF) in aircraft and other vehicles. Comparison of the multimode mod- els with measurements made in a screened room showed that as the electromagnetic losses increased, the transmission between two in- ternal antennas was reduced, and there were fewer turning points in its frequency response. The autocorrelation of this frequency response provided a useful estimate of the composite Q-factor of the resonances and showed that the Q of the chamber was reduced from a value of the order of 10 000 when emptied to 1000 when windows were added and when filled with people to 100. The TLM simulation provided further useful information about the statistical variation of electric field strength with position. Index Terms—Aircraft, cavity resonators, electromagnetic com- patibility, electromagnetic fields, propagation, Q-factor, statistics, transmission line matrix methods, vehicles. I. INTRODUCTION W HEN considering the propagation of electromagnetic fields (EMF) in large conducting enclosures, it is only feasible to apply analytical methods so long as the number of resonant modes does not become too great. For a highly over- moded enclosure, it is necessary to adopt a statistical approach to propagation and shielding problems [1], and any statistical model will need to take into account the various mechanisms for electromagnetic loss that will affect the cavity resonances of the enclosure. In this paper, we consider enclosures that are large enough to hold one or more people, and we describe two methods of investigating the internal propagation at radio fre- quencies: a statistical combination of Lorentzian resonances that reproduces important features of the frequency response and a computational electromagnetic model that gives the statistical variation of field strength with position. A motivation for this research was to study propagation in air- craft, trains, and other forms of transport. It has been suggested that when sources of radio frequency emissions, such as mobile phones, are used in aircraft cabins or on trains, the metallic en- closure acts as a resonant cavity, leading to “hot spots” where the EMF are enhanced. This could increase the severity of elec- tromagnetic compatibility (EMC) problems from such devices. Manuscript received August 5, 2005; revised November 25, 2005. The authors are with Physical Layer Group, Department of Electronics, Uni- versity of York, Heslington, York YO10 5DD, U.K. Digital Object Identifier 10.1109/TEMC.2006.873856 However, in reality, the Q-factors of resonances will be reduced by the windows, which allow electromagnetic energy to escape, and by the presence of the passengers, whose bodies will absorb some of the energy. An understanding of these effects will aid the investigation of potential interference from mobile communica- tion devices [2] and also of human exposure to EMF. Statistical methods have been used before by Panaretos et al. [3], who used a mechanical 1:20 scale model of a Boeing 757 fuselage and measured its shielding effectiveness from 400 MHz to 13 GHz (corresponding to 20–650 MHz in a real aircraft). Their model showed the usefulness of statistical analysis of a frequency re- sponse, but did not include the effects of the passengers. It is therefore appropriate to investigate the damping effect of human bodies and apertures on the internal fields generated in an enclosure by a small source. Human bodies are known to reduce the Q-factors of resonances in aircraft [4]. We wished to quantify this behavior, and its effect on the statistical vari- ation of the electric field strength E. To do this, we utilized a screened room, which is a suitable model for a preliminary study because it has a similar cross section to a small commer- cial airliner and can be loaded with the same density of people per unit volume as a real aircraft. It can be easily fitted with antennas, and being rectangular is convenient for computational electromagnetic modeling (although nonrectangular enclosures can also be studied by such techniques if a stepped approxima- tion to the surfaces is used). In this study, we disregarded the possible electromagnetic losses in wiring looms, seats, fabrics, and thermal insulation, and considered just the contributions from bodies and windows. We applied our two models to experimental measurements made over the frequency range 900–920 MHz, which is of in- terest because it is used for GSM mobile (cellular) phones in many countries and is also a frequency range at which aircraft cabins and train carriages are electrically large. II. THEORETICAL BACKGROUND The presence of a human body in a large resonant cavity is known to perturb its resonances, lowering both the resonant frequencies and their Q-factors [5], [6]. At 59 MHz, the presence of a single person in a screened room (details given below) reduces the Q from 8000 to between 5 and 700, depending on the orientation of the body with respect to the E-field. This is due mainly to the body’s water content, because water has a much higher permittivity than other constituents of the body. For the lower order modes, it is possible to use resonant perturbation 0018-9375/$20.00 © 2006 IEEE