Preprint of paper for LAPC 2009, Loughborough, UK, November 2009. Not for further distribution. 1 of 4 Computed SAR Levels in Vehicle Occupants Due to On-Board Transmissions at 900 MHz Alastair R. Ruddle Advanced Engineering Department, MIRA Limited Watling Street, Nuneaton, Warks, CV10 0TU, UK alastair.ruddle@mira.co.uk Abstract—Numerical simulations have been carried out to assess the specific absorption rate (SAR) in vehicle occupants due to on- board 900 MHz transmissions. The sources considered include a simple external roof-mounted antenna and internal transmitters that are not in contact with the body. The simulations employed homogeneous human simulants to obtain induced SAR values, and investigated a variety of occupancy configurations, ranging from only the driver to a driver and three passengers. Both local and whole-body average SAR levels are found to vary widely with the number and placement of occupants, as well as with antenna configuration. Nonetheless, the computed SAR levels are found to be well below recommended limits for general public exposure at typical radiated power levels. I. INTRODUCTION The passenger compartments of vehicles are resonant environments containing geometrically complex internal structures. Consequently, electromagnetic energy that is coupled into these regions results in spatial field distributions that are highly inhomogeneous, with localized hot-spots that may be associated with much higher field magnitudes than would result in open environments at the same power level. For example, simulations of in-vehicle field distributions due to 400 MHz transmissions [1] indicate that the maximum fields inside the empty passenger compartment of a car may be more than two orders of magnitude higher than the average. The specific absorption rate (SAR) resulting in vehicle occupants have been investigated experimentally at 146 MHz and 460 MHz [2] for external antenna installations. It is reported that although the empty vehicle field levels exceeded FCC recommendation at power levels around 100 W, the SAR values (measured in a physical phantom located in a car with the front half cut away) did not exceed recommended limits. While the additional losses introduced by adding occupants will act to reduce overall field levels, the number and location of occupants within a vehicle can also be expected to modify the resonances that are present within the passenger compartment. Thus the impact of passengers may not be straightforward to predict. In fact, localized increases in field strength [1] and cable coupling [3] due to the addition of human simulants have been found in vehicle simulations. In addition, numerical investigations at 400 MHz for a roof- mounted antenna [4] and an internal source [5] indicate that the worst-case SAR is not obtained for the driver alone or with three passengers, but arises when one or two passengers are present. Thus, SAR values in the occupants of a car may vary depending on the number and relative position of the occupants, as well as on the location of the source. This paper summarizes computed in-vehicle SAR results for a representative range of vehicle occupancy configurations at 900 MHz, for both a roof-mounted antenna and three internal sources that are not in contact with the human body. II. NUMERICAL MODELS The numerical SAR simulations were carried out using a commercial tool [6] based on the transmission-line matrix (TLM) technique, a full-wave 3D field modelling method [7]. A. Vehicle Model The vehicle model used in these investigations was derived from CAD geometry for the body-shell and doors of a medium-sized passenger car, as well as major metallic parts in the passenger compartment such as the seat frames, interior steering components and rear window heater. Simulations based on similar metal-only vehicle models have been shown to provide a good correlation with spatial electric field measurements carried out on a complete vehicle at 390 MHz [7], and with measured broadband coupling to points inside a complete vehicle at frequencies up to 1 GHz [9]. At higher frequencies, however, it is likely that window glass would also need to be included in the vehicle model since this is found [10] to increase average internal field strengths in simulations, although other dielectric materials, such as seat foam and interior trim, are expected to have little impact on the internal field statistics [11]. B. Human Simulants Seating positions vary between vehicle types, as well as between vehicles of similar type, and even between different seats of the same vehicle. For a phantom representing the driver of a car the arms and feet should be positioned to represent use of the steering wheel and pedals. The front passenger is likely to be in a very similar posture, although with the arms down and feet flat. For the rear passengers, the arms and feet should also be down, but the positions of the