Near-Field Radio Theory

Progress

Periodically, I have been working on a radio simulator for ATEMU, TOSSIM, and AVRORA. These simulators provide only a very simplified radio model, and I am attempting to enhance them with a more realistic model. All simulator radio models account for radio attenuation due to distance, and AVRORA adds some additional capabilities, but for serious radio-based sensory networks, the radio model must account for a lot more.

Theoretical Model - General consensus believes that the radio radiation pattern from a vertical antenna is spherical. This is actually the "theoretical radiator" and does not exist. Most motes use a vertical or dipole radiator, and according to ham radio antenna design texts, the radiation pattern is toroid. The diagram, above, shows a conceptualized cross-section of a such a pattern, where each circle is a higher intensity radiation. The antenna (not visible) runs through the axis of the torus, so receivers above or below the drawing would receive no signal regardless of polarity.
Other antennas have different radiation patterns. For example, the Yagi antenna (somewhat like a television antenna) creates multiple lobes along the polar axis (called the 'beam' of the antenna). The examples, below, show examples of such radiation patterns.

Mathematical Model - A simulation of radio wave propagation shows an approximation of the theoretical model. The contoured maps of a dipole and a Yagi antenna (above) outline the changes in intensity (the inner lobes have higher levels of radiation than the outer ones). In both cases, the lobes expand disproportionately with the distance from the center. The appears to be a better representation of the "inverse square law" from electromagnetic physicals. There does appear to be a skewing, however, from this law where radiation loss is greater than the inverse square.
The left graph shows a dipole or loop antenna radiation pattern; similarly, the right graphs shows the pattern for a Yagi antenna.

Density Model - These images shows the false color propagation density. Note that the colors are arbitrary, and careful observation is required to see contours from earlier drawings. There are three observable artifacts:
First, the inner circles on the left image appear triangular instead of circular as expected. This may be due to inverse-square skewing or even model aliasing from imprecise numbers.
Second, the right image shows radiative lines from the center to the outer edges. The cause is negative/positive bands starting from the center. In fact, according to the image, a properly polarized antenna should receive a signal from anywhere circumscribing the transmitting antenna, but the strongest signal is perpendicular to the beam.
Third, the centers of the images are falsely represented with a moderately 'hot' value. The simulation chopped those values to keep from overshadowing the rest of the image. Near field propagation asserts that the intensities in the center regions are actually very hot, because the inverse square of a distance (d) less than one amplifies the signal in the model to high values.


Copyright © 2001-2006 Intelligent Algorithmic Solutions and Sean Walton	eMail