Naval Radar Analysis with UTD

Read about how the UTD is used to analyse radar antenna placement on a frigate.


Problem setup:

  • Consider a frigate of length 120 m
  • An antenna (e.g. for radar purposes) is located in front of the forward mast, to the starboard side and operating at 10 GHz
  • The objective is to assess the effect of the deck environment on azimuth scanning performance
  • Expected issues to take note of are mast blockage and the path gain effect due to an elevated position above the deck

EM analysis procedure:

  • Evaluate the isolated antenna radiation pattern
  • Import the pattern into the ship deck model and place as radiation pattern point source
  • Use Uniform Theory of Diffraction (UTD) to calculate the far‐field radiation pattern
  • Rotate the source and repeat the analysis

Characterising the Antenna

  • A 3D radiation pattern is needed — the higher the resolution, the better
  • Pattern can be obtained in two ways:
    (1) FEKO simulation
    (2) measurement
  • Here Option 1 is taken
  • As an example, a rectangular waveguide‐fed, E‐plane horn is considered. Note, this can be any antenna of interest.
  • This antenna is analysed in FEKO at 10 GHz, using the Method of Moments (MoM). Results are shown below.
  • The radiation pattern is exported to an *.ffe file (can also export to a spherical mode file).

E‐plane horn antenna Maximum gain is 13.6 dB.

Rectangular waveguide‐fed, E‐plane horn antenna radiation pattern at 10 GHz. Maximum gain is 13.6 dB.

Antenna Placement on the Deck

  • Import the antenna pattern as a point source to any desired location with desired orientation
  • The ship model consists of flat polygonal plates, amenable to UTD analysis
    UTD is based on ray tracing, taking into account reflections, edge diffractions, corner diffractions, multiple bounces
  • Imported to location 50 cm above deck, as shown below.
  • Specify far‐field calculation in a sector in front of the antenna (‐60°, +60°)
Radiation pattern point

Radiation pattern point source imported into deck environment. The far field calculation sector is also shown.

Radiation Pattern vs. Azimuth Angle

  • Computational cost of UTD is proportional to the number of field evaluation points
  • Evaluate effect of azimuth rotation angle of source on pattern
  • Evaluate quickly with a coarsely sampled far‐field calculation at various angles
  • At 340° the most tower blockage is experienced (peak pattern gain shown in brackets in the results figure).
  • Computational cost (laptop): 2 Mbytes, +- 1 minute per angle

Azimuth scanning procedure for the radar pattern point source

Azimuth scanning procedure for the radar pattern point source. The source orientation is shown, as well as the far field patterns evaluation sector, which is rotated with the source.

Far field pattern results of the azimuth scanning procedure. The peak gain value in each pattern is indicated in brackets.

More Detailed Analysis of the Pattern at 340°

  • Recalculate at 340°, using much finer far‐field sampling points
  • Lobes due to path gain are now all resolved (antenna is 16.67λ above the deck)
  • Computational cost (laptop): 2 Mbytes, +- 1 hour
 340° scan angle

Finely sampled pattern at 340° scan angle. Height of the source is 50 cm above deck.

Effect of Lowered Position on Pattern

  • Still pointing at azimuth 340°, lower the position of the antenna to 9 cm (3λ) above the deck
  • As expected, lobes due to path gain are reduced in number
  • Effect of tower on azimuth gain variation is still essentially unchanged
  • Computational cost (laptop): 2 Mbytes, +- 1 hour
340° scan angle

Finely sampled pattern at 340° scan angle. Height of the source is 10 cm above deck.


  • The UTD, combined with radiation pattern source importing is ideally suited to the analysis of radar antennas in a naval environment.
  • Computational cost is dominated by CPU time, rather than by memory requirements.
  • UTD is ideally suited to parallelisation, which will reduce the runtime by a factor equal to the number of processes. The UTD is fully parallelized in FEKO.

Visit the website