The measurements of the human torso for two individuals are presented via the generation of the Huynen polarisation fork technique and plotted on the Poincaré sphere, to ascertain characteristics that could be used to remove the effects of the torso when concealed weapons are placed against it. Measurements are taken with a frequency modulated continuous wave (FMCW) mono-static millimetre wave full polarimetric radar, operating at k-band (18 to 26 GHz). The system has been designed to explore the feasibility of using full polarimetry for the detection of concealed weapons, and person borne improvised explosive devices (PBIED). The philosophy of this scheme is a means to extract the maximum information content from a target which is in the nominally single spatial pixel (sometimes sub-pixel) configuration of stand-off (tens of metres) and crowd surveillance scenarios. The radar comprises a vector network analyser (VNA) and an orthomode transducer.

The Huynen polarisation fork as a representation of the full polarimetric radar signature represents a unique and natural description of a target. This paper investigates the use of full polarimetric radar operating over the band 18-26 GHz to measure the Huynen target parameters of size, orientation, helicity, skip angle and fork angle from a range of canonical polarimetric radar targets and classic concealed weapon surrogates. Measurement will determine how accurately the Huynen target parameters represent the geometries of the canonical targets and surrogate weapons such as concealed metal and ceramic guns, shrapnel and plane sheet dielectrics. Target backgrounds will be large area absorbers and the human body to enable assessment of the capability for stand-off concealed weapons detection. The system used for the measurement comprises a dual channel vector network analyser, a Turnstile orthomode transducer (OMT) and a conical horn antenna. The OMT has an isolation better than-35 dB between orthogonal polarisations. This system measures and calibrates the Sinclair matrices of targets, from which the Huynen target parameters are derived. A simple model of targets based on the original work in Huynen thesis back in 1970 will be presented, enabling comparison between measured and simulated Huynen target parameters to be made. Conclusions are that experimental measurements of the Huynen target parameters of canonical and surrogate targets agree well with the basic theory of the technique and simple model simulations.

Surrogate explosives and shrapnel weapons at a range of 2 metres have been measured using a full polarimetric radar operating over the band 18-26 GHz. Measurements of these items were made as they were standing by themselves and as they were placed on the body, under light clothing. These measurements were compared with measurements made of the divested human body. The polarimetric radar comprised a vector network analyser, and orthomode transducer and a waveguide conical horn antenna. The measurements were analysed using the Euler/Huynen decompositions and the Cloude/Pottier decompositions working on the coherency matrix, as derived from multiple time sequence measurements. The results conclude that the signature of threat items changes considerably when they are placed on the human body. The measured signature of the threat item on the body appears to be somewhere between that of the threat item when it is by itself and that of the divested human body.

Finite-difference time-domain simulations have been made of a security screening polarimetric radar over the band 18 GHz to 26 GHz, comparing the results with a proof-of-concept system operating over the same band. The proof-of-concept radar is presented together with its calibration and measurement set-up. Measurements indicate the cross-polarisation returns from a human subject are approximately 10 % to 25 % of the co-polarisation returns. A simulation model has been built using the openEMS software to simulate the body of a human, using realistic primitive shapes and electrical properties appropriate for these frequencies, indicating cross-polar returns are in the region of 15 % of the co-polar responses, with the duration of the reflections lasting around 2 ns. The comparisons between the measurements and simulations are good and provide a qualitative understanding of what happens when security screening radar radiation impinges on the human body. The simulation is extended to two simple enclosures, a cubic box and a short cylinder having dimensions of 300 mm and wall thicknesses of 5 mm, which could be made of wood, cardboard, paper or plastic. Results indicate the cross-polar reflection ranges from 3 % to 75 % of the co-polar and bursts of reflections are commensurate with reflections from the front and back surfaces, these being separated in time by 2 ns.