Human polarimetric micro-doppler

Abstract

Modern radars can pick up target motions other than just the principle target Doppler; they pick out the small micro-Doppler variations as well. These can be used to visually identify both the target type as well as the target activity. We model and measure some of the micro-Doppler motions that are amenable to polarimetric measurement.Understanding the capabilities and limitations of radar systems that utilize micro-Doppler to measure human characteristics is important for improving the effectiveness of these systems at securing areas. In security applications one would like to observe humans unobtrusively and without privacy issues, which make radar an effective approach. In this paper we focus on the characteristics of radar systems designed for the estimation of human motion for the determination of whether someone is loaded.Radar can be used to measure the direction, distance, and radial velocity of a walking person as a function of time. Detailed radar processing can reveal more characteristics of the walking human. The parts of the human body do not move with constant radial velocity; the small micro-Doppler signatures are timevarying and therefore analysis techniques can be used to obtain more characteristics. Looking for modulations of the radar return from arms, legs, and even body sway are being assessed by researchers. We analyze these techniques and focus on the improved performance that fully polarimetric radar techniques can add. We perform simulations and fully polarimetric measurements of the varying micro-Doppler signatures of humans as a function of elevation angle and azimuthal angle in order to try to optimize this type of system for the detection of arm motion, especially for the determination of whether someone is carrying something in their arms. The arm is often bent at the elbow, providing a surface similar to a dihedral. This is distinct from the more planar surfaces of the body and allows us to separate the signals from the arm (and knee) motion from the rest of the body. The double-bounce can be measured in polarimetric radar data by measuring the phase difference between HH and VV. Additionally, the cross-pol and co-pol Doppler signatures are analyzed, showing that the HH polarization may perform better on dismounts in open grass.

Radar micro-Doppler simulations of classification capability with frequency

“…The ability to consistently isolate arm motion from other body parts using range or frequency has been shown to be challenging. Additional information, like polarization information [34] may be required. Alternatively, the look-down angle can be used to change the range profile, more effectively isolating the body parts using range.…”

Micro-Doppler Based Target Recognition With Radars: A Review

“…Polarimetric micro-Doppler analysis can also aid in separation of different components of human motion. The double bounce of radar signal from dihedral joints like elbow and knees, can be separated from planar body surfaces by measuring the phase difference between HH and VV [126,127]. Use of dual polarization radar can be used to improve the classification accuracy for body gesture recognition, by combining the micro-Doppler signatures obtained using HH and HV [128].…”

Review of micro‐Doppler signatures

“…Fully polarimetric radar data can potentially improve the separability of different parts of human motion. One of the techniques to determine whether people are carrying something in their arms is looking at their arm motion [84, 85]. The arm is often bent at the elbow, providing a surface similar to a dihedral, which creates a double bounce of the radar signal.…”

“…The use of multistatic and micro‐range micro‐Doppler signatures may significantly improve the basis for mapping signatures to motions, and multistatic signatures have already been shown to improve the classification of armed against unarmed humans [122, 123]. Polarimetric capabilities may provide another method for recognising dihedral‐like locations in motion such as elbows and knees [84], but is not yet clear that the improvement would be significant. Utilising newer sensor and processing techniques to help separate and disambiguate the micro‐Doppler signals has been leading to significant improvements in the field, including sub‐signature differentiation such as a horse from a horse with a rider or a damaged against a fully functional helicopter.…”