Detection of human activity around corners with radar- Measurements in a realistic environment

The idea to use radar for seeing behind corners has been launched by FOI. Our new method and the experimental research results have been met with great interest when presented internationally. The military benefit of this activity is that future systems can be provided with an ability that allows one's own sensor system and soldier to be kept off the line of sight of the enemy, which decreases the risk of being detected and eliminated. This report gives an account of the measurements carried out at FOI with the purpose of investigating if human activity can be detected behind corners with radar under realistic conditions. Earlier experiments carried out on the FOI test range Lilla Gåra showed that the Doppler signature of a moving human behind a corner is unmistakably detectable, both by taking advantage of diffraction in the corner and of multiple reflections in an opposite wall. It has been possible to detect even small movements like breathing from an immobile sitting person. However, these measurements were made under very controlled conditions. The measurement described in the present report takes place in a real environment where the target signal is mixed with natural interference in the scattering environment. The purpose of the report is primarily to describe the measurement itself, but it also contains a first preliminary analysis of the data. The analysis has been made in two ways. In one case a two-dimensional Fourier transform was used, with the analysis made with a manual method, which generates a Doppler spectrum as its final result. The other analysis was carried out with an MTI-like method (MTI = Moving Target Indicator), which could be made automatic. This method uses change detection between successive registrations in order to detect phase changes and, hence, movements. In the analysis, the possibility to detect a concealed walking human was studied. The analyses demonstrate that the human is detectable when pathways from both one and two wall bounces are taken into account. Taking advantage of diffraction turns out to be more problematic since the signal levels of this component are often considerably lower compared with the wall reflection case. Still, we have shown that this detection is possible under special conditions, but it is very uncertain whether it is of practical use in operative system applications.