Signal Processing Techniques for Spatial and Frequency-Varying Wave Propagation Through Multi-Layer Stack-Ups

Abstract

Radar is classically used over optical sensors to sense objects regardless of weather or daylight conditions. In this case, a waveform is transmitted through the air and reflects off the outside of an object back to the radar. Radar has since been extended to sense within objects. Common applications are non-destructive evaluation, ground penetrating radar, and remote sensing. In this case, the incident wave is reflected by the contrast of electrical properties between materials. More recently, radar has been leveraged for biomedical applications, such as vital sign sensing and imaging within the body. Biomedical imaging radar (BIR) is a promising non-ionizing method to sense within the body. Some potential features to extract could be tumors, brain bleeds, or foreign objects.

A general assumption in radar operation is that the reflected waveform has the same structure as the transmitted waveform. The wave's velocity of propagation is dependent upon the medium and the frequency. Thus, if the reflected wave has traveled through a medium other than air, the received waveform is more stretched out in time than the transmitted waveform. Applying a traditional matched filter in this case yields a degraded range profile, and the returns do not appear at the correct physical location. If the wave only travels through air and one other non-dispersive medium, the range calculation can be easily adjusted. This scenario is commonly encountered in remote sensing. However, more complex scenes such as the human body, are composed of several media with electrical properties that vary across frequency at different rates. Existing techniques are not able to fully leverage radar's pulse compression gain in this case.

In this research, the challenge of radar wave propagation through multiple media is addressed. First, the wave propagation mechanics are studied to understand how the received waveform is distorted. Then, a matched filter is adapted to compensate for this spatial and frequency-dependent distortion in the frequency-modulated continuous wave (FMCW) radar case. The compensation scheme is demonstrated in simulation, and then an FMCW prototype system is built to apply the velocity correction to measured data. The proposed compensation technique is successfully applied to measure a scene with a metal plate placed immersed in a box of oil at various ranges, and more advanced range profile enhancement is explored. The proposed technique is shown to overcome a crucial challenge faced by a BIR.