Guided FMCW radar offers a way to simplify millimeter-wave and terahertz sensing hardware: a dielectric waveguide carries signals between a radar transceiver and the measurement target, avoiding the lenses and mirrors used in conventional quasi-optical arrangements. For industrial automation engineers and inspection teams, the approach is relevant to non-destructive testing (NDT) where compact sensing, distance information and imaging are valuable. Its practical performance, however, depends on how the radar couples to the guide and how the emerging beam interacts with the sample.
How guided FMCW radar measures distance
A frequency-modulated continuous-wave (FMCW) radar repeatedly transmits a signal whose frequency changes across a defined bandwidth. A reflected signal returns after a delay. Mixing that return with the transmitted reference produces a beat frequency related to the signal’s round-trip travel time and, therefore, the reflector’s distance.
This distance encoding distinguishes FMCW reflectometry from continuous-wave measurement. A continuous-wave receiver may combine reflections generated at multiple points along a sensing path. FMCW processing can instead separate returns by their delay, allowing the measurement to focus on a selected distance range. In a guided setup, that means reflections from the waveguide and its surroundings can be distinguished from a return associated with the probe end or sample, subject to the system’s resolution and signal quality.
Range resolution is tied to sweep bandwidth: a wider frequency sweep supports finer separation of targets along the propagation direction. It is important to distinguish this longitudinal resolution from lateral imaging resolution, which depends on the beam profile and the geometry at the waveguide output.
Why use a dielectric waveguide?
Conventional millimeter-wave and terahertz measurement systems often use lenses or parabolic mirrors to direct energy between a transceiver and a target. Those components can require careful alignment and constrain the mechanical layout. A hollow-core dielectric waveguide offers an alternative transmission path, potentially simplifying the connection between the radar head and the sensing location.
The demonstrated configuration used a thin-wall polypropylene pipe waveguide. Its low-loss guiding behavior relies on the interaction between the air core and the dielectric cladding, including anti-resonant reflections. The design must account for cladding thickness, material properties and diameter. Diameter affects how effectively the radar couples energy into the guide, while the guide’s dimensions also influence the output beam and achievable imaging resolution.
Field simulations showed that more than one propagation mode can be present. Alongside energy guided in the air core, fields can extend into or around the cladding. This leakage may support sensitivity to nearby perturbations, but it can also contribute to unwanted reflections or image artefacts. Waveguide selection and mechanical positioning are therefore part of the sensing design, not just packaging decisions.
Coupling and dynamic range shape system performance
The reported setup was evaluated with two transceiver types: a 100 GHz III-V-based unit with a conical horn antenna and a compact 122 GHz silicon-based transceiver with separate transmit and receive patch arrays. The comparison illustrates how the front-end antenna architecture influences coupling to the same general class of waveguide.
In the horn-based configuration, the guide was inserted into the antenna for centered, relatively direct coupling. Simulations estimated a coupling efficiency approaching 70% once guided propagation reached a steady state. With the patch-array transceiver, the broader antenna pattern and transmit-receive separation created a less symmetric interface; the estimated coupling efficiency was 18%. These are configuration-specific simulation results, not general performance guarantees for other radar heads or waveguides.
Coupling losses affect the signal available for sensing. Another constraint is the reflection from the open end of the guide, which can act as a background return. Simulations estimated maximum dynamic ranges of about 27 dB for both evaluated configurations, despite their different coupling efficiencies. In practice, the usable measurement range depends on the complete setup, including the transceiver, waveguide, target and signal processing.
A foam absorber was placed near the coupling region in the experiments to reduce parasitic reflections from signals that did not enter the waveguide. This highlights a practical point for integrators: simplifying the optical path does not eliminate the need to manage stray energy and reflections at interfaces.
Imaging resolution and artefacts
Raster scans of a resolution target demonstrated imaging with both guided configurations. The beam emerging from the horn-coupled guide had two main lobes close to its output. At a short imaging distance, this uneven profile produced duplicated features in the image. After propagation in free space, the lobes spread and the profile became more uniform; the reported optimum lensless resolution was 4.5 mm at a 5 mm working distance.
The patch-array setup produced an asymmetric output field, but its single-lobe profile avoided the same type of prominent duplicated image features. The reported lensless resolution was 2.8 mm at the waveguide output. The differences show why antenna selection alone does not determine image quality: coupling geometry, output-field shape and working distance all matter.
To improve lateral resolution, the study also tested a hemispherical high-density polyethylene solid immersion lens at the waveguide end. With the 100 GHz configuration, the reported image resolution improved to 2 mm at a working distance of around 2.5 mm. With the 122 GHz setup, a 1.4 mm resolution was reported using the lens. The lens adds a focusing element while retaining a guided transmission path, but its placement and the desired working distance must be considered in the probe design.
What this means for industrial inspection design
Guided FMCW reflectometry is a promising architecture to assess when a measurement system needs a more direct, compact connection between a radar transceiver and a sensing point. Distance-resolved returns can help separate sample information from reflections along the path, while raster scanning can turn measurements into images. The reported work establishes feasibility, rather than a universal specification for industrial deployment.
Before integrating this approach into an inspection station or portable instrument, engineering teams should evaluate:
- Transceiver and guide compatibility: antenna pattern and positioning affect coupling efficiency.
- Range resolution: the FMCW bandwidth must suit the separation of features or interfaces being measured.
- Beam profile: output lobes and asymmetry can affect image fidelity and create artefacts.
- Working distance: back-coupling and lateral resolution can change as the target moves away from the guide.
- Reflections: interfaces and uncoupled signals may create background returns that limit usable dynamic range.
The practical takeaway is that guided FMCW radar can reduce reliance on free-space optical components while retaining distance-sensitive measurement and imaging capabilities. Successful implementation still requires coordinated choices in radar front end, waveguide geometry, coupling, signal processing and probe termination. For industrial NDT, these factors should be validated against the target material, inspection geometry and required resolution before the system is treated as a production-ready solution.