Considerations on Sensor Fusion of Multiple Ultrawideband Radar Sensors Operating in Non-Adjacent Frequency Bands

Abstract

This paper presents a novel sensor fusion approach to enhance radar measurement bandwidth and range resolution by integrating data from non-adjacent frequency bands. While ultrawideband (UWB) radar systems offer high resolution, they are often constrained by regulatory limitations and hardware bandwidth restrictions. To overcome these challenges, we investigate on merging intermediate frequency signals from multiple frequency-modulated continuous wave (FMCW) radar sensors operating in separate bands. Though this effectively broadens the usable bandwidth, challenges arise from uncovered spectral regions in between the sensor bands. A frequency domain model is employed to address these systematic challenges in multi-band fusion and to quantify side-lobe-levels and pulse-widths. Furthermore, we discuss the establishment of a common phase reference by means of calibration. The investigations are validated through simulations and experimental measurements using W-band (68–93 GHz), D-band (122–170 GHz), and J-band (205–248 GHz) FMCW sensors. Finally, it is shown that model-based interpolation between the bands significantly removes undesired distortions. Results demonstrate a significant enhancement in range resolution, particularly benefiting applications such as non-destructive testing and high-precision material characterization. In these applications, the approach provides a viable alternative to photonic and optical measurement techniques, leveraging the advantages of compact, MMIC-based radar sensors while overcoming inherent bandwidth limitations.