Radar and Radar Imaging of Cooperative and Uncooperative Modulated Targets

Abstract

Radar and radar imaging technologies are widely used in the exploration of: (1) cooperative modulated targets that intentionally change their reflectivity to backscatter encoded signals, (2) uncooperative modulated targets that unintentionally modulate radar signals due to micro-motions, and (3) unmodulated targets that only reflect but do not modulate radar signals. In this work, a general model is proposed to encompass all three types of targets. This model unmasks the system frequency response when a general modulated target is included in the channel. It treats this target as a group of hypothesized point scatterers that modulate the radar signal with time-variant reflection coefficient values and time-variant point scatterer distribution. Additionally, the model accounts for stationary or moving radar platforms. This allows backscatter signals from different targets to be modeled and processed using the same framework. To use the model on cooperative targets, we first develop a low-cost X-band backscatter testbed and fabricate sensor tags used as cooperative targets. The system shows a good agreement between measured and simulated free-space path loss at up to 4.1 m and demonstrates a BPSK backscatter modulation at 10 Mbps. Then an extension of synthetic aperture radar (SAR) techniques is proposed to enable simultaneous imaging of reflective objects (unmodulated targets), sensor tag (cooperative target) localization, and backscatter data uplink from multiple tags in a cluttered environment. SAR measurements are implemented at 10 – 13 GHz with unmodulated targets and two sensor tags that monitor temperature changes and encode data with balanced orthogonal SAR code words in each packet. The resulting point-spread functions (PSFs) of tags at ranges of 4.4 m and 4.7 m demonstrate a range resolution of 4.7 cm, a cross-range resolution of 9.1 cm, and a maximum localization error of 9 mm. In addition, 1-bounce multipath ghosts (from a large reflective plane) of the tags are predicted by using virtual point scatterers with a maximum error of 1.4 cm. To apply the model on uncooperative targets, backscatter signal features of an off-the-shelf quadcopter UAV are extracted to characterize the reflectivity, static and dynamic radar cross sections, and the micro-Doppler signatures of the quadcopter using a custom K-band (15 – 26.5 GHz) laboratory radar system. The exploration of micro-Doppler signatures includes interference analysis between multiple spinning propellers and Doppler spectra analysis based on controlled rotation rate measurements. The magnitude of Doppler frequencies and Doppler bandwidth differ, even for a fixed rotation rate, and depend on the propeller orientations relative to the radar unit. In the modeling of a single propeller, such differences are explained by correlations between the simulated backscatter signals and the measured receive signals in different orientations. With only twelve point scatterers modeling the propeller, several sets of reflection coefficient values are found to reconstruct the measurement results with a maximum root mean square (RMS) error of 8 dB across all harmonic components in the asymmetric upper and lower sidebands.