Ryan K. Hersey

Dismount modeling and detection from small aperture moving radar platforms

Future advanced radar systems must detect targets of diminishing radar cross section (RCS) at low radial velocity, in demanding clutter and interference environments. Presently, a deficiency in radar detection performance exists between the capabilities of synthetic aperture radar (SAR) for fixed target indication and space-time adaptive processing (STAP) for ground moving target indication (GMTI) of targets with low ground track velocity. Dismounts, individuals or groups running, walking, or crawling, constitute a class of targets that falls into this netherworld between SAR and STAP. While possessing low RCS levels and radial velocities, dismount detection is rendered even more challenging due to their complicated non-linear phase histories that give rise to significant micro-Doppler energies. In this paper we develop a physiological human-gait model for multi-channel moving radar platforms. We characterize the dismount detection performance of a notational UAV system using linear phase, quadratic phase and sinusoidal phase filters. Finally, we summarize our results and present areas of future work.

Real-time implementations of ordered-statistic CFAR

Ordered-statistic constant false alarm rate (OS-CFAR) detectors provide improved robustness over cell-averaging CFAR (CA-CFAR) detectors in multiple target and heterogeneous clutter environments. However, this benefit comes at the cost of generally increased processing time due to the need for a rank-ordering of the CFAR training data. Realtime implementations of OS-CFAR must consider this additional processing burden. In this paper, we present real-time FPGA and CPU/GPU implementations of OS-CFAR. A novel sorting architecture that scales linearly with window size is presented alongside traditional compare-and-swap and rank-only architectures in an FPGA. A rank-only GPU implementation is demonstrated alongside multi-threaded sorting and rank-only CPU implementations. Effects of training window size on throughput and power consumption are considered.

Adaptive filtering for conformal array radar

Conformal arrays possess certain desirable characteristics for deployment on unmanned aerial vehicles and other payload-limited platforms: aerodynamic design, minimal payload weight, increased field of view, and ease of integration with diverse sensor functions. However, the conformal arraypsilas nonplanar geometry causes high adaptive losses in conventional space-time adaptive processing (STAP) algorithms. In this paper, we develop a conformal array signal model and apply it to evaluate the performance of conventional STAP algorithms on simulated ground clutter data. We find that array-induced clutter nonstationarity leads to high adaptive losses, which greatly burden detection performance. To improve adaptive performance, we investigate the application deterministic and adaptive angle-Doppler warping techniques, which align nonstationary clutter returns. Through the application of these techniques, we are able to nearly fully mitigate the nonstationary behavior yielding performance similar to that of a conventional planar array.

Clutter-based array calibration for enhanced geolocation accuracy

Accurate target geolocation is critical to ground moving target indication (GMTI) performance. Large geolocation errors can make interpreting GMTI detection results difficult, particularly in dense target environments. Geolocation errors can result from a combination of antenna channel errors and platform inertial navigation system (INS) biases. In this paper we develop clutter-based array calibration techniques that remove unknown, channel-to-channel, phase and amplitude errors. These clutter-based techniques utilize distributed clutter for calibration and require no a priori knowledge of the scene or calibration targets. We also apply these techniques to estimate the platform orientation, which significantly improves cross-range geolocation accuracy as compared to platform INS orientation estimates. We present results on measured X-band data.

GPU-based space-time adaptive processing (STAP) for radar

Space-time adaptive processing (STAP) utilizes a two-dimensional adaptive filter to detect targets within a radar data set with speeds similar to the background clutter. While adaptively optimal solutions exist, they are prohibitively computationally intensive. Thus, researchers have developed alternative algorithms with nearly optimal filtering performance and greatly reduced computational intensity. While such alternatives reduce the computational requirements, the computational burden remains significant and efficient implementations of such algorithms remains an area of active research. This paper focuses on an efficient graphics processor unit (GPU) based implementation of the extended factored algorithm (EFA) using the compute unified device architecture (CUDA) framework provided by NVIDIA.

Adaptive conformal array radar

This paper considers the novel application of space-time adaptive processing (STAP) to conformal array radar. Using numerical simulation, we characterize the performance potential of two candidate conformal array designs: a tapered, belly-mounted canoe and a conformal array taking the shape of a chined radome. We find the nonlinear nature of the conformal array design induces clutter angle-Doppler nonstationarity. This nonstationarity leads to covariance matrix estimation errors and a consequent degradation in STAP performance potential. We find these additional losses reside in the range of 4-10 dB for the two array designs under consideration. Finally, we briefly investigate several ameliorating solutions based on localized processing and time-varying weights, achieving performance gains on the order of several decibels to fully mitigating nonstationary behavior over regions of the detection space.

Radar processing architecture for simultaneous SAR, GMTI, ATR, and tracking

Combined synthetic aperture radar (SAR) and ground moving target indication (GMTI) radar modes simultaneously generate SAR and GMTI products from the same radar data. Furthermore, the SAR and GMTI data products can be further exploited for target signature extractions, automatic target recognition (ATR), and feature-aided tracking. This hybrid mode provides the benefit of fused imaging and moving target displays along with enhanced target recognition and ground target tracking. The Air Force Research Laboratory (AFRL) Gotcha radar has collected wide-bandwidth, multi-channel data that can be utilized for these hybrid mode applications. This paper presents a processing architecture for simultaneous SAR, GMTI, ATR, and tracking, and includes the results of applying this processing to the AFRL Gotcha data.

Sub-band processing for grating lobe disambiguation in sparse arrays

Combined synthetic aperture radar (SAR) and ground moving target indication (GMTI) radar modes simultaneously generate SAR and GMTI products from the same radar data. This hybrid mode provides the benefit of combined imaging and moving target displays as well as improved target recognition. However, the differing system, antenna, and waveform requirements between SAR and GMTI modes make implementing the hybrid mode challenging. The Air Force Research Laboratory (AFRL) Gotcha radar has collected wide-bandwidth, multi-channel data that can be used for both SAR and GMTI applications. The spatial channels on the Gotcha array are sparsely separated, which causes spatial grating lobes during the digital beamforming process. Grating lobes have little impact on SAR, which typically uses a single spatial channel. However, grating lobes have a large impact on GMTI, where spatial channels are used to mitigate clutter and estimate the target angle of arrival (AOA). The AOA ambiguity has a significant impact in the Gotcha data, where detections from the sidelobes and skirts of the mainlobe wrap back into the main scene causing a significant number of false alarms. This paper presents a sub-banding method to disambiguate grating lobes in the GMTI processing. This method divides the wideband SAR data into multiple frequency sub-bands. Since each sub-band has a different center frequency, the grating lobes for each sub-band appear at different angles. The method uses this variation to disambiguate target returns and places them at the correct angle of arrival (AOA). Results are presented using AFRL Gotcha radar data.

Radar modeling and validation of human gaits using simultaneous collections of motion capture and radar data

Previous research in radar modeling of the human gait has implemented motion defined by a kinematic walking model. While this model captures the fundamental walking characteristics, it is limited in practical application due to the generality of its definition. To increase the fidelity and realism of human gait modeling, we utilize motion capture and radar data measurements that are simultaneously collected. We utilize the motion capture data collections to provide accurate body segment locations over time. We incorporate this information into a radar model that calculates the coherent radar return from each segment at the time of each radar pulse. We validate this model using Ku-band radar measurements that were simultaneously collected with the motion capture data. We present and analyze results for walking, jogging, and throwing motions. Comparisons of the simulated and measured results show radar modeling with motion capture data accurately depicts the radar response in both range and Doppler.