Speaker: Prof. Chiung-Shen Johnson Ku (Department of Electrical Engineering, NTUT)<\/p>\n\n\n\n
Time & Venue: 14:20, S101<\/p>\n\n\n\n
Title: Toward Indigenous SAR Capability in Taiwan: Polarimetric Simulation and X-Band UAV SAR Development<\/p>\n\n\n\n
Abstract:<\/p>\n\n\n\n
Frequent cloud cover in Taiwan’s subtropical monsoon climate often limits optical remote sensing, while slope hazards, land subsidence, and maritime surveillance demand all-weather, day-and-night synthetic aperture radar (SAR) observation. SAR is an active microwave imaging system. Its range resolution is governed by signal bandwidth, and its azimuth resolution is determined by the synthetic aperture, independent of range. The retained phase further enables interferometric (InSAR), polarimetric (PolSAR), and tomographic (TomoSAR) applications. However, Taiwan’s first indigenous SAR mission, FORMOSAT-9, comprises two X-band active phased-array satellites (best resolution \u2264 1 m, maximum swath > 50 km), and its first launch is not scheduled until 2028, one year later than previously planned. Until then, domestic users remain reliant on foreign commercial SAR data.<\/p>\n\n\n\n
This talk begins with a review of SAR imaging geometry, resolution, and processing workflows, followed by a systematic overview of Taiwan\u2019s SAR research and development capabilities across industry, government, academia and domestic industry. The assessment indicates that domestic InSAR applications are internationally competitive, but it also identifies three structural gaps: (1) the absence of an indigenous data source before 2028; (2) the concentration of spaceborne payload integration in a single supplier, with transmit\/receive (T\/R) modules and GaN power amplifiers not yet space-qualified domestically; and (3) the lack of a shared validation infrastructure among the hardware, data, and algorithm communities. Accordingly, a three-phase roadmap toward indigenous capability is proposed, with the two FORMOSAT-9 launches as milestones. The first phase emphasizes pooled procurement of commercial data, calibration-site establishment, and airborne pathfinder validation. The second focuses on in-orbit calibration and a second payload supplier. The third targets a small SAR constellation and multi-baseline tomographic missions.<\/p>\n\n\n\n
Our group addresses these gaps from both the simulation and the system perspectives. On the simulation side, a GPU-accelerated, fully polarimetric shooting-and-bouncing-ray\/physical-optics (SBR\/PO) echo simulator has been developed. It models the polarimetric calibration chain using TerraSAR-X orbit and imaging geometry and incorporates a spherical-wave PO kernel for near-field airborne echoes. This enables imaging algorithms, radiometric calibration, and calibrator designs to be verified before measured data become available. On the system side, a dual-mode X-band (9.65 GHz) unmanned aerial vehicle (UAV) SAR has been designed for an operating altitude of 110 m. The High and Low modes use range bandwidths of 424 MHz and 212 MHz and provide ground swaths of approximately 100 m and 495 m, respectively. Under conservative assumptions, including an antenna efficiency of 0.4 and a system loss of 6 dB, the worst-case noise-equivalent sigma zero (NESZ) is \u221234.2 dB, exceeding the \u221230 dB requirement by about 4 dB. In the processing architecture, an RFSoC performs waveform generation, sampling, digital down-conversion, and range compression. An embedded system\u00a0 executes backprojection imaging, motion compensation, and phase gradient autofocus, with an estimated compute margin of 20\u201380\u00d7. Future work will complete hardware integration and conduct flight tests over a corner-reflector calibration site for cross-validation against simulation. Together, the simulator and the UAV SAR provide an airborne pathfinder for FORMOSAT-9 and help close the validation gap in Taiwan’s pursuit of indigenous SAR capability.<\/p>\n\n\n\n
Keywords:<\/strong> synthetic aperture radar (SAR); polarimetric simulation; SBR\/PO; UAV SAR; noise-equivalent sigma zero (NESZ); FORMOSAT-9<\/p>\n\n\n\n