CCSS: Physics-informed Radar Backscatter Modeling for Through-Canopy Forest Remote Sensing Using RFID Technology

Understanding the forest understory is essential for effective forest management, biodiversity conservation, wildfire prevention, and environmental monitoring. However, traditional satellite and aerial remote sensing technologies are fundamentally limited in their ability to observe vegetation beneath the forest canopy due to occlusion and signal attenuation. As a result, critical indicators of ecosystem health, such as aboveground biomass carbon pools, combustible understory fuel loads, and other measures of biodiversity in the understory, remain largely unmeasured at scale. This project addresses this critical gap by envisioning a low-cost, scalable sensing system that uses radar and wireless communication technologies to detect and characterize the forest understory, complementing existing orbital and suborbital remote sensing systems. The project advances radar sensing and physics-aware modeling by demonstrating an IoT-powered forest observatory in real-world scenarios, and creates new educational and outreach materials, including open-source software and student training modules, to broaden the impact and accessibility of the research. The project outlines a new approach that bridges radar sensing and backscatter communication for environmental sensing via a novel channel modeling approach through vegetation and generalizable physics-aware models that characterize the effect of biomass on radar signals. The key intuition is that vegetative dielectric and moisture content will alter the RF signatures in both frequency and time domains as they penetrate through the vegetation medium. These signatures can be learned using complex physics-aware models as long as the RF reflections that carry this signature can be reliably separated, modeled, and interpreted. The project team will realize this vision through four inter-connected tasks: (i) formalizing the radar backscatter through multi-layer forests, which results in a radar forest synthesizer (ii) proposing physics-aware radar backscatter models that offer spatial characterization and mapping of forest understory biomass (iii) introducing an RF-coded tag design that can pair up with off-the-shelf radar platforms and serve as ground references for accurate understory characterization; (iv) Fully evaluating the end to end system in wide-area testbeds and demonstrating the accuracy of characterizing forest understory. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.