SBIR Phase I: Software-Defined Sub-Terahertz Imaging Radar for Algorithmic Agility and All-Weather Transportation Safety

点击翻译按钮获取中文摘要

The broader/commercial impact of this Small Business Innovation Research (SBIR) Phase I project is the development of a universal, affordable, and sustainable sensing solution to enable perimeter security and transportation safety under all weather conditions. Current sensing solutions available today are based on a single modality, expensive to deploy, and not robust to adverse weather conditions. Current solutions also employ proprietary sensor processing interfaces, do not provide the quality of data needed for decision-making by continuous learning, are hard to upgrade, and have poor size, weight, and power specifications. In contrast, the proposed technology leverages the strengths of multiple sensing modalities on a single, converged, open compute platform to enable robust perception in adverse weather conditions while offering significant advantages to the total cost of ownership. The technology has a wide range of applications in sectors as diverse as automotive, robotics, enterprise, aerospace, and defense. The solution developed under this project has the potential to save lives by reducing the number of road accidents, improving the driver reaction time, protecting vulnerable road users such as pedestrians and bicyclists, reducing the downtime for a shipping company, minimizing the costs associated with collision claims and repairs, and detecting, classifying, alerting, and tracking threats on the ground and in the air. This Small Business Innovation Research Phase I project develops a novel, scalable, centralized sensing platform and a multi-spectral sensor prototype to realize ultra-fine resolution in range, Doppler, azimuth, and elevation domains using commercial, off-the-shelf processing elements. Advanced compression algorithms are employed to transport sensor measurements over secure, open, low-cost, and low-latency interfaces to the centralized processing unit to enable multi-modal sensor processing, early sensor fusion, and high-dimensional perception for higher-level decision-making. The de-coupled sensing and processing architecture produces unprecedented access to measurement-level data to enable artificial intelligence and machine learning-based algorithmic discovery. False-alarm-constrained global object detection algorithms are employed to enable localization, navigation, and mapping for operation under adverse weather conditions. Proof-of-concept sensor hardware is developed with laboratory and field experiments to validate the architecture, technology, algorithms, and software. Some of the key technology risks addressed in this proposal are antenna design and fabrication at millimeter frequencies and above, cascading of multiple radio frequency transceivers to realize a large number of spatial channels, and hardware-level synchronization across the sensors.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.