Low-SWaP EO sensor-based Robust Aircraft Detection For DAA Technology in UAS Integration

摘要

Detect-and-Avoid (DAA) technology is the key technique for UAS integration in the National Airspace System (NAS). To substitute radar for DAA purpose, Electro-Optical (EO) sensor-based aircraft detection method is proposed in this paper. The typical Low-Size, Weight, and Power (Low-SWaP) sensor, EO sensor, has many limitations but the proposed detection method helps robust detection of other aircraft. The main problem of EO sensor-based aircraft is short detection range and weather dependencies. However, the key problem of EO sensor has overcome to find moving obstacles not only for other aircraft because aircraft in the distance do not have aircraft shape and it looks small dot. Therefore, the main strategy of robust aircraft detection is detecting moving something using an optical-flow method as region-of-interest (ROI) for other aircraft. The suggested ROI is a big hint having high probabilities of aircraft detection. Deep learning-based aircraft detection could be implemented, at this step, to identify whether aircraft or not in the magnified ROI image. In this research, light-sport categorized aircraft, CTLS and Cessna, are used to get image data and flight information. Installation of onboard EO sensors obey the rule to cover ±110 degrees following Detect-and-Avoid MOPS. In the conclusion, a robust aircraft detection method using EO sensors is proposed for safe UAS operation and tested in real flight test using CTLS and Cessna.