Black Flight Identification Using Radar Cross Section (RCS), Speed, and Altitude from RADAR Data using Supervised Machine Learning

Black flight or unknown and unidentified air objects entering sovereignty airspace can be hostile and endangering the nation's Center of Gravity (CoG), so they must be tackled quickly and precisely by the National Air Operation Command (NAOC). Intercepting a hostile object needs an accurate estimation of its type and characteristics to minimize the risks that might be experienced during the execution of an air defense operation. However, such an object will cover its identity as much as possible by turning off its Identification Friend or Foe (IFF) transponder. Therefore, the only way to recognize and identify such an object is by using its characteristics extracted from the Radar system, such as Speed, Altitude, and Radar Cross Section (RCS) displayed on the Position Plan Indicator (PPI) console. We did experiments using several Machine Learning techniques to carry out object recognition and identification of several aircraft. From the experimental results on a self-made dataset consisting of five aircraft types, namely F-16C, Su-30MKK/MK2, C-130 Hercules, Boeing 737–200, and Boeing 737–900, with 100 data samples, the RandomForest classifier achieved the highest identification performance with the accuracy of 82
^%
for the aircraft type and 100% for the aircraft role.