Implementation Issues of Automatic Modulation Classification on Software Defined Radio Platform

摘要

Automatic modulation classification (AMC) is used in many military and civilian applications such as surveillance, spectrum monitoring, security, emitter identification, cognitive radio etc. Typically the receivers for the AMC is blind in nature i.e there is no pilot channel or known training sequence between the transmitters and receivers. This requires some non-data aided (NDA) signal processing in the receiver side. The detection algorithm becomes more complex in an environment where emitters use both analog and digital modulation schemes. AMC unit must estimate the carrier frequency, bandwidth, symbol rate etc. for accurate detection of the modulation types. In this demonstration, a feature based AMC algorithm will be introduced in the software defined radio (SDR) platform by using universal software peripheral radio (USRP), GNU radio and some open-source software.