Doppler frequency shift and angle of arrival measurement system with machine learning assistance

A microwave photonic Doppler frequency shift (DFS) and angle of arrival (AOA) measurement system based on Mach–Zehnder modulators (MZM) and optical frequency shift modulator (OFS) is proposed and demonstrated. The dual-branch architecture of the system can effectively avoid the frequency shift error introduced by the OFS, thereby improving the DFS measurement accuracy and distinguishing the DFS direction. Furthermore, this scheme constructs multiple amplitude comparison function (ACF) curves by introducing multiple phase shifts to the echo signal, thereby extending the measurement range of AOA and reducing the power dependence effect. Simulation results show that the DFS measurement error is less than 0.1Hz within 5∼15GHz. After introducing machine learning, the AOA prediction error is stabilized within ±0.15° within −70.8°∼70.8°.