Online Least-Squares Learning for Monopulse Angle Tracking and Waveform Estimation

Monopulse technique is the most widely-adopted method for angle estimation in radar systems, whereas performance improvement is required. To that end, a monopulse angle tracking algorithm named online least-squares learning (OLSL), which can also provide waveform estimation, is proposed in this paper. By establishing a least-squares based joint optimization problem of the target angle and the waveform, OLSL fully exploits the previous data for performance improvement. The estimate is updated in an online manner for acceleration. The memory scheme is further introduced to avoid loss of accuracy in the case of time-varying angles. Compared to conventional mono-pulse estimation, OLSL only requires 4 additional real-number calculations (2 additions and 2 multiplications) and the storage of 2 real-numbers, but enjoys remarkably improved accuracy and robustness against outliers. Moreover, the proposed algorithm can be simply applied to 2D angle estimation, and is compatible with various existing amplitude-comparison monopulse methods. Simulation results verify its effectiveness and superiority.