An improved algorithm for LFM signal frequency modulation slope estimation

The Linear Frequency Modulation (LFM) signal is large time-bandwidth product signal. It is usually used in Low Probability of Intercept (LPI) radar. So it is important to study the parameter estimation of LFM signal. This paper introduces an improved algorithm for LFM signal frequency modulation slope estimation based on Cubic Phase Function (CPF). CPF can form peak through the nonlinear transform of signal in time domain. The LFM signal frequency modulation slope can be estimated through the peak search in time domain. So CPF can be used to estimate the LFM signal frequency modulation slope. We can choose step length according to estimation precision. So if we choose small step length, CPF can have high estimation precision and the algorithm provides better estimation performance in lower SNR condition compared with High-Order Ambiguity Function (HAF). If we don't have any priori information, at the same time we need very high estimation precision, CPF has to search in large range and choose very small step length. So the calculation quantity is too large to apply to practical engineering. Based on it, an improved algorithm for LFM signal frequency modulation slope estimation is put forward. So this paper focuses on reducing the calculation quantity of CPF to apply it to practical engineering which uses dichotomy for LFM peak searching in time domain to reduce calculation quantity. And the algorithm uses smaller step length which can let the estimation precision of LFM signal frequency modulation slope estimation be higher. Besides the algorithm has better estimation performance in lower SNR condition compared with CPF. Simulations are provided to verify the theoretical claims.