Recursive Deconvolution Algorithm Based on Least Mean Square Error

This paper presents a recursive deconvolution algorithm to address the problem of low computing efficiency in previous deconvolution algorithms. This study implements a recursive algorithm that may decompose the deconvolution process into signal sampling time intervals based on the matrix inverse lemma. This approach runs in parallel with the signal sampling process, lowering computing complexity and enhancing the efficiency of deconvolution computation while also laying the theoretical groundwork for real-time deconvolution computation. Next, the algorithm can be applied to the deconvolution solution of unknown signals with fixed-length variables and variant-length variables. Furthermore, this approach decomposes its computational complexity into each recursive link, which makes it possible for high-complexity deconvolution calculation to be applied on low-end chips. Finally, the computational complexity of the proposed method is analyzed theoretically, and the simu-lation results illustrate the calculation error and noise immunity performance of the algorithm.