Real-time Signal Processing of UWB Fuze Based On 1D-CNN Model

Abstract

The main source of the burst-height error of UWB fuze is the random error related to the fuze signal processing method. Therefore, the use of reliable algorithms to reduce the error as much as possible to avoid bursting early and misfire is very important to improve the burst-height precision of fuze. In this paper, taking the output of the fuze sampling and integrating circuit as the research object, a real-time signal processing method based on the 1D-CNN model is proposed. Simulation results show that the target recognition rate of the 1D-CNN algorithm is as high as 100%, with no bursting early or misfires, and the mean absolute time error is as low as 0.0202s. The burst-height error is effectively reduced.