Research on FPGA Pulse Laser Ranging Method Based on Deep Learning

Xiaobin Xu,Yi Chen,Kaiyuan Zhu,Jian Yang,Zhiying Tan,Minzhou Luo

DOI: https://doi.org/10.1109/tim.2021.3096281

IF: 5.6

2021-01-01

IEEE Transactions on Instrumentation and Measurement

Abstract:To reduce the ranging error, a field programmable gate array (FPGA) pulse laser ranging method based on deep learning is proposed. By simulating the echo waveforms, the deep learning sample data are constructed to train the ranging convolutional neural networks (CNNs), and the influences of different convolution kernels numbers and noise levels on the performance of the ranging neural network are analyzed. The ranging accuracy and stability of the deep learning pulse laser ranging method and the traditional pulse laser ranging method are simulated and discussed. The FPGA transplantation of ranging CNN with limited resources is realized by three modules of preprocessing, ranging CNN, and distance calculation. The experimental platform has been built to collect echo data of different distances, feed the echo data to FPGA, and use the deep learning ranging method to perform the waveform range calculation. The simulation and experimental results show that the deep learning pulse laser ranging method has higher ranging accuracy and stability than traditional methods. The ranging method has been successfully implemented on FPGA, which provides the possibility for the engineering implementation of the deep learning ranging method in the future.