基于四象限光电探测的光导引技术研究  

Research on optical guidance technology using four-quadrant photodetectors

作　　者：卢帅华 钟志[1] 杜春燕 单明广[1] 于蕾[1] LU Shuaihua;ZHONG Zhi;DU Chunyan;SHAN Mingguang;YU Lei(College of Information and Communication,Harbin Engineering University,Harbin 150001,China;Beijing Institute of Computer Applications and Simulation Technology,Beijing 100000,China)

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001 [2]北京计算机技术及应用研究所,北京100000

出　　处：《实验技术与管理》2025年第1期45-51,共7页Experimental Technology and Management

基　　金：国家重点研发计划项目(2022YFC3103401);2021年度黑龙江省高等教育教学改革研究项目(SJGY20210167)。

摘　　要：水下航行器中相关导引回收技术不断发展,为满足水下高精度的导引需求,设计了一套基于四象限光电探测器的无人水下航行器(UUV)回收可见光导引装置。通过建立“两灯两探”的光导引模型,采集大量实验数据,并结合深度学习技术搭建精准模型,用于计算水平面上的相对间距、偏移量和偏转角。在发射端,采用水平放置的两颗LED光源,接收端利用四象限光电二极管(QP)收集灯源信息,对灯源信息进行电路处理并采集形成有效数据。随后将数据输入预先训练完成的模型,以实时获取位置和姿态数据。经过实验测试,水平距离测量误差在2%以内,左右偏移误差在2 cm以内,角度测量误差控制在2°以内。[Objective]With the continuous advancement of underwater vehicle guidance and recovery technologies,acoustic and visual systems have proven effective for both long-distance and short-distance navigation.However,significant limitations persist in mid-range guidance.To address these blind spots in underwater mid-range guidance,this paper proposes a visible light guidance device based on a quadrant photodetector.[Methods]By leveraging the advantages of underwater optical technology,such as high precision,strong real-time performance,and low power consumption,the device provides an efficient and reliable technical solution for the mid-range recovery of unmanned underwater vehicles.A system prototype was implemented for horizontal guidance,with the same principles applicable for vertical guidance.Future work aims to integrate spatial theoretical models to address the challenge of accurately determining the position and orientation of underwater vehicles after extended missions as they approach a docking station.This paper begins by introducing the fundamental principle of using visible light and quadrant photodetectors to measure position and attitude data.In a simple setup with a single light source and detector,small changes in position or angle become difficult to detect over longer distances.To overcome this limitation,a“two-light,two-detection”model is proposed,utilizing two sets of LEDs(light emitting diodes)as light sources,with double-point analysis achieved through time-division illumination.The receiving end consists of two sets of quadrant photodetectors,enabling synchronized data collection from each light source.This model forms a network on the horizontal plane,facilitating effective measurement of relative position and attitude data.Extensive experimental data was collected using the prototype system,and a neural network model was developed by applying deep learning techniques.Finally,the two quadrant photodetectors on the horizontal plane simultaneously captured weak light signals instantaneously.T

关 键 词：水下光导引 四象限光电探测 FPGA 深度学习 

分 类 号：TN29[电子电信—物理电子学]