基于注意力机制的室内可见光定位算法  

Design of Indoor Visible Light Positioning Algorithm Based on Attention Mechanism

作　　者：任进[1] 刘跃奇 于淼 REN Jin;LIU Yueqi;YU miao(School of Information Science and Technology,North China University of Technology,Beijing 100144,China)

机构地区：[1]北方工业大学信息学院,北京100144

出　　处：《无线电工程》2023年第7期1612-1618,共7页Radio Engineering

基　　金：北京市优秀人才培养资助青年骨干个人项目(401053712002);2022年北京市大学生创新创业训练计划项目(22XN238)。

摘　　要：由于室内环境错综复杂、诸多障碍遮挡等因素的影响,导致无法保障高稳定性的室内定位需求。针对该问题,提出了注意力机制与神经网络融合的室内可见光定位算法,实现在视距(LOS)链路场景下的高精度室内定位系统。建立室内可见光信道模型参数后,将携带位置信息的光强信息输入至神经网络中进行训练。在神经网络算法方面,进行反向传播算法、极限学习机算法与注意力机制算法结合,提高定位精度。实验结果表明,添加注意力机制的算法预测准确率与基础算法相比有明显提升。Due to the complex indoor environment and the influence of many obstacles and other issues,it is impossible to guarantee the high stability demand for indoor positioning.The indoor visible light positioning algorithm based on the fusion of attention mechanism and neural network is proposed to achieve a high-precision indoor positioning system in Line of Sight(LOS)scenarios.After the indoor visible light channel model parameters are established,the light intensity information carrying the location information is input to the neural network for training.Meanwhile,in terms of neural network algorithms,back propagation algorithms,extreme learning machine algorithms and attention mechanism algorithms are combined to improve the positioning accuracy.The experimental results show that the accuracy of the algorithms that adds attention mechanism are significantly improved as compared with basic algorithms.

关 键 词：可见光定位 室内定位 反向传播 极限学习机 注意力机制 

分 类 号：TN929.1[电子电信—通信与信息系统]