Research On Visible Light Indoor Positioning Algorithm Based On LED

As demand for location-based services grows,indoor positioning has become a hot research area.Compared with the traditional wireless positioning technology,indoor positioning technology based on visible light communication has been extensively studied due to its advantages such as simple equipment,high positioning accuracy,no electromagnetic radiation,and high safety.This paper models and simulates the indoor visible channel,and studies the visible light indoor localization algorithm based on the received signal strength.Aiming at the problem that the positioning error of the boundary region is larger than that of the internal region in the diffuse reflection optical channel,a sub-regional localization algorithm based on multilayer extreme learning machine is proposed.By establishing multilayer extreme learning machine,according to the magnitude and distribution characteristics of the positioning error of the whole experimental region,the whole region is divided into boundary region and internal region,and the boundary region with larger error is separately trained.The position coordinates of the receiver of the boundary region are updated into the overall position coordinates to achieve the positioning.It effectively solves the problem of larger positioning error of the boundary region in indoor environment and improves the overall positioning accuracy.In order to further optimize the localization performance,a sub-regional localization algorithm based on series particle swarm optimization extreme learning machine is proposed to solve the problem of randomly generating input weight matrix and threshold matrix between the input layer and hidden layer by the extreme learning machine.The first layer regression model,the second layer classification model and the third layer regression model of particle swarm optimization extreme learning machine are established to determine the receiver position coordinates of the whole region,divide the boundary region and the internal region,and determine the receiver position coordinates of the boundary region,respectively.The position coordinates of the boundary receiver are updated to the overall position coordinates to achieve the positioning.The input parameters of the extreme learning machine are optimized,and the problem of large positioning error in the boundary region is solved.Through comparison experiments,the positioning performance,effectiveness,time complexity and robustness of the proposed algorithm are verified,and the positioning error in the boundary region is reduced,which can meet the needs of indoor positioning.