A Visible Light Indoor Localization Method Based on Improved Artificial Neural Network

Abstract

The invention discloses a visible light indoor positioning method based on an improved artificial neural network. The steps are as follows: first, a photodetector samples several different known positions, records the real position P and the received signal strength R at the same time, and constructs a training set ; Then, input the training set into the neural network to be trained, and use the proposed loss function to transfer the update coefficients in reverse to complete the network training; finally, sample the RSS at the unknown position to be located, and input the RSS into the trained artificial Neural network to obtain localization results. The invention utilizes a new loss function to overcome the disadvantage that the traditional artificial neural network needs a large amount of training data to ensure the accuracy. On the basis of the loss function, only a few sets of data can be trained to obtain a high-precision artificial neural network, which improves the working efficiency of the visible light positioning system.

Description

technical field [0001] The invention belongs to the technical field of optical communication and optical sensing, in particular to an indoor positioning method of visible light based on an improved artificial neural network. Background technique [0002] Visible light positioning (visible light positioning, referred to as VLP) is a new positioning technology that uses a light source (such as a light-emitting diode LED) to achieve positioning while illuminating. Compared with traditional indoor wireless positioning methods (such as WIFI, Bluetooth, NFC, etc.), VLP technology based on optical communication has outstanding advantages such as high positioning accuracy and anti-electromagnetic interference. The commonly used algorithms for VLC indoor positioning often assume that the light source is a Lambertian radiator, and analyze the characteristics of the received signal (such as intensity, signal phase, etc.) Finally, based on the above relationship and the actual detectio...

AI Technical Summary

Problems solved by technology

[0004] The main purpose of the present invention is to overcome the shortcoming that VLP based on the traditional artificial neural network needs to use a large amount of training data to obtain a high-precision model, and propose a visible light indoor positioning method based on the improved artificial neural network, which has the advantage of using a small amount of training data. The advantages of high-precision positioning can be obtained

Images