| With the increase of people’s demand for location information,location technology has been paid attention to in various fields such as military,emergency rescue and vehicle tracking.At present,the mainstream pedestrian Positioning systems include Global Positioning System(GPS),Bei Dou Navigation Satellite System(BDS),etc.,which can realize high-precision positioning in outdoor environment.Is widely used in aircraft,vehicles and portable devices to provide real-time positioning and navigation.Indoors,the system suffers from signal blocking and multipath effect,which seriously degrades its positioning performance.In fact,indoor applications require more accurate positioning than outdoor applications.At present,domestic and foreign scholars have proposed a variety of indoor positioning technologies,which are either too complex and expensive,such as infrared and ultra-wideband technologies,or susceptible to multipath effects and signal interference,such as Wi-Fi and radio frequency identification(RFID)technologies.The emergence of Visible Light Communication(VLC)technology breaks the limitations of traditional indoor positioning technology.It has the characteristics of economy,environmental protection,anti-electromagnetic interference,high positioning accuracy,strong security and so on,becoming one of the most potential indoor positioning technologies.Therefore,the research direction of this paper is visible light indoor positioning technology,and the main work contents include:Ⅰ.The basic principle of visible light communication is studied.It mainly includes the construction of LED light source and the luminescence principle of LED.On this basis,the visual-distance propagation and non-visual-distance propagation of visible light communication are compared and analyzed in detail.The research shows that the visualdistance propagation of visible light accounts for about 95% of visible light transmission,so this paper mainly studies the visual-distance propagation.Next,the light distribution of different light source arrangements in a 6 m×6 m×3 m room is analyzed.Based on the consideration of light intensity and cost,it is finally determined that this paper studies the indoor positioning of visible light based on the distribution of four light sources.Secondly,the visible light geometric ranging positioning technology and fingerprint positioning technology are analyzed.Geometric ranging and positioning technologies include TOA(Time of Arrival),TDOA(Time Difference of Arrival),AOA(Angle of Arrival)and RSSI(Received Signal Strength Indication)technology,as well as their pros and cons and scenarios.Finally,this paper introduces the common performance evaluation indexes such as root mean square error,average positioning error and error accumulation distribution function,and discusses their significance and application range.Ⅱ.In order to improve the accuracy of visible light geometric ranging and positioning technology,The paper proposes a Seagull Optimization Algorithm combining Golden Sine and Sigmoid Continuity,(Seagull Optimization Algorithm combining Golden Sine and Sigmoid Continuity,GSCSOA)received signal strength of visible light indoor positioning technology.The traditional Seagull Optimization Algorithm(SOA)has slow convergence speed and low optimization accuracy.The Sigmoid function is introduced in the seagull search process to make the algorithm converge quickly in the later period,and the golden sine mechanism is added in the seagull hunt process to improve the optimization ability of the algorithm.Four LED lights were uniformly arranged on the roof of a 6 m×6 m×3 m room.The experimental results show that the average positioning accuracy obtained by the traditional RSSI positioning estimation algorithm is 1.28 m,and the average positioning accuracy obtained by the improved gull optimization algorithm combined with the positioning algorithm of RSSI is 7.17 cm.Obviously,the improved indoor positioning algorithm has higher accuracy and can be used in most indoor positioning places.Ⅲ.In view of the difficulty,high cost and high complexity of geometric measurement,this paper proposes the Extreme Learning Machine Optimized by Seagull Optimizer Algorithm,Visible light fingerprint localization technology of SOA-ELM.The traditional Extreme Learning Machine(ELM)algorithm randomly generates input weight matrix,which is prone to problems such as redundant or insufficient nodes of hidden layer and weak recognition ability of position input parameters.Therefore,this paper proposes the Seagull optimization algorithm.Four LED lights are uniformly arranged on the roof of a 6 m×6 m×3 m room.The experiment shows that the average positioning error of the positioning algorithm in this paper is 4.5cm,which is 40% higher than that of ELM neural network and 59% higher than that of BP neural network.Obviously,the positioning accuracy of the improved algorithm is higher.Ⅳ.In order to verify the feasibility of fingerprint visible light indoor positioning algorithm based on SOA-ELM algorithm in the actual environment,this paper designed the visible light indoor positioning system,including the design of transmitter and receiver,as well as the positioning experiment in the established visible light indoor positioning environment.In the positioning environment of 4 m×4 m×2.9 m,the SOAELM algorithm is used for experiments,and the average positioning error of 5.73 cm is obtained,which proves the effectiveness of SOA-ELM algorithm in visible light positioning.Figure [30] Table [5] Reference [116]... |
PDF Full Text Request