Solar Intelligent Street Lamp System With Visibility Prediction And Self-learning Function

Street lamps is an indispensable part of people’s life.With the increasing popularity of new energy,the utilization of solar street lamps has become an inevitable trend.However,due to the limited energy storage of solar street lamps,energy saving control is needed.At the same time,considering the influence of different lighting colors on people’s comfort level,it is also significant to add lighting color selection mode based on the current solar street lamp system.Therefore,the intelligent design of solar street lamps based on visibility has obvious practical value.Based on the data of Xi ’an Meteorological Bureau,this paper conducts the following research on the solar intelligent street lamp system with visibility prediction and self-learning functions:1)Based on the data of Xi ’an Meteorological Observatory and pollutant monitoring station in 2018,this paper analyzes the factors that affect visibility and pulls out the four factors that are relatively correlated.PSO-BP and LSTM methods are used to establish the corresponding real-time visibility estimation an d prediction model.Simulation results show that the modeling method based on LSTM has higher accuracy and better prediction performance.2)In view of the complex and changeable atmospheric environment,in order to further enhance the performance of real-time visibility estimation and prediction,this paper introduces an adaptive enhancement strategy,proposes a model based on Adaboost-LSTM,and periodically updates the weight parameters of the model.The simulation results show that the performance of real-time estimation and prediction based on Adaboost-LSTM is better than that of LSTM model,and it is more suitable for the prediction of the visibility.At the same time,in order to have the self-learning function of the model,the weight coefficient of each weak predictor is regularly updated by means of sliding window to further improve the model accuracy.3)A smart solar street lamp system with visibility prediction and self-learning functions is designed in this paper.Firstly,design the network structure of intelligent solar street lamp system.Secondly,the color,power saving mode and power saving strategy of street lamps are controlled by the output value of Adaboost-LSTM visibility model.Finally,the corresponding hardware system and visibility monitoring software system are designed,and the effectiveness of the system is verified by experiments.The intelligent solar street lamp system with visibility prediction and self-learning function designed in this paper can implement different lighting modes according to the current visibility and future visibility,which not only extends the lighting time of street lamps but also increases the intelligence of solar street lamps.