Adaptive intelligent street light edge computing model based on CS-ECA-BILSTM and KubeEdge
An adaptive intelligent street light edge computing model based on efficient channel atten-tion network(ECA)and bidirectional long short-term memory neural network(BILSTM)is proposed.Firstly,on the basis of BILSTM,CS-ECA-BILSTM visibility prediction model is constructed by combining Cuckoo Search algorithm and channel attention mechanism to realize road visibility predic-tion.Secondly,illuminance and color temperature factors are introduced to the problem of single control factors of ordinary street lights to improve control efficiency and reduce energy consumption of street lights.Finally,the cloud native concept is introduced at the edge end,and the street light control model is deployed to the edge end using KubeEdge framework and container technology,so as to achieve multi-street light control.The experimental results show that the performance of the CS-ECA-BILSTM model is superior to other comparison models,which can effectively improve the energy efficiency of street lamps,and can realize the operation at the edge.
smart street lightedge computingbidirectional long short-term memory neural networkattention mechanismcontainer technology
万方数据
维普
