Research On Reconstruction Methods And Application Of Illumination Field In Buildings Based On Genetic Algorithm And CMAC Neural Network

Wireless sensor network can realize the change of illumination of buildings within the real-time monitoring and illumination field is reconstructed through illumination of collected data which can provide a secondary means of security for the building by analyzing the changes in the building field and the illumination of the incident occurred and the situation judged. In some of the key prevention place or the high level of security requirements region, while through the reconstruction of illumination field can monitor the changes of illumination of these areas and can better judge the protected object in a safe state, illegal intrusion and timely with the monitoring center to form a linkage alarm which play more solid security.Sensor nodes can not be deployed densely for their high costs and we need to reconstruct the building within the field of illumination, this paper studies the method of using CMAC neural network to reconstruct building illumination and optimizes the CMAC neural network’s learning rate based on genetic algorithm for reconstructing illumination in the building. We can get the main points of interest illumination data within the building by limited nodes and firstly use the numerical interpolation methods to reconstruct the illumination field within lower accuracy; and then put forward the illumination data reconstruction method based on CMAC neural network and compare with the numerical interpolation methods which the experimental results show that the CMAC neural network method can get higher accuracy and meet the accuracy requirements set. While using d algorithm to adjust weight value has a certain blindness learning rate in the choice for CMAC neural network implement process.According to the learning rate selection exists uncertainly, this paper proposes the optimization of CMAC neural network’s learning rate based on genetic algorithm to avoid the blindness of choosing learning rate. Because genetic algorithm has the global search and optimize calculation ability based on evolution conception which looked upon as an effective tool to solve complex problems that traditional methods can not work. Through the experimental comparison finds that the use of genetic algorithm optimization of learning rate more accurate than randomly selected in the reconstruction of illumination based on CMAC neural network and achieve the desired target accuracy.Finally, we research on the design and application on the system of reconstruction illumination in building, which is elaborated on the system from outline to structure and function, then to implement the software’s modules. This system can meet the request of design and provide visualization for the security monitoring staff or end-users through reconstruction data and the illumination field of contour maps. While informing the changing illumination status or the information about security has an important research value and significance in practical application.