The Design Of Intelligent Control System In Green Building Based On Machine Learning Theory

With the improvement of society and advancement of science technology, the energy consumption is growing year by year. In the order to achieve the goal of builing sustainable development society, green building which has features of low energy consumption, people-oriented, acclimatization and environmentally friendly becomes quite popular among specialists and ordinary people. In green building, the intelligent control system which combines the methods of control theory, communication technology and architectural design is used to improve the intelligence of the building and monitor the subsystems of home security, indoor environment, appliances control and user location. In order to meet customer demands and reduce learning costs of users, an intelligent control system which is based on the machine learning theory is developed in this dissertation. The main research contents of this dissertation are as follows:(1) The whole framework of the intelligent control system and specific functions of all subsystems are designed. The goals of home security, indoor environment, appliances operating and user location are achieved which make the green building healthy, safe and comfortable for users.(2) In the control process for the environmental factors, the theory of gray correlation is used to select the main factors which influence the choices of satisfaction environmental parameters. These factors are chosen as the inputs of the prediction model which is based on the algorithm of Support Vector Machine (SVM) and K-Nearest Neighbor (KNN). Two different modified optimization methods which are Adaptive Particle Swarm Optimization (APSO) and Adaptive Genetic Algorithm (AGA) are used to optimize the key parameters of the prediction models and the performance of these models are compared.(3) In the scenario of multiple users who control the same system, the system can distinguish between different users by their ID. It provides services according to their daily preferences. In the dissertation, a location engine which is based on the improved trilateral positioning principle is developed and the regional level positioning error range is within 1.5 to 3 meters.(4) In the control of appliances, a mathematical model which is based on the algorithm of Space-Time Customary-Based Reasoning (STCBR) is applied. By observing and recording the time partition and space partition of operating different appliances, the habit body is established on the meet and the join operation of the user behavior sets. The intelligent control of the appliances is based on these habit bodies.(5) To achieve the goals of response quickly and reducing energy consumption, the background operation of the system is an embedded computer and front end is based on an Android application.(6) A video decoder is used to get the monitor video stream which is under MJPEG format and a video capture algorithm is applied to record the motion video. What’s more, the video can also be recorded under the manual mode.The test results show that the intelligent control system which meets the design requirements can predict the environment parameters and provide services of controlling appliances.