An Application Research Of Multi-Sensor Data Fusion In Intelligent Control Of Greenhouse

Multi-sensor data fusion is an emerging technology applied to Department of Defense (DoD) areas such as automated target recognition, battlefield surveillance, and guidance and control of autonomous vehicles, and to non-DoD applications such as monitoring of complex machinery, medical diagnosis, smart buildings and remote system (RS). Techniques for multi-sensor data fusion are drawn from a wide range of areas including artificial intelligence, pattern recognition, statistical estimation, and other areas. This paper mainly discusses the architecture and algorithm of multi-sensor data fusion, presents multi-sensor data system two-level fusion model and analyses virtues and defaults of all kinds of algorithms, emphasized on the basic connotation of D-S theory and Expert System. Whereas D-S theory and Expert System both have its defaults, a hybrid algorithm based on D-S Theory and Expert Systemwas presented in this paper and this algorithm will be applied to intelligent control of greenhouse.The existence of computer control systems for environmental management in greenhouses provides the opportunity for better productivity and energy saving. The choice of a proper control algorithm and strategy adjusted to local conditions and requirements is also a matter of great importance. The aim of greenhouse environment control is to overcome partly or totally restricts of outdoor climate environments and impersonal conditions according to physiological characters of all kinds of plants themselves, thus we can create optimal growth conditions of plants. Main environment condition of plant growth includes: temperature humidity light intensity and so on. All of them are interacted and connected and plant-growth depends on their synthesis action. At present, most of greenhouse environment control system uses Industrial Personal Computer (IPC) which more emphasizes on computer technology in its development. Because no agriculture expert participates in it, this kind of system lacks expert system which includes physiological information about all kinds of plants and limits its application.This paper analyses the basic Principle of the combination ofevidences and the D-S combination formula, proposes a method to solve the invalidation of the implement of D-S evidential reasoning theory, defines an amendatory combination rule. The construction of the basic probability assignment function value in the D-S model is given to fuse the information from different sensors which use to collect parameters of greenhouse environment. Fusion results will be stored in knowledge base of expert system. Expert System is used to make final decision according to knowledge base which includes experience of expert and fusion results. This hybrid algorithm is used to recognize the decision information of the greenhouse environment monitor system. Through collecting real-time information of the greenhouse environment parameters and fusing data, we can get precise results to regulate all kinds of parameters of greenhouse in time. As a result, we can improve decision-making veracity of greenhouse parameter-regulate, control the environment of greenhouse effectively.