Research On Intelligent Greenhouse System Based On Neural Network Control Strategy

With the development of artificial intelligence technology,artificial intelligence algorithms show unique advantages in implementing adaptive control of complex systems.The country develops"smart agriculture"in order to achieve the goal of developing a strong agricultural country.Using artificial intelligence algorithm to realize adaptive control of agricultural greenhouse system is an important embodiment of"smart agriculture".This study aimed at the climate characteristics of shanxi province,studied the needs of greenhouse crop growth,designed a greenhouse control system integrating greenhouse environment collection,intelligent decision and making control.The research mainly includes the following three parts:1)decision algorithm research;2)control decisions design;3)the hardware and software of system design.In the aspects of decision algorithms research,taking into account the characteristics of nonlinear and time-varying in greenhouse system,this paper analyzes the fully online sequential extreme learning machine(FOS-ELM)algorithm which has the advantages of fast training speed,small initialization influence and strong generalization ability,combining the whole online learning ability of the algorithm,the system model would be predicted precisely and work well.The FOS-ELM controller usually ignores the impact of external disturbances on the controller's decision in practical applications.Based on the expected system out to decide the system output has not met the requirements of system control.Therefore,an incremental compensation FOS-ELM algorithm was proposed and the corresponding controller was designed.Considering the effectiveness of fuzzy control algorithms in the control of greenhouse systems,the corresponding fuzzy controller is designed.In the aspects of control decisions design,the incremental compensation FOS-ELM algorithm designs the main controller of the greenhouse system,and the fuzzy control algorithm designs auxiliary controller of the greenhouse system.Auxiliary controllers provide samples for online training of the main controller.The greenhouse system control mainly by the main controller after the master controller reaches control requirements by training in order to guarantee the control effectiveness and adaptability of system.In terms of system software and hardware implementation,the system is mainly composed of three parts:the management system on the local computer,the single chip of STM32F407 is taken as the greenhouse controller and the sensors and actuators of the greenhouse system.The upper computer mainly completes the function of human-computer interaction,it includes the data and status display,basic parameter settings,data processing and decision-making,and the configuration of the communication part of the greenhouse sensors and actuators.The decision module uses the incremental compensation FOS-ELM controller and the fuzzy controller.The STM32F407 is mainly used for the real-time environmental parameter display of the greenhouse system,the actuators control and the function of over-limit alarm.The results show that the controller designed by incremental compensation FOS-ELM algorithm can meet the requirements of system control when the temperature and humidity control error is±2?and±3%,the incremental compensation FOS-ELM algorithm has characteristics of simple design,fast learning,and strong generalization.It demonstrates the effectiveness and feasibility of this algorithm in practical applications.Combined with the human-computer interaction interface and the stable hardware control unit,the system can meet the needs of greenhouse control and achieve the purpose of self-adaptive greenhouse control.It reduces the reliance on expert systems,improves the stability of the system,and has important significance for the promotion of greenhouse adaptive control systems.