Research On Intelligent Environment Control System Of Livestock And Poultry Breeding Based GRU Neural Network

The environmental control of livestock and poultry houses is a complex multivariable control problem.How to make the environment in the house be able to achieve complete automatic adjustment according to the growth cycle of livestock and poultry is a problem that needs to be solved urgently in the current livestock and poultry breeding industry.In view of the shortcomings of existing environmental control systems in wireless communication,remote upgrade and intelligent control,this thesis studies an intelligent environmental control system for livestock and poultry breeding based on GRU neural network,and designs and develops a set of environmental parameter acquisition nodes and controller equipment based on STM32.In this thesis,an intelligent environmental control strategy based on a GRU model is proposed to realize a complete intelligent environmental control system for livestock and poultry breeding.The main work of this thesis includes:（1）Software and hardware design of terminal equipment.To address the challenges of remote locations,large monitoring areas,complex environments,and significant signal interference in large-scale breeding houses,a terminal equipment solution was developed.The solution is based on the high-performance STM32 MCU and is designed to provide high stability,robust anti-interference capabilities,simple deployment,and support for multiple communication modes.In view of the shortcomings of the existing data acquisition nodes in the industry that are mostly wired transmission,inconvenient installation and high maintenance costs,an ad hoc network mode based on Zig Bee acquisition nodes and controllers is designed to solve the existing problem of complex wiring.Aiming at the problem of unstable data transmission caused by remote location,a dual network communication mode supporting wifi and 4G is designed,and the communication protocol is flexibly adjusted according to the network quality to achieve stable interaction between the cloud platform and the controller terminal equipment.Aiming at the unstable problem of OTA（Over The Air）remote upgrade,the congestion control algorithm is applied to the upgrade process to realize the dynamic adjustment of the upgrade packet bit rate.The experimental results show that the controller terminal device can run continuously for at least 7 months,and can interact with the cloud platform stably to achieve stable remote upgrade.（2）Intelligent environmental control strategy based on GRU model.Aiming at the defect that the traditional environmental control strategy based on threshold can not effectively avoid the abnormal environment,an intelligent environmental control strategy based on GRU neural network is proposed.In order to improve the accuracy and convergence speed of the GRU model,the isolated forest algorithm,moving average method and max-min standardization method are used to preprocess the data set.The experimental results show that the accuracy of the isolated forest algorithm is up to 99%when the percentage of outliers is less than 5%.In order to accurately adjust the output power of the equipment,the change curve of each environmental parameter is drawn based on the prediction results of the GRU model,and then the slope of the curve is used to map the intensity of the change of each environmental factor,and then the minimum output power of each equipment is derived based on the relevant parameters of the breeding house and environmental control equipment（such as the volume of the breeding house,the effective power of the fan,the cooling efficiency of the wet curtain,etc.）.Experimental results show that compared with the threshold-based environmental control strategy,the environmental control strategy in this thesis greatly reduces the abnormal environmental situation.（3）An intelligent environmental control system for livestock and poultry breeding has been realized.The system is divided into two parts:cloud platform and terminal equipment,the terminal equipment includes a plurality of environmental parameter acquisition nodes and a controller,the acquisition node collects temperature,humidity,light,CO₂and other special gas data through the sensor,and interacts with the controller through Zig Bee;The controller uploads environmental data through 4G or Wi Fi,and realizes the control of the device according to the control strategy;The cloud platform adopts the technology of separating the front and back end,the front end adopts the VUE framework,and the back end adopts the Go-Zero microservice framework.The cloud platform is mainly responsible for storing sensing data,training GRU models,formulating environmental control strategies,and realizing fully automatic adjustment of environmental parameters according to the growth cycle of livestock and poultry through the interaction between the cloud platform and the controller.At present,the system has been running smoothly for 7 months since its deployment,and the expected results have been achieved in terms of data collection and environmental regulation.