Research On Key Technologies Of Intelligent Aquatic Oxygen System Based On Multi-Agent

In recent year, with the improvement of daily life, the nutrition requirements forfood are constantly improving. People like aquatic products, especially fish and prawn.The demand and scale freshwater aquaculture of Vannamei increased, because ofVannamei as the main product. Dissolved oxygen as a key factor affecting freshwateraquaculture production, the intelligent control of dissolved oxygen becomesincreasingly important. Currently, the forecasting work of dissolved oxygen has froma single control integrated into the multi-factor.In order to achieve the real-time control of dissolved oxygen, this paper takes theVannamei as research object, has accomplished single loop control model design andmulti-loop control linkage model. This auricle put multi-agent technique into bothcontrol models, the design of agent reasoning module for multi-agent and the internalstructure for each agent, which complete the intelligent control process of Vannameifarming.To the small ponds, which only need one aerator to complete the control ofdissolved oxygen, this monitoring environment composed by a dissolved oxygenprobe with own collection of DO and one aerator in the small ponds. The other waterquality parameters (PH, NH, ORP and temperature) collect by the addition watercollection node. This paper establishes a fuzzy control algorithm under the influenceof multiple factors by analysis the relationship between various factors. Due to theshortcoming that the expert system has low ability of processing distributedknowledge, unable to apply in environment, it did not have social features. so thisarticle put multi-agent technique into expert system, do use Do predictive Agent,control Agent of aerator and decision-making Agent in the oxygen control process. Doforecast Agent call fuzzy control algorithm to get the predicted value which influenceeach other under multiple factors. Decision Agent select the high priority controlcommands sent to the aerator control Agent after comparing the single DO controlvalue and the DO predictive value. Aerator control Agent completed aerator control according to the control command. This system have the remarkable advantage inreal-time aspects, energy consumption and the accuracy of control than the single DOcontrol system.Now with the increasing intensification of farms, lack of dissolved oxygen inlarge ponds easily and single aerator cannot meet the entire demand for dissolvedoxygen in ponds. This article proposed circulating aerator system based onmulti-agent, which uses the autonomy, interaction and communication collaborationfeatures to create a multi-intelligent integration control model aerator. In the linkagecontrol model, buoy system used to collect the water quality parameters (DO, ORP,NH, PH, temperature), and the same time use fuzzy control algorithm to predictive theDO predictive value. In this model use multi-agent technique too, do use Dopredictive Agent, control Agent of aerator and decision-making Agent. DO acquisitionnodes sent the water quality parameters to DO forecast Agent and DO forecast Agentcall fuzzy control algorithm to get the predicted value which influence each otherunder multiple factors. Compare the forecast value and the actual acquisition value,When the difference between the two numbers is too large, DO forecast Agent triggerdecision support Agent, and decision support Agent send the aerator controlcommands to Aerator control Agent. According to this process, complete the aeratorcontrol. When the water quality parameters are not abnormal, buoys support decisionAgent send the data to the database in the terminal within a predetermined period,which is facilitate data offline analysis and processing. This model is designed tocollect data upload by underlying buoy system, terminal control system send controlcommands to the bottom aerator through the analysis of water quality parametersthroughout the breeding environment. Among them, terminal control system used arouting algorithm which is based on multi-hop of ZigBee cluster tree structure, toselect a small, short latency communication link to complete the control command.This system achieves a multi-aerator linkage control.Finally, the article introduces the development technology, analysis the realizationof Agents in small ponds, compared the switch number of aerator, the time of switch,and the dissolved oxygen content. This article summarizes the research of this fieldand proposes the further improvement of the system.