Research On The Self-adaptive Control Mechanism Of Nutrient Solution For Soilless Cultivation

Soilless Cultivation, as an important development direction of modern agriculture, is an agricultural mode that utilizes man-made root environment to substitute the soil environment. It has been widely applied in developed countries and regions such as Europe, America, Japan, and Israel, and demonstrates superior advantages and great potential. Soilless Cultivation could be an essential technique to raise the agricultural productivity and alleviate the problems such as land shortage, and would play a significant role for the economic & social development.In soilless cultivation, the control of vegetable-growth environment is an important tool to enhance productivity-efficiency and reduce production costs. The ideal control mode of nutrient solution is to achieve the precise control of nutrients compositions. However, the recent researches on optimal control are mainly based on EC and pH, which is not efficient on the composition control. Furthermore, for the correlations between vegetable growth and nutrient solution compositions are extremely complicated, it is hard to model these correlations and to achieve optimal control of nutrient solution. To handle the above challenges, this thesis proposes an optimal control method for nutrient solution based on Q-Learning and develops a self-adaptive nutrient solution control system according to this method.The main contents of this thesis are as follows:1. Research on control algorithm of nutrient solution based on Q-learningThis part is the core of the thesis. The main work includes: defining the S-A (status and action) representation, and taking the photosynthetic rate to evaluate the vegetable growth status; proposing a nutrients control algorithm based on Q-Learning.2. Hardware design of self-adaptive control system of nutrient solutionHardware makes use of the monitoring result about vegetable growth status and nutrient compositions to carry out the control decision. The hardware design work includes hardware structure design, electrical device drive circuit and so on.In addition, we conduct experiments to test the performance of this control system. Experimental results show that the proposed method is more efficient than the method based on the control of EC and pH.