Studies on Control Systems of Bioproduction Machinery and Devices by Means of Artificial Intelligence

本研究旨在利用人工智能技术设计和控制耕作、播种、间伐除草和灌溉以及组织培养过程中的机械和设备。通过一系列土槽试验和田间试验，对少量耕作进行了研究，揭示了与普通硬齿机相比的特点。在此基础上，构建了一个专家系统，用于在给定的条件下选择合适的少耕机具和设计最小耕作机械。为了研制精量播种机，对大豆进行了三维图像分析，并对其形状特征进行了描述。对除草机器人进行了田间作物和杂草的图像分析，在不同条件下对杂草和杂草的识别率为68%~95%，并利用自回归模型对灌溉植物进行了最优控制，当最优调节器的权重函数在1：10范围内变化时，获得了满意的结果。利用形状记忆合金设计了一种软操作夹持器，成功地实现了原球茎的移栽，移栽间隔为7 mm，移栽速度为8s。

This study aimed at design and control of machines and devices by means of artificial intelligence in tillage, seeding, thinning weeding and irrigation as well as tissue culture process.Minimum tillage was studied by carrying out a series of soil bin tests and field experiments on spring tines, which revealed specific features in comparison to ordinary rigid tines. Then, an expert system was constructed to select proper minimum tillage tools under given conditions and design a minimum tillage machine.To develop a precision seeder, three dimensional image analysis of soy- beans was conducted and their shape characteristics were formulated. Image analysis of field crops and weeds was done for a weeding robot and was able to distinguish weeds from crops with a probability of 68-95% under various conditions.Optimum control of irrigating plants was tried by means of self regression model and satisfactory results were obtained when the weight function of the optimum regulator was varied within the range of 1:10.A robot for tissue culture operation was developed to handle small protocorm under 1mm. A soft handling gripper was designed using a shape memory alloy and succeeded in transplanting protocorms in raw with 7mm spacing at with speed of 8 seconds for each protocorm.