苹果采摘机器人末端执行器无损主动抓取技术研究

发布时间：2017-12-31 01:38

本文关键词：苹果采摘机器人末端执行器无损主动抓取技术研究　出处：《江苏大学》2017年硕士论文　论文类型：学位论文

【摘要】：在果蔬生产过程中,采摘环节是最耗时和耗力的环节,人工采摘的成本约占整个生产成本的50%~70%。开展采摘机器人研究,不仅有利于解决人口老龄化和社会城镇化导致的劳动力日益短缺问题,还能促进我国农业科技进步,加速实现农业现代化。由于果蔬组织柔软、易损伤且生长程度不一,相互差异较大,导致采摘的损伤率较高。因此开展末端执行器抓取规划和控制策略的研究对实现采摘机器人的无损采摘具有重要意义,本文从采摘机器人末端执行器稳定抓取规划以及抓取力主动柔顺控制两个角度出发对采摘机器人无损抓取技术进行研究,主要研究内容如下:(1)基于苹果的形态结构特性和力学特性,设计了末端执行器,确定了果梗切割方式。为提高末端执行器环境感知能力,设计了包括力传感器、碰撞传感器和视觉传感器在内的末端执行器传感器系统,并建立了末端执行器的数学模型。(2)以力封闭作为理论基础,分析了两指稳定抓取条件,制定了两指稳定抓取策略。同时根据三指实物模型,建立三指抓取数学优化模型,通过神经网络对优化模型进行拟合分析,得到抓取位置与抓取性能之间的关系。(3)根据苹果采摘机器人末端执行器对抓取力的要求,设计了改进阻抗控制器。通过对阻抗控制的稳定性分析得知力误差与环境刚度和环境位置密切相关,为了实现对参考力的快速跟踪,在力控制器中,加入了递推最小二乘法(RLS)实现环境刚度参数的在线辨识,根据辨识得到刚度参数,调整位置输入信号,减小稳态力误差。根据遗忘因子在RLS算法中的作用,对RLS算法进行了修正,提出了一种基于误差变遗忘因子的RLS算法,克服了环境参数未知对阻抗控制的影响。在Matlab中对所提出的控制算法进行验证,仿真结果表明了所设计的阻抗控制器具有良好的性能。(4)分别利用改进的阻抗控制算法和普通阻抗控制算法控制末端执行器进行苹果抓取试验。试验结果表明,相对于普通阻抗控制,改进的阻抗力控制在抓取过程中能够快速跟踪给定力,提高抓取的稳定性,减少抓取损伤,能够实现主动柔顺抓取,更加符合采摘机器人抓取的应用要求。
[Abstract]:In the process of fruit and vegetable production, picking is the most time-consuming and labor-consuming link, the cost of manual picking accounts for about 50% of the total production cost. It can not only solve the problem of labor shortage caused by aging population and urbanization, but also promote the progress of agricultural science and technology, accelerate the realization of agricultural modernization, because of the soft organization of fruits and vegetables. It is easy to damage and grow with different degrees, which leads to high damage rate of picking. Therefore, it is very important to carry out the research on the planning and control strategy of end effector grab in order to realize the non-destructive picking of picking robot. In this paper, the acquisition technology of the harvesting robot is studied from two aspects: the stable grab planning of the end actuator and the active flexibility control of the grab force. The main research contents are as follows: (1) based on the morphological and mechanical characteristics of apple, the end effector is designed, and the cutting mode of fruit-stem is determined. The end actuator sensor system including force sensor, collision sensor and vision sensor is designed, and the mathematical model of end actuator is established. The condition of two-finger stable grab is analyzed and the strategy of two-finger stable grab is worked out. At the same time, according to the three-finger physical model, the mathematical optimization model of three-finger grab is established, and the optimization model is fitted and analyzed by neural network. The relationship between the position of grab and the performance of grab is obtained.) according to the requirements of the end actuators of the apple picking robot for grasping force. Through the stability analysis of impedance control, it is found that the force error is closely related to the environment stiffness and the environment position. In order to realize the fast tracking of the reference force, the force controller is used in the force controller. Recursive least square method (RLSs) is added to realize the on-line identification of environmental stiffness parameters, according to which the stiffness parameters are obtained and the position input signals are adjusted. According to the function of forgetting factor in RLS algorithm, the RLS algorithm is modified, and a RLS algorithm based on error variable forgetting factor is proposed. It overcomes the influence of unknown environmental parameters on impedance control and verifies the proposed control algorithm in Matlab. The simulation results show that the designed impedance controller has good performance. The improved impedance control algorithm and the common impedance control algorithm are used to control the end actuator for the apple grab test. The experimental results show that. Compared with the ordinary impedance control, the improved impedance force control can quickly track the fixed force, improve the stability of the grab, reduce the grab damage, and realize the active compliant grasp. More in line with the picking robot grab application requirements.
【学位授予单位】：江苏大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】