丁一, 姬伟, 许波, 陈光宇, 赵德安. 苹果采摘机器人柔顺抓取的参数自整定阻抗控制[J]. 农业工程学报, 2019, 35(22): 257-266. DOI: 10.11975/j.issn.1002-6819.2019.22.031
    引用本文: 丁一, 姬伟, 许波, 陈光宇, 赵德安. 苹果采摘机器人柔顺抓取的参数自整定阻抗控制[J]. 农业工程学报, 2019, 35(22): 257-266. DOI: 10.11975/j.issn.1002-6819.2019.22.031
    Ding Yi, Ji Wei, Xu Bo, Chen Guangyu, Zhao Dean. Parameter self-tuning impedance control for compliance grasp of apple harvesting robot[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2019, 35(22): 257-266. DOI: 10.11975/j.issn.1002-6819.2019.22.031
    Citation: Ding Yi, Ji Wei, Xu Bo, Chen Guangyu, Zhao Dean. Parameter self-tuning impedance control for compliance grasp of apple harvesting robot[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2019, 35(22): 257-266. DOI: 10.11975/j.issn.1002-6819.2019.22.031

    苹果采摘机器人柔顺抓取的参数自整定阻抗控制

    Parameter self-tuning impedance control for compliance grasp of apple harvesting robot

    • 摘要: 为了实现苹果机器采摘过程中的柔顺抓取以减小果实损伤,该文在对苹果抓取过程的力学特性变化规律分析的基础上,提出了苹果采摘机器人柔顺抓取的参数自整定阻抗控制方法。首先,利用Burgers黏弹性模型表征苹果的流变特性,将抓取过程分为匀速加载、夹持减速、应力松弛3个阶段,在此基础上求解获得苹果形变量随时间的变化规律和果实接触力与变形量的变化关系。然后,求解出所设计的基于力的阻抗控制系统的期望输入以及抓取环境接触力模型。最后,针对阻抗控制器参数对接触力的影响,构造阻抗参数自整定变化函数,完成改进阻抗控制系统设计。仿真及试验结果表明:依据果实抓取模型及变形规律求解期望位置的方式来模拟末端执行器对苹果的抓取过程是可行的,所建立的抓取环境接触力模型在一定程度上能够避免将环境模型简化为一阶模型而产生的误差。改进阻抗控制得到的期望抓取力更加平顺,其超调量约为2.3%,接触力调节时间减小到0.48 s,接触力的超调量约为2%,较未改进阻抗控制的接触力超调量减小了37.5%。研究结果可为苹果采摘机器人的柔顺控制方法提供参考。

       

      Abstract: Abstract: The picking operation of agricultural fruit and vegetable harvesting robot is the key link to realize the automatic harvesting of fresh fruits and vegetables. The stable holding of fruits can be achieved by controlling the output grasping force of the end actuator. However, due to the complexity and particularity of the working object, the harvesting robot is easy to cause fruit damage when grasping fruit, and the success rate of grasping is low. Therefore, to reduce the mechanical damage, improve the quality of fruit picking, and realize the compliance grasping of robot picking is one of the key points of the research of harvesting robot. In order to reduce the damage caused by apple harvesting and achieve a more compliance harvesting, the mechanical characteristics of apple during the process of grasping and the improvement of impedance control algorithm were studied. Firstly, Burgers viscoelastic model was used to characterize the rheological properties of apples, 10 groups of apple samples were tested for uniaxial compression creep, and through Burgers creep model to fit the test data, the viscoelastic parameters of the creep model of apple were obtained. The grasping process was divided into three stages: constant loading, overload deceleration and stress relaxation. On this basis, the change equation of apple deformation with time was obtained, and the change curves of apple deformation with time and the relationship between fruit contact force and deformation under different grasping velocity were made by using the change equation of apple deformation with time in three stages. Secondly, the end-effector drive control system was modeled, and the mathematical model of the end-effector was obtained. Considering the contact force and deformation of apple change with time in the process of grasping, the apple deformation was equivalent to the forward displacement of the end effector finger, and it was the expected input of the impedance control system. Aiming at the uncertainty of the target apple's stiffness and position parameters, the contact force model under grasping environment with the grasping speed of 3 mm/s was solved. Finally, the influence of inertia parameters, damping parameters and stiffness parameters on the contact force was analyzed. Based on the hyperbolic secant function and the influence of three impedance parameters, three impedance parameters self-tuning functions were constructed to complete the design of the improved impedance control system. The simulation and experimental results show that it was feasible to simulate the grasping process of apple with the end-effector by solving the desired position according to the fruit grasping model and the change law of deformation with time. Besides, to a certain extent, the established contact force model of grasping environment could avoid the error caused by simplifying the environment model to the first order model, and the three functions designed could meet the requirements of adaptive impedance control. The desired force obtained by the improved impedance control was smoother, the overshoot was about 2.3%. The response speed was faster, and the adjustment time of contact force was shorter about 0.48 s. The contact force overshoot was about 2%, which was 37.5% less than that of the original force-based impedance control, and had obvious advantages.

       

    /

    返回文章
    返回