【作者】 陈刚；

【导师】 张立彬；

【作者基本信息】 浙江工业大学 ， 机械电子工程， 2003， 硕士

【摘要】 基于柔性气缸，设计了一种新型的气动弯曲关节。在一定气体压力作用下，研究弯曲关节的径向变形和轴向变形。建立了弯曲关节的静态模型，并通过实验论证了该模型的合理性。基于悬臂梁模型，建立了弯曲关节的动态模型，仿真结果表明，通过选择合理的结构参数，可以使弯曲关节具有很好的动态特性。 基于比例伺服技术，设计了弯曲关节的伺服控制系统。建立了伺服控制系统的数学模型。采用PID算法和模糊控制算法，实现了弯曲关节任意弯曲角的控制。采用计算力矩法和神经PID算法，实现了弯曲关节的轨迹跟踪。 用两个弯曲关节构造了一个两自由度手指，建立了手指的运动学方程。提出基于雅可比矩阵的反馈控制算法。在要求误差及其误差变化率极小的条件下，采用了误差性能最优控制算法。在考虑鲁棒性的条件下，给出了滑模控制算法。 讨论了三关节冗余手指的运动学方程，并且建立了手指的动态数学模型。运用具有三个神经元的Hopfield网络求得了冗余手指的逆运动学解。仿真结果表明，该手指具有很好的动态特性，而且运用Hopfield网络算法，手指的轨迹跟踪误差很小。基于psuedo逆，提出了冗余手指的动态控制算法。并且对手指在抓取时的夹持力进行了分析。 针对在某些场合对机械手的作用力提出的特殊要求，本文提出了一种能适应不同刚度环境、对力进行自适应控制的基于神经网络带积分环节的控制算法。仿真结果证明该算法的合理性。 随后介绍了一个三爪气动机械手的设计原理以及基于PWM原理的伺服控制系统。 最后，讨论了弯曲关节的应用前景，对未来的研究方向进行了展望。更多还原

【Abstract】 A new type of pneumatic bending joint based on the flexible cylinder is designed. We study the racial deformation and the axial deformation created by the certain gas pressure. We establish the static model of the bending joint and prove the rationality of the model by the experiment. We also establish the dynamic model of the bending joint based on the model of cantilever. The simulating results indicate that the bending joint could possess good dynamic characteristic by choosing the rational structure parameters.We design the servo control system of the bending joint based on the proportional servo technology. The mathematic model of the servo control system is established in this paper. We fulfill the control of any angle degree by using the algorithms of PID control and fuzzy control. The paper also realizes the trajectory tracking of the bending joint by adopting the algorithms of calculating moment and neural PID.We construct a finger of two DOF by using two bending joints. The kinematics equations are established. The paper presents the feedback control based on Jacobi matrix. We bring forward the error optimal control when the error and the rate of the error are required to be the least. The sliding model control is considered when the robust characteristic of the system is considered.The paper discusses the kinematics equations of the three-joint redundancy finger and establishes the dynamic mathematic model. The solution of the inverse kinematics equations is acquired by using a Hopfield NN with three neural cells. The simulating results indicate that the finger possesses good characteristic. Furthermore the error of trajectory tracking is very small by using the algorithm of Hopfied NN. We present the dynamic control algorithm based on inverse-psuedo and analyze the grasping force of the finger.According to the special requirements of the action force of the finger, the paper brings forward a algorithm which possesses force adaptation based on NN with integral controller. The simulating results prove the rationality of the algorithm.Subsequently the paper introduces the designing principle of a three-claw pneumatic manipulator and the servo control system based on PWM principle.At last, we discuss the application foreground of the bending joint and put forward the future study directions.更多还原