微创手术机器人震颤机理及其抑制方法研究

发布时间：2018-03-02 05:17

本文关键词： 微创手术机器人震颤机理零相位自适应模糊卡尔曼滤波器模糊神经网络滑模控制器震颤抑制　出处：《天津工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：微创外科手术机器人具有灵活的自由度、运动缩放和3D视觉等优点,是高新科技与先进医学结合的典型产物。然而,在手术过程中,手术器械末端存在不同程度的震颤现象,降低了手术器械的运动精度,影响了手术质量及安全性,增加了如缝合、打结等精细动作的难度。针对上述问题,本文分析了机器人震颤产生的主要机理,提出了相应的震颤抑制方法,构建了微创手术机器人主从控制系统实验平台并进行验证。论文的主要研究内容如下:首先,本文试图从医生的不光滑操作指令、机器人动力学行为等角度研究震颤产生的机理,通过仿真及实验得出导致手术器械末端震颤现象的主要因素。其次,针对医生手部生理震颤现象,提出了基于新型零相位自适应模糊卡尔曼滤波器(ZPAFKF)的震颤抑制方法。通过对震颤信号的零相位采集滤波并以相同幅值相反相位实时叠加至原始指令信号中,得到光滑无生理震颤的理想指令,避免了传统低通滤波器容易造成有效信息丢失及相位延迟的缺点。再次,针对微创手术机器人动力学行为特别是关节内非线性摩擦导致的"粘滑行为",提出了基于模糊神经网络滑模控制器(FNNSMC)的震颤抑制方法。在传统滑模控制器的基础上,通过对各关节摩擦力矩的离线辨识以及神经网络、模糊逻辑控制器对无关干扰的在线补偿,得到平滑的末端运动轨迹点。避免了传统PD控制器在低速状态下"爬行"及"平顶"现象和传统滑模控制器在控制律切换时的"抖振"现象,同时构造李雅普诺夫函数,证明了控制器的稳定性。最后,为验证以上两种震颤抑制方法的有效性,进行了仿真和实验研究。结果表明,两种方法能有效地抑制微创手术机器人手术器械末端震颤现象,提高末端点的运动精度。
[Abstract]:Minimally invasive surgical robot has the advantages of flexible freedom, motion scaling and 3D vision. It is a typical product of the combination of high and new technology and advanced medicine. However, in the process of surgery, there are different degrees of tremor at the end of surgical instruments. The motion accuracy of surgical instruments is reduced, the quality and safety of operation are affected, and the difficulty of fine movements such as suture and knot is increased. In view of the above problems, the main mechanism of robot tremor is analyzed in this paper. The experiment platform of master-slave control system of minimally invasive surgery robot is constructed and verified. The main contents of this paper are as follows: firstly, this paper attempts to use the doctor's non-smooth operation instruction. In this paper, the mechanism of tremor generation is studied from the point of view of robot dynamic behavior, and the main factors leading to the terminal tremor of surgical instruments are obtained by simulation and experiments. Secondly, aiming at the phenomenon of physiological tremor in the doctor's hand, A new method of vibration suppression based on a novel zero-phase adaptive fuzzy Kalman filter (ZPAFKF) is proposed. The zero phase acquisition and filtering of the tremor signal and the real-time superposition of the same amplitude and opposite phase into the original command signal are carried out. The ideal instruction of smooth non-physiological tremor is obtained, which avoids the disadvantages of traditional low-pass filter which can easily lead to the loss of effective information and phase delay. Again, Based on fuzzy neural network sliding mode controller (FNNSMC), a new method for vibration suppression of minimally invasive manipulators, especially the "sticking-slip behavior" caused by nonlinear friction in joints, is proposed. By off-line identification of joint friction torque and neural network, the fuzzy logic controller can compensate the irrelevant disturbance on line. The smooth end motion locus is obtained, which avoids the phenomenon of "crawling" and "flattening" in the low speed state of the traditional PD controller and the "buffeting" phenomenon of the traditional sliding mode controller when the control law is switched. At the same time, the Lyapunov function is constructed. The stability of the controller is proved. Finally, in order to verify the effectiveness of the two methods mentioned above, the simulation and experimental research are carried out. The results show that the two methods can effectively suppress the terminal tremor phenomenon in the surgical instruments of the minimally invasive surgical robot. Improve the kinematic accuracy of the end point.
【学位授予单位】：天津工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】