Control of a Robotic Manipulator Using Artificial Neural Networks with On-line Adaptation

An adaptive neural system for positioning control of a PUMA 560 manipulator is presented". The computed torque method was implemented with a Multi-Layer Perceptron with on-line learning. The control scheme is implemented into two phases. The first one is the off-line phase in which the neural network is trained with previously known control actions. The second one is the on-line phase in which the neural network parameters are adapted while controlling the manipulator. The control system is able to respond to changes in the manipulator model and to load disturbances. As will be shown, control system performance is improved with the on-line learning strategy presented in this paper.     

基于PUMA OPEN1．2的发动机动态特性研究

简述发动机性能测试、动态测试的重要性,给出发动机动态特性的定义,详细介绍了AVL公司的发动机动态测试系统PUMA OPENl．2的组成和测试方法。对某型汽油发动机进行稳态和动态测试,并对测试结果进行了比较分析,为发动机的动态控制设计提供了指导依据。     

新一代GPS测量不确定度与管理分析

在新一代产品几何技术规范(Next Generation Geometrical Product Specification and Verification,NGGPS)不确定理论基础上,分析各种不确定度之间的关系,重点讨论不确定度管理(Frocedure for Uncertainty Management,PUMA)测量不确定度管理程序所涉及的贡献因素、评定模型和评定关系等关键环节.设计不确定度管理程序,实现贡献因素分析、不确定度自动计算和系统评价,为实现不确定度评定管理的规范统一奠定基础.     

RBF网络在PUMA560机器人运动学逆解中的应用

介绍了RBF网络及RBF网络基函数中心选取的重要性,将进化优选算法用于网络中心的选取.并利用D-H方法对PUMA560机器人进行研究,将正运动学推导结果作为训练样本.采用6个相同的12输入、单输出的RBF网络,实现PUMA560机器人运动学逆解计算,利用该方法大大减少了传统方法中的公式推导.计算结果表明,用RBF网络解决机器人运动学逆解问题精度高,收敛速度快.     

RNA Interference Targeting Slug Increases Cholangiocarcinoma Cell Sensitivity to Cisplatin via Upregulating PUMA

Slug is an E-cadherin repressor and a suppressor of PUMA (p53 upregulated modulator of apoptosis) and it has recently been demonstrated that Slug plays an important role in controlling apoptosis. In this study, we examined whether Slug's ability to silence expression suppresses the growth of cholangiocarcinoma cells and/or sensitizes cholangiocarcinoma cells to chemotherapeutic agents through induction of apoptosis. We targeted the Slug gene using siRNA (Slug siRNA) via full Slug cDNA plasmid (Slug cDNA) transfection of cholangiocarcinoma cells. Slug siRNA, cisplatin, or Slug siRNA in combination with cisplatin, were used to treat cholangiocarcinoma cells in vitro. Western blot was used to detect the expression of Slug, PUMA, and E-cadherin protein. TUNEL, Annexin V Staining, and cell cycle analysis were used to detect apoptosis. A nude mice subcutaneous xenograft model of QBC939 cells was used to assess the effect of Slug silencing and/or cisplatin on tumor growth. Immunohistochemical staining was used to analyze the expression of Slug and PUMA. TUNEL was used to detect apoptosis in vivo. The results showed that PUMA and E-cadherin expression in cholangiocarcinoma cells is Slug dependent. We demonstrated that Slug silencing and cisplatin both promote apoptosis by upregulation of PUMA, not by upregulation of E-cadherin. Slug silencing significantly sensitized cholangiocarcinoma cells to cisplatin through upregulation of PUMA. Finally, we showed that Slug silencing suppressed the growth of QBC939 xenograft tumors and sensitized the tumor cells to cisplatin through PUMA upregulation and induction of apoptosis. Our findings indicate that Slug is an important modulator of the therapeutic response of cholangiocarcinoma cells and is potentially useful as a sensitizer in cholangiocarcinoma therapy. One of the mechanisms is the regulation of PUMA by Slug.     

Adaptive Learning with Large Variability of Teaching Signals for Neural Networks and Its Application to Motion Control of an Industrial Robot

Recently, various control methods represented by proportional-integral-derivative (PID) control are used for robotic control.To cope with the requirements for high response and precision, advanced feedforward controllers such as gravity compensator,Coriolis/centrifugal force compensator and friction compensators have been built in the controller.Generally, it causes heavy computational load when calculating the compensating value within a short sampling period.In this paper, integrated recurrent neural networks are applied as a feedforward controller for PUMA560 manipulator.The feedforward controller works instead of gravity and Coriolis/centrifugal force compcusators.In the learning process of the neural network by using back propagation algorithm, the learning coefficient and gain of sigmoid function are tuned intuitively and empirically according to teaching signals.The tuning is complicated because it is being conducted by trial and error.Especially, when the scale of teaching signal is large, the problem becomes crucial.To cope with the problcm which concerns the learning performance, a simple and adaptive learning technique for large scale teaching signals is proposed.The learning techniques and control effectiveness are evaluated through simulations using the dynamic model of PUMA560 manipulator.     

A QP Artificial Neural Network inverse kinematic solution for accurate robot path control

In recent decades, Artificial Neural Networks (ANNs) have become the focus of considerable attention in many disciplines, including robot control, where they can be used to solve nonlinear control problems. One of these ANNs applications is that of the inverse kinematic problem, which is important in robot path planning. In this paper, a neural network is employed to analyse of inverse kinematics of PUMA 560 type robot. The neural network is designed to find exact kinematics of the robot. The neural network is a feedforward neural network (FNN). The FNN is trained with different types of learning algorithm for designing exact inverse model of the robot. The Unimation PUMA 560 is a robot with six degrees of freedom and rotational joints. Inverse neural network model of the robot is trained with different learning algorithms for finding exact model of the robot. From the simulation results, the proposed neural network has superior performance for modelling complex robot’s kinematics.     

A Workcell for the Development of Robot-Assisted Surgical Procedures

This paper describes a robotic workstation for the development of new robot-assisted surgical procedures. This work is motivated by the difficulties and cost associated to the development of surgical robots, often requiring large investments and several re-designs which limit wider use of this technology. The approach presented here consists of using a general purpose robotic workcell to develop the hardware and the surgical aspects of new robot-based surgical systems, before committing to a completely new system design. The workcell is based on a clean room PUMA 260 manipulator, suitably enhanced to expand and improve its capabilities, and on a vision-based operator interface. Two new robot-assisted surgical procedures have been developed and tested using this set-up: percutaneous discectomy and knee osteoctomy. By using the robotic workcell, engineers and surgeons are able to define many aspects of the two procedures, such as surgical gestures, workspace of the robot, and calibration procedures, without incurring a large, up-front investment. First, the article describes the configuration of the workcell, the enhancements to the PUMA manipulator and the surgical procedures developed with this setup. Then the results of the tests and the lessons learned using the workcell are discussed in some detail.     

Photochemical trajectory modelling of ozone during the summer PUMA campaign in the UK West Midlands

A Photochemical Trajectory Model (PTM), coupled with the Master Chemical Mechanism, is set up for summertime conditions and used to simulate pollutant levels measured in Birmingham, UK during the summer Pollution in the Urban Midlands Atmosphere (PUMA) campaign. In general, ozone is modelled well by the PTM, with reasonably good correlation and ≈ 90% of afternoon and evening concentrations within a factor of two of measurements. The contribution of local biogenic emissions in the West Midlands to ozone formation during this period was also assessed and found to be of minor importance. Initially, the ozone episode of 26th June 1999, where levels of up to 87 ppb were observed was not reproduced by the UK PTM model. Sensitivity studies showed that the major cause for this was transport uncertainty in the origin and pathway of the air during conditions of slow moving anticyclonic conditions. The results indicate that the ozone episode was caused by recirculation of air polluted within the UK, with additional precursor emissions over the Netherlands, Belgium and France.     

Adaptive learning with large variability of teaching signals for neural networks and its application to motion control of an industrial robot

Recently, various control methods represented by proportional-integral-derivative (PID) control are used for robotic control. To cope with the requirements for high response and precision, advanced feedforward controllers such as gravity compensator, Coriolis/centrifugal force compensator and friction compensators have been built in the controller. Generally, it causes heavy computational load when calculating the compensating value within a short sampling period. In this paper, integrated recurrent neural networks are applied as a feedforward controller for PUMA560 manipulator. The feedforward controller works instead of gravity and Coriolis/centrifugal force compensators. In the learning process of the neural network by using back propagation algorithm, the learning coefficient and gain of sigmoid function are tuned intuitively and empirically according to teaching signals. The tuning is complicated because it is being conducted by trial and error. Especially, when the scale of teaching signal is large, the problem becomes crucial. To cope with the problem which concerns the learning performance, a simple and adaptive learning technique for large scale teaching signals is proposed. The learning techniques and control effectiveness are evaluated through simulations using the dynamic model of PUMA560 manipulator.