机械手外文翻译

本科毕业设计(论文)外文翻译（附外文原文）学院：机械与控制工程学院课题名称：搬运机械手的结构和液压系统设计专业(方向)：机械设计制造及其自动化（机械装备）班级：学生：指导教师：日期： 2015年3月10日Proceedings of the 33rd Chinese Control Conference July 28-30, 2014, Nanjing, ChinaThe Remote Control System of the ManipulatorSUN Hua, ZHANG Yan, XUE Jingjing , WU ZongkaiCollege of Automation, Harbin Engineering University, Harbin 15000E-mail:Abstract: A remote control system of the 5 degree of freedom manipulator was designed. This manipulator was installed into our mobile robot to constitute a remote rescue robot. The Denavit-Hartenberg method was used to establish the kinematic models and the path planning of the manipulator was researched. The operator could remote control the manipulator by the interactive interface of PC which could display moving picture and various data of the manipulator. The servos of the manipulator were controlled by the slave FPGA controller. In addition, the slave FPGA controller communicated with the PC via the wireless communication module. Owing to the embedded Nios II program and IP (Intellectual Property) core generating PWM waves in FPGA, the system could control the multiple servos fast and flexible. In order to achieve real-time operation and simulation, the interactive interface was established by the mixed programming of VC and MATLAB. Key Words:The manipulator; Remote control; Denavit-Hartenberg; FPGA; Human-computer interaction1 IntroductionWith the development of the microelectronic technique and the computer technology, the manipulator has become essential equipment in the manufacturing industry. As we all known, the manipulator is usually applied to accomplish dull, onerous and repeated physical work, especially used to substitute the manual operation under the dangerous and the hazardous environment such as the corrosion and the high temperature.In this paper, the manipulator was installed our mobile robot. The tele-operation system of this manipulator was designed. The whole system is onstituted by PC and slave FPGA. The operator can remote control the manipulator by PC. The wireless communication was used for transmitting data between PC and FPGA. FPGA is controller of the the manipulator in the mobile robot. FPGA has the abundant internal resource and IP cores.And a central control option was built via an embedded Nios II program and an IP core in FPGA. Furthermore, Verilog language was adopted to design the IP core which generated digital PWM waves for controlling the manipulator. Therefore, this system could reach higher precision and easy to debug.MATLAB software was adopted to build the kinematic models of manipulator. And using D-H (the acronym of Denavit-Hartenberg) method to solve the forward and inverse kinematic equations of the manipulator, to analyze the motivation, to plan and track the motion’s path.In addition, a good interface of human-computer interaction was enhanced in the remote control system of the manipulator in PC. Moreover, the manipulator simulation technology was built by using the mixed programming of VC and MATLAB. Thus, the motion choreographs was got quickly and easily, also greatly saved time and cut the cost.2 Manipulator Model and Path PlanningAt first, the motion model of the manipulator was built. Then, the kinematic simulation and its path planning were researched. These works provided the foundation for the design of the remote control system of the manipulator.2.1 Motion Model of the ManipulatorThe manipulator was regarded as an open loop kinematic chain. It was constituted by five rotary joints. And its one end was fixed on a base while the other end was used to achieve the ability of grabbing. Therefore, it is better to establish a chain coordinate frame as shown in Fig.1. The terminal position and attitude was determined via using forward kinematic equation after knowing the rotating angle of every joint. The D-H parameter table shown as Table 1 was established by using the frames in Fig.1.Fig.1 Coordinate frames of mechanical armTable 1 D -H Parameters of the Robot ArmDue to D -H method:T =A n+1n+1n =(Cθn+1−Sθn+1Sθn+1Ca n Cθn+1Ca n 0a n −Sa n −Sa n d n+1Sθn+1Sa n Cθn+1Sa n 00Ca n Ca n d n+101) Where Cθn+1=cos θn+1 , Sθn+1=sin θn+1 , Ca n =cos a n , S a n =sin a n . The transformation matrix of every joint was given by equation (2).T 10=(cos θ1sin θ1sin θ1cos θ1000000001001) T 21=(cos θ2−sin θ200001d 1−sin θ2−cos θ2000001) T 32=(cos θ3−sin θ3sin θ3cos θ3000000001d 201) T 43=(cos θ4−sin θ40000−1−d 3sin θ4cos θ4000001) T 54=(cos θ5−sin θ5sin θ5cos θ5000000001d 401) T 50=(n x n x n y n y n x n x n y n y n z n z 00n z n z 01)=T 10T 2∗1T 3∗2T 4∗3T 5∗4 (2) Where unit vector n →,o →,a → in equation (2) was n →=normal , n →=orientation ,n →=approacℎ, n→=position . Parameters of mechanical arm were given by d 1=85mm , d 2=116mm , d 3=85mm , d 4=95mm . Therefore the forward kinematic equation was determined by taking every parameter in equation (3).P 50=(180Cθ1S (θ2+θ3)+116Cθ1Sθ2180Sθ1S (θ2+θ3)+116Cθ1Sθ285+116Cθ2+180C (θ2+θ3)) (3)In practical application, the manipulator was adopted to grab objects. This required thatthe fixed position was given from terminal to target location. That was the inverse kinematic analysis of manipulator. Inverse transformation was used to determine angle of every rotary joint toward the established coordinates. And the used method of inverse transformation was the common method to solve such problem (this method also known asalgebraic method).Using inverse transformation T n n−1−1 separately to the left multiplicationwith T =50T 10T 2∗1T 3∗2T 4∗3T 5∗4 , the angle of every rotary joint （θ1 θ2 θ3 θ4 θ5）wasdetermined. Owing to these results, the rotary angles （θ1θ2θ3）at terminal position of manipulator were totally decided by the target position [P x P y P z ]. Angle θ4 was used to change terminal attitude of the manipulator and it was changed by the known normal vector. However, angle θ5, was decided by the size of target object.2.2 Motion Simulation of the ManipulatorThe manipulator model was built and simulated via MATLAB toolbox. We could verify the rationality of the mathematical model. While the MATLAB model was established by table 1 and shown as Fig.2Fig.2 MATLAB simulation of the manipulatorComparing to the Fig.1 and Fig.2, the simulation model of the manipulator was coincided to the reference frame model. That was to say, the given coordinate frame was correct. These results also could be proved by the determined inverse kinematic equations via MATLAB shown in the table (2) and table (3).The target position was solved by forward kinematics. After that, the rotary angles were calculated by inverse kinematical equation. It turned out that these rotary angles coincided to the given angles. Therefore, these results verified the correctness of forward and inverse kinematical equation.Table (2) Forward Kinematics AnalyzeTable (3) Inverse Kinematics Analyze3 Path Planning of the ManipulatorThe total displacement of joint was calculated by inverse kinematical equation when the manipulator moved to new position. Thus, the manipulator could move to new position. Although the manipulator finally moved to the expected position in such condition, the motion of the manipulator between these two points was unknown. Due to space limitations, motion and some certain position requirements, the manipulator was often unable to move as the above mentioned method. Therefore, the motion path was designed to coincide with the limited conditions.In this paper, we could use these certain limitations to decide some expected points. Andthese expected points were used to match the planning path of the manipulator’s movement. Owing to the planning path, coordinate in every part could be calculated. The rotary angle of every joint was calculated via inverse kinetical equation and these angles realized the movement of planning path. Movement of the manipulator was shown in Fig.3 (Where‘?’ represented the points would be passed by the manipulator; ‘*’represented the expected points of every segment; ‘-’represented path planning of the manipulator). In Fig.3, we could see that the motion of the manipulator passed every planning point and the movement path coincided to the planning path.Fig.3 The path planning simulation of the manipulator4 Remote Control System of the Manipulato rThe remote control system of the manipulator contains the main PC and the slave FPGA controller using DE2 Board of ALTER Company. The motors of the manipulator were controlled by multipath PWM waves. And the PWM waves were generated by IP core. The FPGA controllerCommunicated with PC via wireless serial port. While in the PC interaction, the operator could observe the move of the manipulator in real-time and tele-control the motion of the manipulator. Also every movement of manipulator could be observed in advance via the simulation technique. The general design of the manipulator remote control system was shown in Fig.4.Fig.4 The block diagram of the remote control system4.1 Control Mode of the ManipulatorThere were two control modes of the manipulator. One mode is that the inverse kinematical equations are calculated by FPGA straightly to determine angle of every rotary joint. Thus, the control of the manipulator was achieved. The advantage of this mode is more direct and independent to finish the control of the manipulator without the external devices. At the same time, this mode has large quantities of calculations, which occupy more internal storage and running time of FPGA. Resources of FPGA are wasted under this mode.The other mode accomplished the control of the manipulator by using VC and MATLAB in PC. Using VC and MATLAB finished a large number of complex calculations and determined angle of every rotary joint. And the angle results were transmitted to FPGA in order to accomplish the control of the manipulator. This manner saved lots of internal storage and running time. In addition, FPGA could finish other works under this mode. But the manipulator was not under fast control in this mode.In this system, a new mode was adopted in the manipulator remote control system depending on the advantages of the two modes. Specifically, when the manipulator accomplished the specified and repeated movement the former mode was adopted under direct control by FPGA. When the manipulator wanted to achieve new motions the latter mode was used to be commanded by orders from PC. This new mode was made good use of advantages of the two modes in the above. And this new mode lightened computational burden and improved working efficiency of the manipulator.4.2 SOPC Design for the Remote Control SystemMovement of the manipulator was controlled by servos. And the servos were controlled by PWM waves with the cycle of 20ms. Pulse width of these PWM waves was 0.5~2.5mscorresponding to the rotary angle of servo with -90 degree to 90 degree. High precision of PWM waves were generated by IP core via Verilog in this system. The results were shown in Fig.5. PWM waves controlled rotary angles of the servos via the servo drivers.Fig.5 The PWM IP coreMultiple of IP cores were able to be downloaded into FPGA. And multiple PWM waves with high precision were generated in the output. As shown in Fig.6, the pulse width of these waves could be settled by program of Nios II. The movement of the manipulator was more flexible and in higher precision in this system.Fig.6 The IP cores generating PWM waveThe movement of the manipulator was accomplished by the duty ratio of PWM waves. Formula (4) inverted rotary angle θn to the corresponding amount of the duty ratio of PWM waves. The duty ratio of PWM waves corresponded to the Nios II output.PWM n=1000000−(θn∗50000+75000)（4）Wireless serial of 9600 baud rate was used to transmit the coordinate and the angle information from host computer to FPGA. After that, the data and orders were analyzed by FPGA Then FPGA transmitted the movement results to interactive interface of host computer via wireless transition model. This communication was realized through adding UVRT communication protocol to FPGA.4.3 The Interactive Interface of the Remote Control SystemThe interactive interface of the remote control system was shown in Fig.7. There were some functions in the interactive interface: video observation, the manipulator control and the simulation modeling.At first, the manipulator video could be seen from camera to interactive interface. The operator could monitor the manipulator in real-time.Secondly, the angle and the coordinate could be set in control zone of the interactive interface. The angle of the manipulator could be set independently to each single joint. In addition, the angle setting could be shown in real-time in the list of interactive interface (as shown in Fig.7). In the set of coordinates, judging of coordinate setting assured that the total coordinates could achieve to the target points. Thus the manipulator could be controlled to move in the settled path depend on the angle information.Lastly, the MATLAB robot toolbox was embedded into this interactive interface. One interface was integrated both the control and simulation of the manipulator. MATLAB robot toolbox was directly used by interactive interface in the manipulator modeling. Each group of information was simulated separately in order to detect whether each movement was correct. And the general simulation could test whether movement arrangement of the manipulator was reasonable. Combining with multiple simulation methods made the movement arrangement more flexible, the operation of the manipulator simpler and interface interaction more perfect.Fig.7 The interactive interface of the manipulator5 Experiment and SimulationIn order to verify properties of the remote control system of the manipulator, experiments of the system were under way and were comparing to the simulation system. To be specific, manipulator modeling was built by interactive interface and a group of coordinates could be designed. These coordinates were transmitted to FPGA, which controlled the servos to accomplish the movement of the manipulator. Joint angles, the terminal coordinates shot by interface video. The simulation results were shown in paring the real movement and the simulation results, we could see that the manipulator modeling and control of the interactive interface design comforted to the design requirement. The comparing between experiment and simulation was shown in Fig.8.Fig.8 The experiment and the simulation6 ConclusionIn the experiment, the 5-DOF manipulator modeling was simulated by MATLAB. In the slave FPGA board, control of the manipulator was accomplished via IP core based on the Verilog language. That greatly reduced design of the peripheral circuit, cut the cost, improved the precision and made the movement smoother without shaking. While in the interactive interface, the mixed programming method of VC and MATLAB was embedded into the MATLAB simulation function. Thus the operability of this manipulator was enhanced. The system had a good ability of interactive interface. The whole system was verified and achieved to the expected effect. One new thing in this system was that embedded the MATLAB robot toolbox in the interactive interface. The D-H modeling, path planning and tele-operation and so on were accomplished by using this interactive interface directly. Compared to the other development tools, this interactive interface had portability, good compatibility, short development cycle and simple operation.References[1] Saeed B. Niku write, Sun Fuchun, Zhu Jihong, Liu Guodongetc translate: Robotics Introduction, Beijing, ElectronicIndustry Press, 2004(1):60-63,132-137.[2] Brady, M.J.M.Hollerbach, T.L.Johnson, T.Lozano-Perez, andM.T.Mason, editors, Robot Motion; Planning and Control,MIT Press, Cambridge, Mass, 1982.[3] Paul Richard P., Robot Manipulators, Mathematics,Programming, and Control, The MIT Press 1981.[4] Li Jian, Design and Research of Multi-DOF Robot, Masterdegree theses of master of university of technology, Chineseacaedemy of sciences, 2009?20-31.[5] Cheng Liyan, Fei Ling, Su Zelang, The 5-DOF ManipulatorKinematics Simulation Analysis Based on MATLAB,Mechanical Research & Application, 2011(06).[6] Zhang Puxing Jia Qiuling, Mechanical Arm Multi-channelServo Control Design based on FPGA, Small and specialelectrical machine. 2011, 39(4)第33届中国控制会议论文集中国，南京，2014年28-30日机械手的远程控制系统孙华、张媛、薛晶晶、吴宗凯哈尔滨工程大学，哈尔滨15000学院，自动化专业电子油箱：sunhuas@摘要：一种5自由度机械手的远程控制系统的设计。